Denmark gone bankrupt
Organisation: Danish Broadcasting Corporation (Data Team) (Denmark)
Publication Date: 03/30/2016
Size of team/newsroom:small
DescriptionWhen companies go bankrupt, to benefits? That was the main question that led us to do the unique investigation based on public data in the search for a valid answer. DR's Investigative Data Team mapped bankruptcies in Denmark in the period 2004-2014. 45.000 Danish companies went bankrupt doing this period. While it is well known that the financial crisis has given bankruptcy-riders and frauds good conditions, it is unknown to the public how the majority of bankruptcies are handled. How long time does it take before a company is dissolved and any assets distributed? What do lawyers, who treat the thousands of bankruptcies, profit? And what will be left for the creditors? The Data team set out to dig into data on bankruptcies to examine the consequences by the wave of bankruptcies that have washed over the country the past 10 years. Earlier, the Bankruptcy Council analyzed datasets with a few hundred bankruptcies to assess how long the proceedings lasted. The Investigative Data Team have pooled numerous of public datasets managed to create a unique overview of more than 32.000 finished Danish corporate bankruptcies. By means of the massive datasets, and additional research as e.g. crowd sourcing, the public has for the first time become aware of the following revelations: - Banks and lawyers eat up estate from bankruptcies. - Lack of supervision of bankruptcies may have cost creditors uncovered debt in thousands of cases. Minister of Justice promises action after DR's disclosure. - Bankruptcy-lawyers has for 20 years been allowed to break the VAT-law by not reporting VAT continuously when there's been sold out of the bankruptcy estate values.
What makes this project innovative? What was its impact?This project is unique - It was actually an impossible task. Originally the Datadesk was asked to assess whether it would be possible at all to gather and calculate data on bankruptcies. Such data is not collected and published systematically from Danish courts or authorities. The conclusion was that such a task would be nearly impossible. Data on bankruptcies would be prohibitively expensive to gather, since it would be dependent on FOIA-requests at 175 DKK per case, and furthermore very complicated to process, since reports are not uniform in format and issued only on paper. We needed to find other ways to gather the data in bulk. This was done by collecting a full list of bankruptcies from the years 200-2014 and using the company details to scrape other databases, including the Danish official public records gazette Statstidende. All collected data were merged in a single database for analysis. The primary source was a list gathered from the public Danish company registry CVR giving a full list of bankrupt companies from 2004 to ultimo 2014. This list contained names and registration numbers of companies and the date of declaration of bankruptcy. This was used as a base for collecting location data on the companies, information on the courts and bankruptcy lawyers that treated the cases, information on how companies were dissolved, including whether funds would be paid to creditors or not (but unfortunately not information on amounts of funds paid) and dates for court meetings, including the final closure of the cases. Data was mostly collected from CVR, from private company registries and from the Danish official public records gazette Statstidende. Much of the collected data was correlated by hand and checked up against FOIA-requests. Journalist Bo Elkjær spent about four months on research and data processing and dissemination to the web and cooperation with the documentary television team.
Technologies used for this project:Getting, analyzing and visualizing data: The initial data-set was scraped from a website using C# to iterate through and parse data from relatively well-formed HTML. A package called HTMLAgilityPack was used to do this easily. Data was saved as a simple CSV-file that later was opened and analyzed in Microsoft Excel, Google Sheets and Google Open Refine. The data-set contained more than 50.000 records and was visualized on a map of Denmark. To plot and animate that amount of points D3 was used - along with TopoJSON, TweenLite, TimeLineLite on a canvas-element.
You have to be connected to contribute
You have to be connected to follow
Leave this project and no longer be informed about this project
By joining this project, you will be informed by email when an update or a new contribution is posted on the website.
Thank you for your active participation !
The GEN Community Team