Data abnormalities, which deviate from general voting trends, are often discovered during analysis of election outcomes. Such abnormalities are usually detected in the process of calculations and data visualization. They can also reveal falsification of voting results or misuse of administrative resources/voter bribery during an election campaign. However, in most cases, these abnormalities can be explained by regional peculiarities, type of polling stations, specifics of an election campaign, etc.
First round of the presidential election
We have analyzed election outcomes in Donetsk oblast and noticed a big number of abnormal polling stations, where one candidate received much higher than the average number of votes in the region. To visualize this abnormality in infographic, we placed polling stations in Donetsk oblast on the axis. Thus, X axis held polling stations numbered as on website of the Central Election Commission, and Y axis - the percentage of Petro Poroshenko's results.
We can see on the infographics that the percentage of votes for Petro Poroshenko sometimes exceeded 70%, while the average number of votes for the candidate in the oblast was below 10%.
To illustrate this abnormality, we visualized the results of Petro Poroshenko in Luhansk oblast, which is similar to Donetsk by its electoral tastes.
We can see on the infographic that there are several exceptions, but most of them - are temporary polling stations at military units, where Metro Poroshenko has higher support than the average in the region. (You can build similar charts for other oblasts in the end of the article).
Not only the data but also the media mentioned abnormal results of Petro Poroshenko in the 1st round in Donetsk oblast. Videoswithballot box stuffing for one of the candidates were shared online. Besides that, the National Police of Ukraine opened a pre-trial investigation into falsification of election documentsat two polling stations in Donetsk oblast.
Second round of the presidential election
The abovementioned abnormality in the Donetsk oblast has almost disappeared in the second round of elections. Only temporary polling stations at military units and several polls in Marinka raion stood apart from the general picture.
However, the second round results of Donetsk oblast are interesting by other features. Donetsk oblast was the only one, where the number of votes for Petro Poroshenko in the second round decreased in comparison to the first round. The average number of votes for Petro Poroshenko increased by 44% compared to the first round, but this percentage decreased by 19% in Donetsk oblast.
Thus, 107,284 voters supported Petro Poroshenko in the first round in Donetsk Oblast, and 86,550 - in the second round. In search of a logical explanation, it can be assumed that a part of the voters in Donetsk oblast, who voted for Petro Poroshenko in the 1st round, changed their choice in favor of another candidate, and Volodymyr Zelensky gained the voters, who voted for other candidates. But we do not believe in such assumption..
Thus, to investigate this abnormality in detail, we made an example to compare votes for Petro Poroshenko in the first and the second rounds. The tool is based on the difference between the number of votes received by Petro Poroshenko in the first and the second rounds. By setting the range of lost/received votes for Petro Poroshenko, one can visualize the data and independently find and explore abnormal polls.
Click on an image to go to the tool:
We tried to estimate the number of abnormal votes for Petro Poroshenko in the 1st round in Donetsk oblast. Thus, we took the number of votes for Petro Poroshenko in the second round - 86,550, and the average change in the number of votes for the candidate in adjacent oblasts - + 30% of votes in the second round.
107,284 – (86,550/1,3)=40,707
Thus, we received a 40,000 votes, or 40% of the votes cast in the Donetsk oblast.