How to Find Company Information in the Balkans

Bosnia and Herzegovina

The front page of the Registers of Business Entities in Bosnia and Herzegovina. Photo: Screenshot

https://bizreg.pravosudje.ba/

This page provides a searchable database of available electronic court registers of business entities in the Federation of Bosnia and Herzegovina, Republika Srpska and the District of Brcko.

The electronic court registers contain information on all business/legal entities that are required to register by the Laws on Registration of Business Entities.

You can search by the name of the business entity, the registration number, abbreviation of the business entity and even a founder of the business entity.

There are even more options if you go with the advanced search option.
However, the information you get is limited: just the basic information about founders, board of managers, branch offices and subsidiaries.

If you need to find specific information on a company from Republika Srpska, you’d get better results if you go to http://bizreg.esrpska.com/Home/PretragaPoslovnogSubjekta.

Bulgaria

The front page of the Bulgarian business registry. Photo: Screenshot

http://www.brra.bg

The Bulgarian registry is extremely useful and has lots and lots of information. If you can’t read Bulgarian, go to http://www.brra.bg choose English language in the upper right corner, and click on the link, “By natural person or legal entity”. 

You need to enter the search terms in Bulgarian cyrilic, or, if you have this information, the company numbers or personal numbers of directors and owners.

Once you’re in, you have access to all company documents, even historic ones, acts of establishment, information on beneficial owners, etc. 

Croatia

The front page of the Croatian business registry. Photo: Screenshot

https://sudreg.pravosudje.hr/registar

There’s no English version and only the basic search is available. You can search by the name of the business entity, the registration number of the subject of entry, personal identification number, but not by the founder or owner of the business entity.

However, in the search results, you have a link to all the related companies, which is useful if you want to check where else the person appears. 

In the results, you’ll find shareholders’ names, addresses, financial reports, amount of share capital, subject of business, the status of the entity, the competent court for the entity, the name of the authorized representative, information on the founding acts and the statute of the entity.

If you need the company’s financial reports for the given year, you’ll find them here http://rgfi.fina.hr 

There you have an English version. You need to register with an email address. The registration is free, just insert your chosen username and email address and they’ll send you the password you need to use to login. 

Kosovo

The front page of the Kosovo’s business registry. Photo: Screenshot

https://arbk.rks-gov.net/

Searches are available in Albanian, Serbian and English and it’s pretty straightforward. Put the name of the company in the appropriate field and you’ll get the owner, the authorized person, registration date, and its capital. 

If you know the personal ID number of the person you’re interested in, you can search through all of their companies. 

Montenegro

The front page of the business registry in Montenegro. Photo: Screenshot

http://www.pretraga.crps.me:8083/

You can only search in Montenegrin and there’s no way to reach it through Google Translate.

With advanced search, you can search owners and directors names, personal numbers, contact data, what kind of business they are registered for, date of establishment, and changes that happened in the company.

For financial reports, you’ll find them at https://eprijava.tax.gov.me/TaxisPortal 

That search is also only in Montenegrin and you can’t reach it through Google Translate. 

Moldova

The Moldovan business search web-page has been under construction for a long time. Photo: Screenshot

http://webinfo.cis.gov.md/company-search

Good luck with finding company info in Moldova. The registry’s website has been under construction for months and the only way to find any info now is to go personally to the State Registration Chamber’s desk and pay for the information. 

The chamber will provide you with up-to-date official and complete info about a company.  Financial reports are not available. 

There are two useful independent websites, https://idno.md/ and https://www.bizzer.md/ where you can search for company data for free. This is not official information, so we can’t vouch for whether it is up-to-date. You can get the ownership, the date of registration and the physical address of the company.

Serbia

The front page of the Serbian Business Registry. Photo: Screenshot

http://pretraga2.apr.gov.rs/unifiedentitysearch

There is an English version of the homepage and parts of the website that do not have search parts. Databases themselves are only in Serbian. 

You can search by the registration number or name of the company. 

There’s no option to search by the names of the people involved.  

Once you get to the search, you’re provided with two search boxes. One says “Матични број”, which is the registration number – an eight digit number associated with the company. 

The other box is labeled “Пословно име”, where you can put the name of the company, if you want to do a search that way. 

Either way, you’ll get a great amount of data – owners’ names, directors, their personal numbers, contact data, capital, branches, beneficial owners (sometimes true, sometimes not), financial reports since 2016, and every change that happened in the company in the last couple of years. 

You can also obtain the original documents used to establish a company, including passport copies, off-shore information, etc. 

However, all the information is in Serbian. 

Searches by the names of people involved are available on commercial databases, such as poslovna.rs or checkpoint.rs.

Dragana Peco: How to Track Offshore Companies?

Offshore companies have got a bad press in the last few years, not surprisingly, since they play a massive role in all kinds of criminal activities.

Legality question

“There are some country destinations, some business registers, where we can’t find information on who the shareholders of the company are,” said Dragana Pećo, investigative reporter for KRIK and OCCRP.

Although establishing an offshore company is legal, people can use it for illegal activities. Pećo illustrates this with what some corrupt Balkan politicians do. “They steal the money from the budget, buy property and want to hide it,” she explained.

Tracking offshore companies has become a daily activity for her. “It’s not that I find some new offshore company every day that I should investigate, but it’s every day … [either] as a part of some cross-border project I’m working on with some colleagues, or something related to local stories,” she said. 

Secrecy, remember?

The most important tools for tracking offshore companies are business registries and databases. Some are free, while some can be pricey. The first task, however, is to find information and documents about the company in an official business registry. Usually, these kinds of information and documents cannot be found in official company registries, since it is an offshore jurisdiction – secrecy, remember? The agent who helped found an offshore company will usually ignore a journalist’s emails.

But some information can be found in other places. These companies usually have subsidiaries, she said. “What I do is, I search for the documents that they should submit to local business registries,” Peco explained. 

Information can be found in court records and different databases worldwide, such as Orbis, LexisNexis, Offshore Alert, Sayari and Arachnys. Journalists can also use opencorporates, ICIJ’s Offshore Leaks database, OCCRP’s tool for records search as well as ask for assistance from the investigators at an online platform run by OCCRP, Investigative Dashboard.

How To Prepare Your Investigation

Here is an easy step by step guide to formulating an investigation.

Step 1: Finding a story

Films such as All the President’s Men have shaped how the public, and even many reporters, view the work of investigative journalists.

Hollywood screenwriters would have you believe it is a shadowy world of clandestine meetings, codenames and tussles with the security services.

A tiny minority of stories do play out like movies – we have, after all, had cinema adaptations of the Edward Snowden leaks and The Guardian’s relations with Wikileaks recently – but most investigative journalism is much more mundane, although no less important.

You will need to work hard to develop your sixth sense – your ability to spot the unusual, the unexplained or patterns that will lead you to your next investigation.

It helps to be inquisitive and curious by nature, but you can teach yourself to sniff out those great untold stories.

Here are a few tips to build up your skills.

Spotting the unusual

To be an investigative journalist you must be able to quickly spot the unusual – what I call “smell a rat”. Without that sixth sense, stories will pass you buy and you will miss key details in your research.

Part of this skill comes from practice, but the tips in this chapter should help you sharpen your eye to those potential story ideas.

Take a careful look at this press release from the EU Rule of Law Mission and ask yourself whether there is more to this story than meets the eye:


Press release – the Son of the Kurdish President:

Indictment confirmed against six suspects in the Spain theft case
10 July 2012

A local judge at the Prishtine/Pristina District Court confirmed the indictment filed by a EULEX prosecutor against six suspects, one of them a high-ranking ROSU officer, in connection with a theft that took place in Barcelona, Spain in July 2009.
The indictment was confirmed against Haki Januzi, Artan Xhaferi-Jelliqi, and Shemsedin Benarba who are charged with the criminal offence of aggravated theft. Artan Xhaferi-Jelliqi is also charged with unauthorized control, possession or use of weapons.
The judge confirmed the indictment against Afrim Ymeri, Abedin Beka and Saim Januzi who are charged with the criminal offence of providing assistance to perpetrators after the commission of a criminal offence.
The confirmation judge found that the evidence supports the grounded suspicion that the suspects committed the criminal offences they are charged with.
The indictment was severed against Sami Makolli because the suspect was not present in court during the confirmation hearing.
The judge did not hold a confirmation hearing for Erhan Berniqani, Dusan Kadijevic, Nebojsa Bojvic and Afrim Ymeri charged with the criminal offence of receiving stolen goods, since these charges are handled through a summary procedure.
The defendants are accused in connection with the theft of approximately 1.3 million Euros worth of jewellery in a hotel room in Barcelona, Spain.
The case is being prosecuted in a mixed team led by a EULEX District Prosecutor.

The press release appears to a pretty standard announcement of an indictment for an unremarkable crime, until you reach the second to last line.

