The data has not yet been verified to be correct!
While some of the results of the findings here are very interesting, they have not yet been properly verified. They could be wrong, they could be right. If you can help everyone out by verifying that the data from Parltrack, and the scripting in the open source projects, is correct - please do! There is currently (2014-01-28) work in progress to require machine-readable sources of data straight from the European Parliament, but until that has been resolved, keep in mind that this data is based upon parsing PDF-files meant for human eyes and paper printer consumption.
The data has not been filtered to reflect any opinions on votings, voters or votes. The data has only been reduced to include only votings with recorded corrections.
During votings in the European Parliament (EP), each elected and present Member of the European Parliament (MEP) record their votes. Choices are For, Abstain and Against. If they discover that they made the wrong choice, they themselves can record their vote as being incorrect. This doesn't change the outcome of the voting, it is only done after the votes have been counted.
My inital thought regarding these corrected votes, is that they are purely erroneous button clicks in the electronic voting process. The number of corrected votes during certain votings, the number of corrected votes some individual MEPs have recorded and the obvious group belongings of the correctors, seem to tell a different story though. I now leave it up to you to dig deeper. I will be here to assist, if there are questions or suggestions.
This data, research, software, and analysis is driven by volunteer work and shared completely free. You even get the source code and data for free. To develop open source software is not without cost though — it takes technical knowledge, time, and effort.
You don't have to be a developer to support data-driven research! If you want to receive personal support, or just feel all warm and fuzzy inside from helping open source development, donations are very welcome.
Donate $5 now Donate $25 now Donate $100 now More options
The power is in your hands. Use the source, Luke!
Votings in the European Parliament that have corrections.
Changes: generated complete data for term 7. Added a summary to each dataset. Also rendered the same data for term 5 and 8, but there are no recorded corrected votes.
Changes: Added two new generated files: one with automated metadata/aggregate data about the dataset, and one with a table of the MEPs' corrections.
Changes: separate output per term in the European Parliament. This will likely be the only time the fifth term is included, because there is no voting data for term 5 available in Parltrack.
Changes: publishing aggregate data in a JSON file, showing how many occurances each MEP name variation has.
Changes: used a new dataset from Parltrack, that removed 2142 duplicate votings, plus has the right MEP database IDs.
Since the statistics, these and before this date, are run over the complete data set, it doesn't take the five year parliament terms into consideration.
Changes: used a new dataset from Parltrack, that removed 2142 duplicate votings.
Changes: improved links to MEPs pages on Parltrack.
Changes: removed empty links to dossiers.
Changes: votings are now listed by voting timestamp, in descending order.
Article in MM.dk, a Danish think tank, by Andreas Baumann, about the fight for Open Data in the EU including the digging for corrected votes.
The fight for digital openness is a movement of data democrats armed with laptops, who bit by bit create order out of the chaos of data that the political processes create.
A growing movement of activists fighting for the rights of citizens to know what is going on in the EU closed law making machine. Armed with laptops and open data they create structure in the myriad of data by EU institutions produce. Thanks to them, you can see that Poul Rasmussen and Britta Thomsen, the EU parliamentarians who most often corrects their votes - after they have been submitted. We try to make the democratic institutions of the EU to breathe easier, says the German data journalist Friedrich Lindenberg to Monday Morning. Data Democracy in Europe is under construction.
Article in Altinget.dk, a Danish political newspaper, by Rikke Albrechtsen, about the Danish social democrat MEPs and their comparatively high number of corrected votes.
Looking for errors: Danish Social Democrats in the European Parliament are those among all EU elected representatives, who most often corrects their votes. Not credible, says the man behind the collected data. Democratic diligence is the answer from the list's top scorer, Britta Thomsen.
After looking at the most recent data from Parltrack, I was surprised to see that the "correctionals" were gone from term 8. The Parltrack developers seem surprised as well, as they can't find the data in the source documents from the European Parliament. We hope to find the reason for this reduced amount of information, but until then feast your eyes on the analysis of the full dataset of term 7, which was previously not published here.
Added two new generated files: one with automated metadata/aggregate data about the dataset, and one with a table of the MEPs' corrections. The list has a built-in simple bar graph, which will make it just a tad bit easier to get an overview of the difference in number of corrected votes.
There have been several requests to split the voting data per term, so now that is done for term 5, 6 and 7. Unfortunately, there are no corrections (recorded or extracted) from 2004, 2005 and 2006 - it's missing from the upstream dataset. Term 5 is only there to verify that there are no votes that are from that term, and will likely be excluded until there is. Hopefully there this will also lead to more automation of data exports.
This evening, at least 2136 duplicate votings were found. They seem to stem from the fact that he has been sent duplicate lists, but under different voting titles or languages. This is yet another reason why being delivered original data directly from the European Parliament is important!
Update: a new, deduplicated dataset has been released from Parltrack. The dumps on this page, from 2014-01-29T23:44:56Z, should be based on it.
Thanks to a journalist, errors propagated to the data with normalized names, which is the source for the lists of "worst offenders", have been discovered!
Since the data source in Parltrack is PDF files, which is far from optimal, there are sometimes problems with parsing them into a usable database. Having incorrect names affects the total count, since the totals have been grouped by name. This is the most recently released list of MEPs (grouped by name) who have made corrections, ordered by number of corrections in ascending order - it contains the incorrect grouping. The other lists have been unlinked, at least for now, and I'll work on a solution.
Until the data has been verified by others, I would take it with a grain of salt. If you are apt at verifying data like this, please help out!
Now that the hackathon is officially over, there is a list of projects people have worked on. These projects are listed there!
The corrected vote data broken out in ep-erroneous-votes is being used to present a long list of votes. Not sure this would be called visalization, but with the help of investigative journalists, I'm sure something will come up!
It would be interesting to look at corrected votes in the European Parliament. How many are there, is there a pattern and such. Here's the start of a narrowed down dataset, with only voting that have corrections and only those who corrected their votes.
Brainstorming and development starts, during the Europarl Hackathon, in preparation for the European elections 2014.