When I read this, the immediate question which jumped out of the page was what kind of person would own at least 1.3 million euro worth of jewellery, travel with it and then leave it in a hotel room? The answer is someone who is grossly rich with perhaps a reckless approach to security.

After obtaining further court documents, BIRN was able to confirm that the victim of this crime was the son of the Kurdish President, who had been gambling in Barcelona’s casino during an extended stay. This raised further questions about how he had accumulated such wealth and what he was doing in Barcelona.

An alternative investigation which reporters could have carried out would have delved into how this Balkan gang was operating in Barcelona, whether this was a one-off, how they operated and whether they were connected with other Balkan organised criminals, such as the notorious Pink Panthers.

Thus, from a simple press release, two excellent and intriguing investigations emerge demanding further attention.

Tackling the Press Release

The Press Release: A vector for much of modern journalism and a symbol of what is wrong with our trade. PR merchants and taxpayer-funded spin doctors spew out ‘media releases’ in the hope that journalists will be too lazy to do more than copy, paste, rewrite.
Press releases, however, can present a useful starting point for an investigation, revealing a great deal in a turn of phrase, its timing and its omissions. Here are a list of questions which are worth asking yourself when picking through one.
Why is the press release being issued? Is it pre-empting bad news (was it issued unusually late in the day for example)? Who issued it (which department or which press officer? Who is quoted (the more powerful the person quoted the more likely it is important even if at first sight it does not appear to be)?
Is the language and vocabulary clear? Is there deliberate use of vague or unusual terms?
What has been omitted from the release? Key details? Names? An important event?
Are there details which are unexpected or unusual which require further investigation?

Spotting the unusual, understanding the usual

As demonstrated by the previous example, a key part of investigative journalism is spotting the untold and unusual. You should be alert to this as you walk to work, attend a press conference or pour over thousands of pages of government paperwork.

Often, what is “unusual” will be very obvious – although it is by no means certain that journalists will follow it up. In a case involving the EULEX press release, we can all agree that it is unusual for someone to travel with 1.3million euro of jewellery. But, sometimes what is common and uncommon isn’t immediately apparent, particularly if you do not know the field well.

To spot the unusual, you must first understand what is usual in a certain context. 

For example, it is usual for companies operating in the construction sector in Albania to be registered locally. It would be unusual if one of these is registered in the British Virgin Islands and if you discover this you should investigate. 

The best way to increase your chances of spotting a story is to be well acquainted with the field you’re looking at and/or speak to someone who is. 

It is impossible to build an exhaustive list of the “unusual” as each case is context-specific and you must let your journalist nous and common sense guide you to some extent but here are three red flags to look out for:

RED FLAG The presence of offshore companies: This is a tell-tale sign that something untoward is going on

RED FLAG Overly cautious, sketchy or aggressive statements from public bodies: In adversity and under pressure organisations, and humans, can react strangely. Remain vigilant for unusual behaviour.

RED FLAG Secretive and opaque processes:  Where there is secrecy, a lack of transparency and money, there tends to be corruption – you only have to look at FIFA, or the scandal surrounding British MPs’ claims for expenses (to get taxpayers to pay for a duck house on a private lake or have a moat dredged!). When there are secrecy and haste in a particular process (the awarding of a mining licence or tender for example), this should also ring alarm bells.

 Case study – Arms dealers arrested in Montenegro

In March 2015, a Serbian-Israeli consortium snapped up Montenegro’s state-owned arms-dealing firm.
While Israeli firms and dealers are major players in the secretive weapons market, it was still noteworthy that it was an Israeli firm, ATL Atlantic, which had partnered with Serbian company CPR Impex to buy Montenegro Defence Industries in March 2015.
This provided BIRN with the impetus to find out more about the deal and who was behind it.
Public records in Bulgaria, where ATL Atlantic had opened a subsidiary, revealed that it had recently been owned by a man named Serge Muller. 
Following a brief internet search, two important leads emerged. The first was that Muller was a notorious arm and diamond dealer who had previously been linked to ATL Atlantic in Belgium. 
The second was that Muller had been arrested under a Belgium arrest warrant in Montenegro at roughly the same time as the deal between the Montenegrin government and ATL Atlantic had been signed. This raised the important questions - was Muller still involved in ATL Atlantic and was he present at the signing ceremony?
Through various interviews, we were able to piece together Muller’s movements and demonstrate that he was arrested immediately after leaving the signing ceremony and that he continued to play a role, albeit an unofficial one, in ATL Atlantic. Following the publication of our story, the case was referred to the prosecutor.
This example demonstrates how a journalistic hunch – a sense that there is something unusual – can lead to important revelations.

Spotting the unreported in the reported

Many great investigations are hidden in plain sight. By that, I mean that all of the information is publicly available, often accessibly online and sometimes already in the media, but the full picture has not yet been pieced together.

This is often because journalists have failed to ask essential questions, hastily dispatching stories for a daily deadline.

You will, therefore, find many great leads for investigations in the unfinished work of your colleagues. 

Case study  – Veselinovic works for PDK-linked firm on Bechtel
It had been widely reported that Zvonko Veselinovic, the Serbian underground figure who had led the 2011 uprising against the Kosovo government, had worked on Kosovo’s 1billion euro highway.
Some sources had even reported that he had been involved in a quarry based in northern Kosovo, near the border with Serbia.
At the same time, other news outlets had reported how senior figures linked to the ruling Kosovo Albanian party of Hashim Thaci had opened a quarry in northern Kosovo.
Placing the two stories side by side, it is remarkable that no one had spotted the potential links between the two.
Using court testimony from Belgrade, publicly available documents and interviews with key figures, BIRN was able to conclude that Zvonko Veselinovic had provided logistical services to the PDK-linked firm, Arena Invest, and that this firm had secured a major contract from the firm constructing the highway.

Spotting a pattern

Another useful technique to hone in on an investigation or to expand the scope of the story is to look for a pattern of events.

To do this you need to dissect the information you have, breaking it down into its most basic form. For a suspect tender, this would mean compiling a spreadsheet of the companies involved, owners, officials, addresses, telephone numbers and lawyers, for example.

You can then look at other tenders involving one or more of the elements above.

Step 2: Initial fact-finding

So you’ve spotted something unusual which has piqued your interest and after some basic research, you are beginning to form a picture in your head of what the story could be. This picture, or set of scenarios, is the beginnings of your investigative hypothesis.

Over the next few days or weeks, you will need to test this hypothesis and ask some searching questions about what the story is, how much it will cost to carry out, how much time is needed, what are the logistics of contacts and travel, and whether it is in the public interest.

At the end of this process, you will form an investigative hypothesis – combined with your minimum and maximum story which we will explain shortly – which will guide you throughout the investigation, providing you with clear demarcations of what your story is about and where it is heading.

Initial research to create a list of facts, and assumptions

Once you have a potential story, you need to refine it through some basic research. This will help you understand if your story is worth pursuing and to frame it. The time needed for this initial scoping will depend on how much you already know about the subject and its complexity.

The research should include:

  • A thorough press review of all available material. You may need to use online media archives for this and, for course, Google. Many media organisations and NGOs now publish the source material for their work – check these carefully for leads.
  • Use social media – Facebook, Twitter, Instagram and Google Plus – to build up an initial profile of your targets (see Chapter IV for more information) 
  • Check other officials, online sources of material – government tenders, company registers, court documents, UN reports – to see if there is a pattern or other potential avenues of research.
  • If there are books or academic and NGO reports about your chosen area of research then it is a good time to start reading. This will give you the in-depth knowledge to understand the world you are entering. You need to grasp the vocabulary and the processes involved in your subject matter if you are to fully investigate.
  • Meet sources who could give you an overview of the subject or an indication of whether you are on the right track. Do not spend too long on this in the initial phase.
  • Research the context of the story asking to determine why this issue important? Are there victims and who could they be?
  • Attempt to work out the costs and time involved.

Create a list of bullet points which summarises your initial findings. These will allow you to later form your hypothesis.

Step 3: The feasibility question

Before you begin your investigation and frame your hypothesis, it is essential to ask yourself what documents are available and where you will find sources.

There is little point aiming for a story which you know will not be provable. This is not a time to be despondent about the lack of information available, but to carefully consider what you can reasonably expect to achieve.

Draw out a rough timeline of events, or a flowchart of how the system works, listing documents and people at each stage who are involved or will have knowledge of this. In this chain of events, look for the weakest links i.e. people who are most likely to talk. 

Here is a helpful list of questions to use as a checklist.

  • What documents are available and related to your subject matter? Can they be obtained through a public database? Freedom of Information request? Which institutions hold copies? Are there relevant court cases? Commercial disputes? What are the costs involved?
  • What sources will have information relevant to the investigation? How many people are involved? Where do they work? How willing will they be to speak and what motivations may they have to blow the whistle or remain silent? Can you easily identify former employees? Who loses out? Who are the victims?
  • How many and which countries are involved in the investigation? How could the international nature of the investigation help by providing extra sources of information or hinder by making the research more complex? Could you obtain information from more transparent jurisdictions?
  • Are offshore entities involved? What are the major complications you are likely to face? Are there potential ways of bypassing tax haven secrecy?

Step 4: The pitch

Set a hypothesis

Your hypothesis is what you are seeking to prove (or disprove) and therefore will guide you through your investigation, who you will contact, which documents you will seek.

It is an extremely useful device to focus your thoughts but also to explain to your editor and collaborators what you are doing. Get your hypothesis right, and pitching your idea should be easy.

Many fields use a hypothesis to frame their work, from scientists to police officers, which is then tested throughout the research or investigation. 

You will need to think carefully about what kind of information is available and how you will obtain it as there is little point setting a hypothesis which cannot be proved through journalist endeavour. 

Your theory should provide answers  to the classic what, when, who, why and how questions based on your initial research and your hunch. 

Remember that it is easy to prove that events took place or decisions were made but very difficult to demonstrate the intentions behind them. For example, it is easy to prove that a minister awarded a contract to a friend, but it is far more difficult to prove that the minister awarded the contract to a friend because of their friendship. 

Your hypothesis, in this case, may be that the contract was awarded because of the friendship, but do think carefully about how you can prove it.

Case Study: The Dahlan Investigation

When we discovered that famous Palestinian politician Mohammed Dahlan had been quietly and inexplicably handed a Serbian passport in 2013, we knew we had a good story idea which required further investigation. 
We set about checking how this had happened and delving into how feasible the story was. Having explored the data we discovered these facts:
Dahlan, his wife and four children had been given citizenship in 2013 and 2014
Six other Palestinians had also been given citizenship over the same period, often at the same time as Dahlan family members. An initial search showed that these figures were supporters of Dahlan and had worked with him in the past
Citizenships tended to be awarded to prominent sporting and artistic figures, not politicians
No other Palestinians had been awarded citizenship in recent years
The decision had been taken by the Prime Minister and/or members of the cabinet
During this period Dahlan had been working with Abu Dhabi investors in Serbia, bringing the promise of billions of investment and had been honoured by the Serbian president for his bridge-building.
Mohammed Dahlan was under investigation at the time for corruption in Palestine.

From this we form this hypothesis:
Serbia’s government lavished passports on a group of controversial Palestinians in recompense for the work of Mohammed Dahlan in bringing gulf money to Serbia.

Define your maximum and minimum story

Investigations are complex and unpredictable, making it impossible to accurately predict the outcome, particularly as you rely on sources and public bodies which are beyond your control.

Even the best reporters using all the tools available to them need a bit of luck, particularly as cost and time constraints must be taken into account.   

As a result, it is important to create a safety net by setting a minimum story which you can publish even if sources fail and documents do not materialise. 

If your minimum story is still worthy of publication, then your project will not fail even if luck is not on your side.

Equally important is setting your maximum story. This provides you with your ultimate ambition if you collect all the documents you planned and all of the sources speak out.

With a hypothesis, minimum and maximum story pencilled out, you are ready to pitch your story to your editor and convince him or her that you really need to be taken off the daily news agenda to.

Step 5: Create a research plan

At this stage you know you have a story and you have successfully pitched it to your boss. It is time to create a research plan. 

You should start by breaking down your story into stages in the process or events in a timeline.

From there, note what documents are created and used at each stage, who is involved in the process and who else would have access to the information. 

Most stories can be broken down in this fashion, but if not then try to imagine all the interactions your person or company has with individuals, other companies and institutions (from the United Nations to small municipal councils). Make a note of these and the potential to obtain information.

You should then analyse each stage to determine to weakest links in the chain. Which institutions are likely to provide you with documents? Which individuals are likely to talk to you?

You should also consider the difficulty and time in obtaining the information or contacting people, (for example, Freedom of Information requests).

Write out your research plan (using a spreadsheet works well) and share it with your colleagues.

This article is originally published in BIRN Albania’s ‘Getting Started in Data Journalism’.

How To Make a Database on Politicians’ Assets

The database on the property of politicians compiled by the Center for Investigative Reporting, CIN, from Sarajevo is the most complete database on the revenues, movable and immovable property of more than 120 officials and politicians from both the ruling and opposition parties in Bosnia and Herzegovina, BiH.

Finding out about politicians’ assets:

CIN published the first such database ten years ago when it processed data for 32 politicians. By 2014, that number increased to 200. Some of the politicians from this database are no longer active or have died. Their profiles can no longer be found in the database.

CIN reporters searched court records, payrolls, stock reports, company registers, land records and other documentation. The information obtained was compared to that reported by the elected officials in the property files. 

These forms are filled in by politicians at the beginning and end of their term of office and are submitted to the Central Election Commission, CIK, the institution that conducts elections in BiH.

While citizens have the legal right to inspect the property status of elected officials, this database has become valuable material for future journalists, not only CIN journalists but also for colleagues in other newsrooms.

In recent years, CIN has published a great deal of research on the illicit acquisition of property by officials and politicians. We discovered a big disparity between the money some of them invested in real estate and the legal income they were earning, thus establishing that the assets they had gained in some cases involved corruption and criminal activities. 

Following the publication of the stories, investigative institutions have sometimes responded. Some officials have since been tried and indicted, their property has been blocked by the courts or they have been dismissed due to public pressure.

One such story concerned Kemal Causevic, a former director of the BiH Indirect Taxation Authority, who bought real estate worth more than 500.000 euros in less than seven years. His official income was not nearly enough to cover the investments in property. Three days after publication of this story, in April 2013 the Court of Bosnia and Herzegovina blocked the property. It is currently taking legal proceedings against him for corruption.

Suspicious property asset of Bosnian Croat chief:

In a second example, the story received two awards for investigative journalism. Reporters revealed in August 2015 that Dragan Covic, president of the Croatian Democratic Union, HDZ, of BiH and a former member of the BiH Presidency, had acquired property in a suspicious manner. 

CIN research showed Covic and his wife obtained their most valuable piece of real estate before 2003. His wife had bought the property worth thousands of euros on behalf of her relatives and later received it back from them as a gift. Investigations and lawsuits were opened on suspicion that the HDZ president had used his public office for personal gain. 

We found that quite a few officials had lied about their income and the real estate they owned when filling out their property records. While we informed the public, punishments often did not follow, as no sanction is foreseen for incomplete or false reporting of property in the property files under the laws in Bosnia and Herzegovina. Also, no institution in Bosnia checks the information written in the property records, which is why this database has special value.

Comparing the property records filed by politicians and the actual situation, journalists found that a large number of them were not accurate. For example, Bakir Izetbegovic, president of the Party of Democratic Action, SDA, and former member of the BiH Presidency, had omitted to report land held in Sarajevo’s elite settlement Poljine, which was worth around 13.000 euros. Izetbegovic did not declare this land in the property file although he was obliged to do so.

In 2012, in the wake of a BiH Court ruling on the protection of personal data, the CIK removed the BiH property files of politicians from their website. The public could only see this information on CIN’s database. Because of this, this database has become one of the most-visited projects on the CIN portal. Content from the database has been published frequently in other media.

The property card analysis is important because it reveals many unusual things. For example, it shows that Milorad Dodik, president of the Alliance of Independent Social Democrats and a current member of the BiH Presidency, in his 2010 property file stated that he had given funds to his son, Igor. This would not be unusual had he not also stated in the same form that his son now had assets worth 1.3 million euros.

This CIN project has also had an impact beyond the BiH border. Journalists of the Crime and Corruption Investigation Network, KRIK, in Serbia have made a similar property base of Serbian politicians, which has also proved a great move.

The property project has, therefore, fully justified its existence and the efforts of the journalists involved, and, due to the interest of the public, the need to constantly update it and add new profiles has been shown.

How did we do it?

To build such a database, documented and verified information is needed. CIN reporters mostly obtained this information using freedom of access to information, FOAI, requests sent to institutions. On this basis, journalists collected almost 3,000 documents that they published in the database. All the published information was confirmed in interviews with those they wrote about – over 100 interviews in all.

Although the laws in BiH allow citizens free access to information, in practice this freedom is limited. It often happens that institutions unjustifiably reject journalists’ requests and thus bar access to information of public importance. In those cases, a fight follows with the bureaucrats within the institutions as well as a search for other sources that may have the desired information.

Given the complicated administrative apparatus in BiH, journalists also face the problem of obtaining documents in one part of the country, while in the other, they faced complete resistance. The land registry office is one example. The Federation of Bosnia and Herzegovina, FBiH, entity offices seamlessly provided documents proving ownership of real estate on the basis of FOIA. But in Republika Srpska, RS, they regularly denied such requests for information, although they were not allowed to do so by law.

One way to get information is to sue institutions that refuse to respond to requests, which is what CIN did. However, these processes are slow and the uncertainty of the court process cannot guarantee receipt of the requested information on time. Therefore, you need to look for other ways to gather information. Given that, under the laws, land registers are public and every citizen has the right to obtain these documents, journalists have sometimes physically gone to offices that get this information, without submitting a request, and received documents on the spot.

In the meantime, the courts also decided in CIN’s favour concerning one lawsuit for failure to provide information under FOIA. This verdict came late, however, after the release of the database. But it allowed us to request this information based on FOIA, so we can update the information or explore new profiles in the database.

Each country should have an online register of businesses, which shows if a company is being run on someone’s behalf. Using media reports and conversations with sources was our best way to find out if politicians or their family members had such businesses. Subsequently, we confirmed this information in the courts, requesting the provision of a history of ownership. We similarly checked the ownership of individuals over securities. There are official portals where you can track the value of stocks on the stock exchanges and who owns them. We then entered the stock exchange reports in the database, stating the exact time of data entry.

Although the database is called “The Property of Politicians”, we have also collected information on court proceedings against politicians. We received verdicts and indictments from courts and prosecutors’ offices. Sometimes, judicial institutions refused to provide this information, but with persistence, we eventually had it in the database.

Each profile also contains biographical information about the officials’ education and political and professional careers. We used press clipping for extra input. Most politicians have published official biographies on the Internet. Based on this information, we were able to confirm with institutions, public companies and faculties what function they performed, during what period and when they had graduated.

We constantly added new stories. By checking the data in two cases, for example, we found that high-ranking politicians lied about their diplomas or college.

Sarajevo Mayor Abdullah Skaka, for example, told CIN that he had graduated from the State Faculty of Economics at the University of Sarajevo, and this information was included in his official biography. However, this faculty told us that this was not true. At the same time, a certificate of Skaka’s diploma came from another address, the privately run Faculty of Management and Business Economics at the University of Travnik.

Gradonačelnik Sarajeva Abdulah Skaka je ranije za CIN rekao da je diplomirao na državnom Ekonomskom fakultetu…

Posted by Centar za istraživačko novinarstvo (CIN) on E premte, 5 Tetor 2018

In another example, the official biography of the RS Minister of Industry, Energy and Mining, Petar Djokic, says that in 1991 he graduated from the Faculty of Economics in Belgrade. He confirmed this to CIN. But the Faculty of Economics at the University of Belgrade told us that Djokic was never one of their students. CIN reporters revealed that Djokic, in fact, graduated from the Megatrend International Expert Consortium Limited Business School in Belgrade.

This was not the end of the process of gathering information. We also needed to find out the salaries and the extra income that officials were earning while performing their careers. Given that we already had enough information about where they worked, we approached these institutions and asked for the exact amounts of their income. 

As in the previous cases, many requests were rejected. But considering that the database is a long-standing project that is regularly updated, it is worth fighting for. This information concerns a specified period in the property records, so journalists can use it until they receive an official response from the relevant institution.

When the main actors refuse to talk:

Of course, it often happens, as you do a story, that the main actors refuse to talk. This should not be discouraging, because as mentioned before, when they notice that you have been collecting data, they often end up agreeing to an interview. Interviews are not only important for the sake of respecting professional standards. Interlocutors often give you information that you won’t find in your documents.

Some of the information collected is covered by proprietary or private information, which is why it is either unavailable, or you can obtain it, but it shouldn’t be disclosed. Care should be taken in how to process such data. For example, while the public has the right of access to the property of politicians and their families, CIN has deleted the actual addresses where they reside in the property files. The other information is available in the database.

To create such a database, it is necessary to have a plan in place at the beginning, to share the responsibilities of the editorial office and to respond to the obstacles that you will encounter in a timely manner.

And, in the end, before publishing, you need to check and compare everything with documents and transcripts to avoid the inconvenience of potentially publishing incorrect information.

How To Make a Database of Proceedings Against Officials

Is the Serbian state interested in fighting corruption? A good question, to which many citizens would probably say: No. But how to show this is true? How to find good examples of unsuccessful fights against corruption without pointing to specific cases? How to detect loopholes in the system?

The answer imposes itself; this systematic problem can only be detected by systematic analysis.

That’s how we at CINS came up with the idea of making a database of proceedings against officials – a creative and innovative way of showing how important and powerful institutions like the Anti-Corruption Agency, ACAS, are not doing their job.

What makes this CINS project special is that we haven’t just written an investigative story in which we list all the problems. We wanted to show our readers, and the public, what the true problem was, and give the problem it’s first and last name.

We wanted to give the public an opportunity to see the data and all the information by themselves and get a chance to have important documents just a click away, so they can believe what their eyes tell them.

To achieve that, this project became a collaboration of journalistic skills and innovative open source technology in the purpose of reporting about corruption.

What the database is

The database of proceedings against officials is a unique database in Serbia. It contains information that you will not find on any government or institutional website.

First published in 2016, it was upgraded in December 2018 and now contains information on all proceedings initiated by the ACAS against public officials since 2010.

By public official, we mean practically anyone appointed to work for the state, from the President to ministers, mayors, municipal heads and directors of public companies. But not only them. An official is also a public prosecutor, judge, faculty dean or director of a school.

They have all had an obligation to send data on their assets and revenues every year to the Anti-Corruption Agency since 2010, and to inform it about any potential conflicts of interest while working for the state.

If they don’t report, for example, that they have an apartment in another city, or own a company, or that their wife or husband owns cars, the agency should initiate proceedings against them.

If they don’t report that they have employed a family member, the agency should also initiate a procedure.

All this information is available in one place, the CINS database.

For several months, CINS collected data, such as:

  • The list of officials that the agency initiated proceedings against, and the reason for initiating the proceedings.
  • The list of procedures initiated.
  • Who the officials that violated the law were.
  • Which officials the agency filed misdemeanour or criminal charges against.
  • What the prosecutors and courts did with a misdemeanour or criminal charges

The results

1. Who are the officials potentially breaching the law?

Are you interested in whether Prime Minister Ana Brnabic, President Aleksandar Vucic, former Mayor of Belgrade Dragan Djilas or former and current finance ministers Mladjan Dinkic or Sinisa Mali broke the law? All you have to do is type in their name in the search IME I PREZIME. The list contains more than 1,700 officials.

What you will find is the following: In the case of Sinisa Mali, today Minister of Finance, for example, we see that the agency launched three proceedings against him over 2014 and 2015, and that in all three cases he received the smallest possible “punishment” – a written warning.

2. In which towns do these officials work?

Do you live in Nis, and are you interested in officials from Nis that have come under the radar of the agency? In the search option MESTO, type the word “Nis” and the database will give you a list of all the officials from this town against whom a procedure has been or is being conducted. In this way, citizens and journalists can quickly get information about officials from their hometown and keep track of what these officials are doing.

The City of Belgrade is in the first place according to the data of the database, which makes sense, as so many state institutions are located in the capital. In second place is Novi Sad, again with a host of provincial officials. The third and fourth place is occupied by Nis and Kragujevac. Behind them comes Vranje. This tells us that the most irresponsible officials come from these towns and cities.

3. Institutions and positions

Do you want to know which current or former minister the agency has investigated? This option is also possible through search option FUNKCIJA (position). Simply write the Serbian word “minister” and the database will list all the ministers from several governments that have been on the radar of the agency. 

Analysis of the position that the officials held when they violated the law shows us that the most irresponsible ones were directors of public enterprises and, in second place, the employees of the highest bodies of local municipalities.

This search option is complemented by the search of a type of institution where officials work – ORGAN. We have thus enabled our readers to search, for example, for who the officials employed by the Agency for Privatization that violated the law are. Or, against which judges of the Basic Courts the agency started proceedings?

4. What were the outcomes of the proceedings?

During our work on this database, we constantly wondered what our readers would want to find, what would be useful search options for journalists, and what information could be important to NGOs. So, the process of organizing search options was long. Without the help and suggestions of our developer who worked with us on developing the database, we would not have been able to do something that is simple, on the one hand, and useful on the other.

We also strengthened and improved the four basic search options to provide data on how the agency “punished” these officials through four basic measures:

  • a warning 
  • a decision on law violation
  • recommended dismissal
  • misdemeanour and criminal charges

It was important for us to have this data because it shows how seriously the agency takes the fight against corruption, and whether irresponsible officials were eventually punished for violating the law. 

The results of the data were disappointing. In most cases, the agency went for the smallest punishment – a warning measure that was merely a written statement and does not entail any special consequences. In most cases, the most severe “penalties” – criminal charges – were dismissed by the prosecutors’ offices. No investigated officials went to jail, the data show.

The right recipe for creating a database

To create a database, more things need to come together to make the magic – a dedicated team, a good organization, clear knowledge of the topic and data you are working with, a programmer and a lot of patience.

When we started to work on this database, we wanted to include young journalists in the project, to give them a chance to work on a big project and learn from the beginning.

So two of our then participants at the CINS School of Investigative Journalism joined the team of experienced investigative reporters already working on this complex project. 

These were key elements:

– For a team to function and work on schedule, good organization is crucial. First, it means a clear division of work and coordination, a balanced distribution of work, regular meetings to monitor what has been done and plan the next steps, continuous communication and solving problems as they come, because they will surely exist even though you think it is clear from the beginning who is doing what.

– Another key element is Excel, an incredibly useful but also complex program that was the beginning and end of every day for us while working on the database. It was necessary for journalists to know how to work in it and to respect the strict rules that we agreed on. This means that if we say that the years are written without a dot, in the end, one person cannot write “2014”, the second “2014.”, and the third “2014. God”. Similarly, if we agreed that the figures should be written only as a number, they should not write “50 thousand dinars.”

Working in Excel obliges you also to something called cleaning and data processing, because errors will appear somewhere. But if you initially create a system of how data is entered, it will be much easier.

– Our programmer was very helpful here, because he marked all the cells and columns in the Excel that contained minor errors and omissions, such as extra space functions or incorrectly typed words.

– In the end, before publishing anything, CINS standard demanded fact-checking. So, one person who did not have anything to do with the work on the database checked all the data entered in the Excel for several days and pointed out potential flaws.

– Time. Databases cannot be done quickly. One of the key reasons in our case is that Serbia is still far from what is called digitization, and even further away from the idea of open data. In some countries, all these data would be publicly available and organized by the state. That’s not the case with us. The Anti-Corruption Agency publishes part of its information on its website in the form of various not-searchable scans and unorganized documents located in different places and not easy to find. The other part of the information is no longer available.

A few months after CINS released the database, the agency pulled key documents from its site. You cannot see any information about proceedings against officials, except on the CIJS website today.

In the end, time is an important factor in collecting data for any database because we had to send dozens of requests for access to information of public importance to institutions throughout Serbia to collect the documentation that we need.

If a person wanted to get information about all the proceedings against public officials in Serbia and to find the outcomes of these procedures, he or she would have to send nearly 1,000 requests for access to information of public importance to various institutions. Even if you have enough journalists and time to do this, it’s still not certain at the end of the process that you will collect all the information. This is because in Serbia institutions simply do not respond in time, do not send out everything you asked for, sometimes send out poorly scanned documents that are almost useless and sometimes send the wrong documents.

Although the deadline for responding to requests is 15 days, some institutions, and in the first place the Anti-Corruption Agency, has not yet responded to the submitted requests.

Despite the obstacles, a Law on Free Access to Information of Public Importance exists in many countries around the world and is important for all journalists who want to collect data from various institutions and create a database of them.

In the end, there were many challenges over the course of several months while we were working on the database, and my role was a sort of coordinator of the entire project in addition to the journalistic work on the research. Having a person working with the whole team and knowing at all times how much has been done and how much more you need to do is very important. That is the only way you can keep up, especially when the project lasts for six months and time can easily slip away.

Why GitHub

Reporting about fight against corruption through open source tools and open data was the main idea when we decided to make the entire database code publicly available at GitHub.

The idea is that anyone who wants to do something similar can make use of our code and make something new. All you need is knowledge in R programming language or a programmer that can re-use the code for you. Therefore, you can go to the CINS profile on the GitHub platform at any time and download the database code.

Database Proceedings Against Public Officials is the first example of data made available by journalists in the form of open-source code in Serbia. The project was also presented at the Open Data Week in early April 2019 and CINS was the first media in Serbia to participate in the event.

In May 2019, the database was nominated for the international journalistic award Data Journalism Awards 2019. It seems to me that we did a good job.

How To Make a Database on Criminal Murders

As murders related to organized crime become more frequent in Serbia, while state institutions remained silent, we decided to determine how many murders in the country were connected to criminal groups. 

So, we made a database, which we have called “The Black Book”. One goal was to show what the authorities were doing to solve such cases or prevent such killings from occurring.

The Black Book is a unique database, containing information on the total number of mafia killings that have occurred in both Serbia and Montenegro from 2012 onwards – and what the authorities have done to solve these cases. 

The information in the database can be easily filtered; it is possible to search by different criteria, such as the place where the murder occurred, the time frame, and the stage of the prosecution investigation or court trial.

The reason for creating the database

The original idea was to collect data on criminal murders, analyze them and see what they show. We started working at the end of 2015, and at first covered both the current and previous year, so that we could compare the data and tell if the situation was getting better or worse. 

By the time we decided to build a database, a year had passed, so we now had information for three years. To make the sample even more representative, we added two more years. In this way, the database became more relevant as it also shows if the rate of those murders is increasing or decreasing. 

A few phenomena attracted our attention. One is that mafia killings have become more frequent, yet state institutions have remained silent. A key aim was to show what the authorities were doing to solve these cases or prevent these kinds of killings from occurring.

Mafia war causes a spike in killings

In the beginning, we didn’t think we would have enough material for the database, so our idea was to publish a feature story with infographics that would explain the data we had collected. We soon realized that the situation was more serious than we had thought. 

The number of murders was constantly increasing because of an event that happened in 2015 when a mafia war erupted between the two major criminal clans originating from neighbouring Montenegro, known as the ‘skaljarski’ and the ‘kavacki’.

The original cause of this conflict, which took on incredibly large dimensions, was 200 kilograms of cocaine that went missing in Spain. The clans started fighting in Montenegro and Serbia, but also in some Western European countries where they also operate. Other important criminal groups from Serbia and Montenegro, we found out, then took sides in this war. 

So, we knew the Black Book couldn’t be complete without covering Montenegro as well. Organized crime groups from Serbia and Montenegro have been linked for decades. They often work together, or, and, have the same smuggling channels, but sometimes they fight over territory and illegal businesses. 

Some of the murders that happened as a result of this war were committed in unusual places. One man was killed in a prison yard while serving his jail sentence. Another was shot dead at a prison gate while waiting to visit his brother who was an inmate. Murders occurred in the middle of the day, in public places, on streets, squares, even in restaurants and cafés. We noted every murder that occurred.

The final decision to create the Black Book was made at the end of 2016. By then, the number of mafia hits in Serbia had escalated to 52. 

We realized that we were not the only ones to notice both the importance of this topic and the frequency of murders. Colleagues from Radio Free Europe also expressed interest, so we joined forces on the project and split the job. RFE focused on Montenegro while we continued to work on Serbia. We brainstormed about the methodology, the visual look and the functionality of the database, but also on the information it should contain.

Choosing the right methodology

We started to collect data – news articles, police and prosecution press releases, and publicly available official documents. Information related to murders up to mid-2015 was collected from the media. From that point onwards, we started doing everything by ourselves.

To make the database relevant, we needed to agree on a methodology. We noted victims who had any criminal background. We also decided that murders had to be committed in mafia-style – using firearms or car bombs, even if the victim himself had no criminal background. It was also important what the authorities had to say about possible motives.

Choosing the methodology and compiling the list of murders, it turned out, was the easy part. Another decision was figuring out what data to use, and what was important for the readers, and for journalists and other researchers who would use the database as a source of information. 

We realized it was necessary to include various things: the place of the murder – as this showed us whether the assassinations were more frequent in certain cities; a description of the incident; the age of the victim (there was a prevalence of victims under the age of 40); the way in which the murder was committed and the stage and progress of the investigation. 

Map of killings in Belgrade. Photo: Screenshot, crnaknjiga.rs

We decided also to compile several categories to show: what the prosecution and police had done; any unknown perpetrators; any possible suspects; on-going trials; case closures (final verdicts).

We needed to fact-check all the information collected from the media. In most cases, the media had already published the full name of the victim. If they didn’t, we had to figure out ways to identify the victim. One was to check the obituaries; almost every time we got full name and age of the murder victim.

We had to be sure also when and where the murder took place, check the person who was killed and get official information on whether the victim was convicted and, if so, for what crime. 

In some cases, this was easy. Certain prosecution offices in Serbia answered our questions quickly. The situation was more complex with the prosecutors’ offices in Belgrade, where it is almost impossible to get such information quickly. Since most of the murders took place in the capital, we had a lot of work to do. 

Besides investigating the victims, we had to monitor what was going on with the police investigations. By mid-2017, just before it was published, the database contained 80 killings. For each of them we asked the competent prosecutor’s office about the investigation and whether there were suspects or the trial had begun. 

If the case was already before the court, using the Law on Free Access to Information of Public Importance, we asked for copies of the indictments. These gave us more details, such as, among other things, the way the murder was committed. 

But because prosecutors and courts in some cases delayed delivery of the requested documents, this often took a while. 

In the meantime, the number of murders increased, so we were working on new cases. In mid-2017, we were ready to publish the database. By then, we had recorded 83 murders and had collected all the necessary information about them.

Most criminal killings remain unsolved

The data were alarming; murders took place on average every three weeks and were often carried out in public places. The database showed also that the authorities were not solving these cases; there were only four final verdicts.

The Black Book has revealed that the lives of citizens of Serbia and Montenegro are endangered by these ongoing criminal conflicts. In two cases, innocent victims lost their lives. Other citizens were injured just because they found themselves at the wrong time and place.

After the database was released, we were attacked and our data was questioned. Senior officials tried to discredit our work. The Serbian Minister of Interior, Nebojsa Stefanovic, and the President, Aleksandar Vucic, both presented their own statistics, which, of course, showed the authorities in a better light – as successfully solving these cases. They also claimed the number of murders was decreasing.

But they forgot to mention that their statistics included all murders, not just mafia-related killings but also “crimes of passion”, caused by domestic violence, quarrels, and brawls. 

They have tried to convince the citizens that they are safe. At the end of May 2018, the Minister of Police in Serbia insisted Belgrade was a safe city, even though 10 murders had occurred in the city that year.

It is a dispiriting fact that, since October 2016, when the state of Serbia declared war on the mafia, 37 more mafia-style murders have occurred.

Two years after the publication of the Black Book, the number of murders has risen to 147, of which 96 were carried out in Serbia. Only 9.5 per cent of these cases have been solved. The suspects remain unknown in almost 64 per cent of them.

A useful barometer of civic safety

Besides citizens, this unique database is also important for other journalists and researchers monitoring the level of citizens’ security in Serbia. 

Its unique value has been recognized beyond the borders of this country. An MEP and Rapporteur for Serbia Ivo Vajgl used its data last year during a debate in the European Parliament on the latest draft report on Serbia, in order to raise awareness about this issue. 

We, meanwhile, will continue to monitor the frequency of such criminal killings – and the willingness and ability of the state to solve them.

How To Make Database of Governments Officials Limos

Bosnian institutions and public companies spend around 10 million euros on cars a year! While most of that spending is for cars needed for public service, great amount and most expensive are government officials cars. 

In 2018 alone over 10.6 million KM – 5 million euros – was spent on purchasing 329 official limousines whose price averaged 32,000 KM, about 16,000 euros. In total, tenders to procure 1,666 official vehicles, worth about 46 million euros, were recorded in 2018 by Balkan Investigative Reporting Network in Bosnia and Herzegovina.

BIRN BiH database proved for the first time a few talking points of Bosnian public: politicians enjoy overpriced luxury limos like no others in Europe; most of the tenders have only one bidder indicating corruption; and also there is serious lack of control of budget spending on cars.

While this was speculated in media, experts discussions and in public, evidence was partial for complete analysis and complete conclusions. BIRN Bosnia and Herzegovina itself published dozens of articles of specific cases realizing in that process two important things: there are far too many examples of overspending and stories soon begin to repeat and look the same for audience lowering their impact. By late 2017 we were already collecting al tenders related to cars from public procurement website, centralized government portal where institutions and public companies are obliged to publish their tenders. 

BIRN published an analysis in December showing that around five million euros was spent on vehicles in 2017.

After BIRN’s reports about violations of public procurement practices, several institutions amended their tender specifications.

We then decided it would be more effective to make database with every tender related to cars.

It took us around six months to work with IT company to develop database structure and to manually input hundreds of tenders for car purchases and to input data on more than 3,000 cars into our car registry – where we publish data on existing cars owned by institutions.

In mid 2018 we published database and data for first half of same year with press conference. When we wrapped a database for whole year, final figure of more than 93 million Bosnian marks or more than 46 million euros of total tenders for car purchases in 2018 was a surprise even we saw that figure will be huge while we were working on the database.

It earned us a republication in almost every significant media in the country. We had enough data to be able to say there is no competition in tenders to buy cars for officials as vast majority of tenders had only one bidder and we could identify who bought most expensive cars and how they did it but also we knew preferences of models and brands for politicians.

We had data for tenders for more than 1,600 cars that year and data on how almost 1,000 cars were bought as not all tenders were finished by the end of 2018.

The database also contains a register for vehicles already owned by institutions and public companies, which shows that the average cost of a vehicle is around 25,000 euros.

Other important part of database was car registry as we now have more than 3.500 cars from numerous government institutions. Its unique database in Bosnia as no official data is available on which institutions owns how many cars and which are they and how much is their worth.

How we collect and input data

We are monitoring every out of tens of thousands tenders in Bosnia and Herzegovina and we do it every day, downloading hundreds of documents. Our source is official public procurement portal and we only use data available on this portal. Each tender contains two main parts, one document containing basic data on who is buying what and document containing detailed information on car details. While first document has template and could be used for automated input, second one is different almost every time making it hard to scrape data from portal directly to the database. In some cases, institutions use scanned document making it harder to copy data. In these cases we have to retype data containing details on car dimensions, type of engine required or what kind of display it needs to have.

As we use only official data which is in most part publicly available, we only fact check our input after we input each tender but also at the end of the year before we publish final report.

As we collect data for car purchase we found a way to make first officials car registry in Bosnia and Herzegovina. It would take years to use Freedom of information request to get data for all cars owned by every institution in the country as there are tens of thousand of cars owned by the different level of government. What we found is that all of them are obliged to use tenders to buy insurance or to repair cars. We are using tender documents as they contain lists of cars.

Tenders for insurance are more useful than ones for maintenance as they contain more data, in most of cases they have to contain data on year of purchase and price of the car.

Sometimes they also list license plates which we used when we tracked who is using budget car for traveling to election rally. 

We use custom made back end of database website for input. There are few main input entities – institutions including couple of sub-entities such as level or type of government (containing basic information on them and allowing different kind of search results); companies (basic data on companies that are bidding for tenders); car registry part where we insert each car as different entity in SQL database; expanses (each expanse tender in instead separately and linked by database to different modules which allows complex search results); cities; and most important part, tenders, which has input module containing several parts with different ways of input thru text boxes and drop down menu’s.

Database structure

Public procurement portal is far from user friendly for general audience and it does not have an option to list only tenders related to cars. We had to replicate parts of portal’s database.

Database on cars is custom made SQL database separated in three main parts – 1) data on purchasing cars; 2) car registry containing data on cars owned by institutions; 3) part where we register expenses on cars such as gas and maintenance costs.

Database is live and interactive as visitors without IT knowledge can surf thru the database and search by numbers of criteria. Its structure is designed to be user friendly and to show data as simple as possible but also to connect the data by different criteria. 

1) Data on car purchases contain several most important information on every car purchase for government institution or public company from 2018 to the date. We are manually inserting data on institutions, type of tender, its current status, number of cars purchased and their information such as which type of cars; what gas they are using and even if buyer instated on heated seats or other luxury item requested.

What is significant for BIRN Bosnia and Herzegovina database is that its showing data as they are happening so we can publish if there is something wrong with the tender. It resulted in canceling couple of tenders or changing requirements as they indicated which brand is request, criteria not allowed by Bosnian law on tenders. This live database is allowing visitors, journalists and researchers to track tenders as they are happening. 

We are also red flagging tenders, around 100 of them in 2018 by criteria made by years of state audits related to cars or Law on public procurement. 

2) Car registry contains thousands of cars and their details. It is possible to how many cars Presidency has and what is their worth. As database front end is interconnected and interactive it is possible to list all Mercedeses for example or who owns most expensive car.

Visitors can check which car their local major drives or to see how often he changes his car.

3) Third segment of database are expanses. As this is by far the biggest database set available on public procurement portal we only insert data on institutions which are already inserted in our database. We plan to automate this part by scraping data in the future.

Visualization

We tried to make front end as simple as possible and to highlight most important parts of database with modules showing latest tenders or most expensive ones, using graphs to show expanses or which company won most tenders. As much effort we put into building database structure that would allow to have complex search results we also put into visualizing hard work we put into maintaining this database.

Each tender has document linked to it making it available to others to use, a feature which made this kind of documents transparent to the general audience.

Challenges

As BIRN’s database improved transparency on car tenders procedures it faced major backlog by procurement agency. Different to other countries in the region, in Bosnia and Herzegovina, tender documentation, a document or more of them containing most important information, is not available to the public as the rest of procurement portal, but only to registered users, companies interested in tenders. 

When we started to monitor public procurement it was possible to register as user as an NGO but after we published our most important stories, our account was disabled in broader effort to ban NGO’s to bid for tenders. Even if BIRN’s account was only used to monitor tenders and even if we never big for any tender or ever wanted to, our account was also suspended with the rest. 

It made impossible to continue with the database update but also disabled notification about changes to tenders we were already monitoring so we had to manually go back tender to tender to see if they were finished and who won the tender. It took us a few week to find a solution and register as private company in order to continue database update.

News stories and impact

BIRN’s car database was quoted in every major news outlet in the country. What it enabled us is to be quoted even if we don’t do story ourselves. Journalists find different story in database by themselves and publish it. 

Other times it is enough to tweet about interesting fact or a tender from the database for other news outlets to write the story. 

As database is searchable by keywords, we once used this search to see how many tenders requested official limo to have ISOFIX, part of the seat which enables mounting child seat. It is feature that should not be listed as requirements when buying official car but it turns out six of tenders had that requirement in 2018. After we tweeted that on database separate twitter account, media made a story out of it.

Database is regularly quoted in most of news stories related to government cars.

It also made an impact on number of tenders as our live monitoring made tender to be changed or canceled after we revealed information. 

Story on public spending on cars in 2018 was one of the story in series nominated for EU Journalism Awards which won BIRN’s journalist Second Prize for best investigative stories for 2018. 

“The significance of (…) investigative series is that it documents the misuse of public procurement through publicly available information,” the jury noted.

How To Search Albanian Websites and Databases Like a Pro

Here is the list of essential websites in Albania and internationally which will help you carry out your research, including practical examples of cases when they have helped dig up important information.

Qkr.gov.al: The official web site of Albania’s National Registration Centre. The website contains a database of all the businesses registered in Albania, from small businesses to large corporations, including information on shareholders, company directors, board members and some financial records.

The database is not trawled by Google and it can be quite temperamental to use.

Although data collected by this database could serve as the basis of a news report, most of the time the data are more helpful in identifying real-world sources for journalists.

The reporters Lindita Cela and Aleksandra Bogdani used this database to identify the key actors that should be interviewed for their exposé of corrupt practices that led to the award of a hydropower plant concession in a protected area in the Lengarica River, southern Albania.

Through the database, the two journalists created a wide list of people who had knowledge of the affair, including former shareholders and directors of the company involved in the building of the hydropower plant.

Top tip: QKR’s search function by company names is temperamental. To get round this search for the company’s unique ID in the international company database opencorporates.com and copy it back into 

QKR.Qkl.gov.al: This is another government-operated website, which contains a database of the licences issued by different government institutions. This database is useful to identify people who operate in different professions or licences awarded to specific companies. 

The website works best with Internet Explorer compared to other browsers.  

Financa.gov.al: This is the official website of Albania’s Ministry of Finance, where the government publishes on a daily basis all the transaction of the treasury, including names of individuals and companies. 

Journalists have used the data, which are published under an .xls format, to identify beneficiaries of state funds and sift through possible conflicts of interests between them and state officials. However, these data sometimes contains mistakes, and the data should be double-checked. 

The data published daily by the Ministry of Finance can be easily searched by using the database spending.data.al, published by the Albanian Institute of Science, AIS, which aggregates all the Excel files. 

The Ministry of Finance publishes also other statistical data on budget expenditures, revenues, debt, etc.

Dogana.gov.al: This is the official website of Albania’s customs. The website also contains a statistical section where every month the volume of the country’s exports and imports are noted, as well as the revenues collected by the custom’s agency.

Qbz.gov.al: This is the website of the Centre of Official Publication (Official Gazette), which publishes all Albania’s legislation, including laws, Council of Minister’s decisions, bylaws from line ministries and other independent public agencies.  Search on this website does not work. Ligjet.org is an alternative mirror website with enhanced search capability. 

Journalists visit qbz.gov.al to read the official gazette, because in many cases the government does not publish its most controversial decisions on its website. 

Instat.gov.al: This is the official website of Albania’s National Institute of Statistics, which publishes periodically an array of official data and statistical reporters. These data are often useful to journalists as background information for their stories. 

Instatgis.gov.al: This database contains visualised statistical data extracted mainly from the national census and divided by administrative entities.

BankofAlbania.org: is the official website of the Bank of Albania. The central bank produces a number of statistical reports centred on the financial sector, which are available in the periodicals section of the website.

Amf.gov.al: This is the official website of the Financial Oversight Authority, an independent institution that publishes a number of reports on the financial markets. 

Ere.gov.al: This is the website of Albania’s Energy Regulatory Agency, ERE, an independent agency that regulates the country’s energy market. ERE publishes data on the energy sector, including detailed yearly reports. 

Parlament.al: This is the website of Albania’s parliament, which publishes the draft laws as well as the arguments sent by the draftees and minutes of meetings of parliamentary committees. The website provides quick access to useful information on the political parties’ stances on different issues. 

Cec.gov.al: This is the official website of Albania’s Central Electoral Commission. The website contains election results on local and national polling contests held in the past two decades as well as information on candidates.

App.gov.al: This is the official website of Albania’s Public Procurement Agency. The website contains a database of current public procurement listings, as well as notices of the winners of various tenders held within a period of a few months. Historical data is continuously archived. 

Kpp.gov.al: This is the website of the Public Procurement Commission, PPC, an appeals body for tenders which have been run by public institutions. The website contains a database of complaints and decision takes by the PPC for procurement disputes.  

Gjykata.gov.al: This is a portal hosted by the Ministry of Justice, which provides access to verdicts of the district and appellate courts in Albania and links to their respective websites.

This article is originally published in BIRN Albania’s manual ‘Getting Started in Data Journalism’.

How To Search Online, Google Basics

The internet has revolutionised investigative journalism. From the comfort of my office, we are able to consult official records of companies, land/cadastral authorities and governments from Panama to Mozambique. We can map the movements of an Egyptian arms dealer using social media.

While the old skills of the trade – such as finding sources and scouting locations – remain as important as ever, you can carry out a huge amount of background research without leaving your office. This means it is possible to test your story and make important breakthroughs without having to expend large amounts of time and money.

Given the importance of online research, it surprising that many journalists fail to understand how the internet functions, and therefore fail to exploit its amazing potential.

If you do not understand how to use Google properly, know what the dark web is, and can trawl social media, you will be hugely handicapping your potential.

The internet of things

Surface Web: Google.com is the most visited website in the world and rightly the gateway for many journalists when they are researching an investigation.

Google, and other search engines, work by indexing (taking copies of) of billions of pages across the web. They use a programme called a search engine spider (sometimes known as a robot) to crawl through webpage after webpage, downloading copies of the pages into their servers.

You can see what Google saves by checking the “Cache” option which drops down if you click on the green arrow on the right-hand side of each result.

Tip – How to find Lost pages: Sometimes you can retrieve lost webpages (or pages that have been deliberately deleted) by finding the copy kept in Google cache. Do remember that Google regularly updates the copies it keeps in cache so this is only a short-term option. You can also use the waybackmachine.com, which we will discuss in more detail later.

The pages indexed by search engines are known as the surface web and represent a small fraction of the overall information held online. Exactly how much is disputed but the perhaps around only 20 per cent of all online information is held on the surface web. Surface webpages have their own URL, the Uniform Resource which is often referred to as a “Web address”. Remember, even within this section of the internet which can be indexed by a search engine, there is no guarantee that Google has indexed every page. 

The Deep Web: Beyond the surface web lies the “deep web” which, if we are to believe the estimates, represents the remaining 80 per cent of the data. The “deep web” is much more than its image of a shady underworld of arms bought with bitcoins and hitmen hired on secret forums.

Here is a summary of some of the information contained the deep web:

  • Databases that are accessed via a search interface. Public databases of companies and land ownership are often free to search but are kept in a “walled garden”. As a result, you will not find results from the archive with a Google search and need to search through each relevant database. 
  • Password-protected data: This includes court registers, such as the US court archive Pacer, and paid-for databases for company databases such as Orbis.
  • A page not linked to by any other page: There may be pages on a particular website which the company does not wish to make public and are invisible to search engines but can sometimes be found through playing with the URL structure.
  • The dark web: The dark web is defined as a section of the deep web which is intentionally hidden from search engines and accessible with a particular type of browser which masks your IP address (that’s the equivalent of the home address of your computer). The best-known browser for entering the dark web is Tor (The Onion Ring). The dark web is where some of the seedier activities of the internet occur (it was where the online drug marketplace Silk Road operated until it was shut down by the FBI). While there are no doubt corners of the dark web for journalists to explore, Tor browser is most useful for protecting journalists’ privacy, of which we will speak more later.
  • Social media: Social media represents parts of the web often referred to as “walled gardens”, such as social media websites like Facebook and Twitter. This means that the treasure trove of information is invisible to Google and needs to be searched directly through the search engine of the social media application. Use the social media’s own search function.
  • Historical pages: pages which have been removed or modified are not searchable but are maintained by various websites, the best being Wayback machine. Here you can find not just deleted webpages but also deleted documents.
  • PDFs and images: Google has become increasingly sophisticated in its ability to turn data such as PDFs and photos, previously arguably part of the “deep web”, into searchable material. Google transforms PDFs into searchable text using optical character recognition, which is improving by the day but still fails to convert all the text particularly when it encounters different languages, unusual characters and poor quality PDFs. As a result, there remain huge quantities of data kept in unreadable formats online. If you inspect the mountains of data and information collected by the UN Security Council committee, it is all stored in the form of PDFs. Some of this is searchable, but swathes of the information, mostly original documents which have been scanned in, are lost to the search engines. Unless you read through the documents carefully you may miss key pieces of information.

Key lesson: Do not rely on Google to find information, no matter how good it is. Check webpages manually, find walled-off databases, read PDFs thoroughly and learn to search social media.

Mastering Google

Google remains the greatest tool for journalistic research and all reporters must learn how to harness its potential.

Here are some key tips, from the extremely basic to more advanced, that you need to make the most of Google.

Rule 1: Quotation marks

Use quotation marks around your search term to find a specific name. If, for example, you are searching for references to Lawrence Marzouk and do not include the quote marks (“Lawrence Marzouk”), Google will return pages with references to Lawrence and Marzouk, but not necessarily the exact phrase “Lawrence Marzouk”

Tip: Tracking an individual

When you are searching in Google for information about a certain person, remember that names can be misspelt or transliterated in various ways. It is particularly important to look for variants if the name is originally written in a different alphabet. For example, in BIRN’s investigation into Damir Fazlic’s business in Albania, it was important to look for stories in a variety of languages. Damir Fazlić is sometimes spelt “Damir Fazlic” and “Damir Fazliq” when it is published in some Albanian media. We would therefore search for the variants of his name with OR (note the capital letters) between each search term.

While researching a story about Euroaxis, a Serbian bank set up in the Milosevic era in Russia, we found that we had to try multiple spellings to uncover all the relevant information. We had a similar problem while researching the Palestinian businessman and politician Mohammed Dahlan (whose name you can find transliterated into the Latin alphabet in more than a dozen ways).

It is very helpful if you understand how letters are transliterated in order to find alternatives. For example, Euroaxis in Cyrillic is officially written Евроаксис with the Euro changed in Evro. If it is transliterated phonetically it appears as Еуроаксес. Both of these spellings occur and searches revealed useful information. Finally, some outlets had transliterated the spellings back into Latin, with some stories writing about “Evroaksis” or “Euroaksis”. 

When researching Mohammed Dahlan, we tried a variety of spellings, but most useful was using Google in Arabic. While my knowledge of the language is limited, I was able to use Google Translate to write out his name in Arabic script. You can select Arabic as your base language and as you type Google translate will automatically transliterate into Arabic script. You may need to ask an Arabic speaker for help in getting this right, but Google Translate often works well.

It’s worth remembering that on member official registers sometimes a person’s surname comes first and often it includes middle names. In this case, a simple search for “Damir Fazlic” or “Mohammed Dahlan” would not return results. You will need to search for “Fazlic, Damir” for some official forms. 

To tackle the middle name issue use the * (known as a wildcard) which represent one or more unspecified word for example “Mohammed * Dahlan” also shows results for “Mohammed Yusuf Dahlan”.

Last, you can use Google Trends (www.google.com/trends), which tracks trends in Google searches, to find out how other people are searching for your target.

Rule 2: Narrowing your search

Using the world’s biggest library has its benefits, but how do you filter the information to what is important to you? Once you have found it, do you trust the information? You can do this by using two Google search commands “site:” and “:inurl” which we will explain here.

Trawling government archives

The best (although not unimpeachable) source of information is from governments and official bodies. Luckily Google allows you to filter results in that way.

Most countries in the world have “gov” in the domain name of their official websites. The US has “.gov”, the UK “.gov.uk”, Albania “gov.al” and Kosovo “rks-gov.net”. There are exceptions within countries and internationally, of course – the Prime Minister’s office in Albania is kryeministria.al and German official websites end with just the country prefix of “.de” – which you should take into account.

By typing your search term followed by site:.gov in Google you will be searching for any webpages with your search term within websites with domain names which ends in gov. As a result, it will return official, US government webpages.

Type: “Lawrence Marzouk” site:.gov

Result: All references to “Lawrence Marzouk” foundon websites which end in .gov

If you type your search term followed by inurl:gov you will be searching for any webpages with your search term where the domain name includes .gov.

Type: “Lawrence Marzouk” inurl:gov

Result: Government webpages from around the world, including gov.al (Albania), gov.rs (Serbia) and gov.uk (United Kingdom), which contain “Lawrence Marzouk”.

Narrowing your searches by country

Beyond looking at government webpages, you may want to limit your searches to a particular country. For example, perhaps you are looking for references to someone in the Spanish media. 

Type:  “Lawrence Marzouk” site:.es

Or for Albania:

Type:  “Lawrence Marzouk” site:.al

Of course, not all webpages are country-specific (there are many webpages which end in .com without being American), but this does give you an option to narrow down your search and is particularly useful if you are looking for media stories from a particular country.

Rule 3: Filetype

It is possible to narrow down your search based on the type of document you are searching for. This is done with the “Filetype:” search term.

This can be useful when you are looking for official material. Often companies, organisations and governments upload important documents (yearly accounts for example) in the form of PDFs, Word documents or presentations.

This function allows you to hone in on these documents.

So, if you type your search term followed by Filetype:PDF and inurl:gov Google will search for all pages with your search term in a PDF and where the domain name includes gov.

Type: “Lawrence Marzouk” filetype:pdf inurl:gov

You can select different types of filetypes (and indeed carry out much more advanced search options) by clicking on the cog at the top right-hand side of the screen and then selecting “advanced search”.

Be creative. 

There are many other useful search commands you can use – too many to list here. But you can check out most of them here: http://www.googleguide.com/ and practise using them in your own time.
The final thing to consider when carrying out a Google Search is that even if you master the commands, you will not get the full use out of Google unless you think creatively and think carefully about what search terms to use. Try to imagine what sort of page you are looking for, what type of format the information you are looking for might come inand alongside what other words. Also remember that a person or company name is not the only thing that identifies them – you might want to try a phone number, email address, website or office address.

Image search

Google Image Search:

Images can also be extremely helpful sources of information for your story and also provide leads to other information.

Google provides a means to see if and when a photo has been used elsewhere on the internet.

If you click on the “image” option on Google, you will notice on the far right-hand side of the search bar a camera icon. 

Click on that, and it gives you the ability either to upload a photo from your hard disk or include the URL (the web address) of an image. 

Once you have done that, Google will search online for any other examples of that same image and will even come up with similar pictures, although the latter is not of much use to journalists.

This allows you to find other webpages which have used the same image, for example, new stories from a variety of countries or different social media profiles using the same photo.

It can also help to check if a company’s claims are authentic.

When BIRN was looking into the New York-based firm Siva Partners, which announced a series of developments in Albania and Kosovo in 2012 meeting the then Albanian PM Sali Berisha, we checked the photos of its offices and boardroom which featured on its website and looked unusually grand.

Is this office really what it seems? By running the photos through Google images, we were able to discover that its boardroom was rented office space and not the grand boardroom of a major company.

Google also has an advanced image search option which allows you to tinker with colours and add search terms to narrow down your options: https://www.google.co.uk/advanced_image_search

Google Maps and Earth:

Do you need to check out the offices of a company in London, New York or Paris? 

Perhaps you would like to check the development of a new building in a capital city? Zoom in on a restricted military base? Find out what a prominent politician’s home looks like behind the tall walls. 

Google Maps, the online search function from Google,and Google Earth, the downloadable application, can help out with all of these. 

Search for particular places using the satellite view or, where it is available, use Street View to get a roadside picture of a location.

On Google Map, using Street View, you will notice at the top left-hand side the option to scroll back through time and look at how a location has evolved. You may notice interesting developments, changes to posters or intriguing vans parked outside.

Google Earth (which you will need to download) offers historical Satellite images by clicking on the timer at the top right on the screen. You can also find street-level 3D images which can be helpful to visualise a location.

Unfortunately, Google Maps offers historical images of Street View and Google Earth of satellite pictures, but not vice versa, so you have to toggle between the two.


Street View hasn’t arrived in Tirana yet but you are able to scroll through historical satellite images and use the 3D street-level view.

Tip: Searching for people in Google maps

It is possible to search for people based on email addresses, twitter handles or just names through Google Maps. The results are a little erratic but it is worth a try.

This article is originally published in BIRN Albania’s manual ‘Getting Started in Data Journalism’

BIRD Community

Are you a professional journalist or a media worker looking for an easily searchable and comprehensive database and interested in safely (re)connecting with more than thousands of colleagues from Southeastern and Central Europe?

We created BIRD Community, a place where you can have it all!

Join Now