As we all know, Zoran Milanović won the Croatian presidential elections and became the fifth president of Croatia. But how it actually happened? How did people vote and whose votes brought victory to Mr. Milanović? What was the people’s vote and which counties helped his victory?
We’ve created a thorough statistical data science-backed analysis of the recent Croatian presidential elections. The presented data is heavily visualized and shows several crucial moments of the electoral battle. There’s a lot of healthy conclusions to be drawn from this analysis, especially when it comes to trends between the two electoral rounds, regional distribution of new voters, regional distribution of invalid ballots, specific relationships between Grabar-Kitarović’s, Škoro’s and Milanović’s voters in both electoral rounds, etc. A lot of hidden gems for those who know how to find them.
All data was taken from the website of the State Electoral Commission of the Republic of Croatia.
In the first round the turnout was 51.15%, with a total of 1,997,149 people voted, while in the second round it slightly increased to 54.89%, with total of 2,151,475 people voted by the number of ballots.
From the graphs below we can see that the turnout was better in the second round in all counties except in Požeško-slavonska. However, the number of valid ballots in second round was smaller in four counties: Osječko-baranjska, Požeško-slavonska, Virovitičko-podravska and Vukovarsko-srijemska.
The number of invalid ballots was higher in all counties in the second round. This can probably be explained with people’s tendency to void ballots if they are unhappy about the remaining two candidates. The county with the highest number of invalid ballots in the second round was Zagreb.
Polling stations with the most invalid ballots are located in Zagreb. If we remove the number of invalid ballots in the first round, we still have the highest increase in ballot annulment on polling stations in Zagreb.
In the first round, Zoran Milanović won 29.3% of all votes, Kolinda Grabar-Kitarović 26.9% and Miroslav Škoro 24.7%.
Purple color on the graphs below represents candidates that got more than 10% of all votes, while orange represents candidates with less than 10% of votes. The same coloring is used for counties where each candidate won more than 10% of all votes in that county.
For more details, we will take a look at just top three candidates. We can see that Zoran Milanović won in Zagreb, Istarska, Zagrebačka, Varaždinska, Međimurska, Primorsko-goranska, Koprivničko-križevačka and Krapinsko-Zagorska. Miroslav Škoro won in Vukovarsko-srijemska, Osječko-baranjska, Brodsko-posavska and Bjelovarsko-bilogorska; while Kolinda Grabar-Kitarović won in all the others.
We want to draw your attention to the fact that people who voted for Miroslav Škoro were mostly from Slavonia, Zagreb and some from Splitsko-dalmatinska. Remembering this will be useful in the rest of the analysis.
As we said at the beginning the number of votes in the second round slightly increased by 3,74%. However, you can see from the graph below that in 5 continental counties fewer people voted in the second round in Vukovarsko-srijemska, Virovitičko-podravska, Požeško-slavonska, Osječko-baranjska and Brodsko-posavska. Perhaps the majority decided not to vote in the second round because their candidate lost in the first round.
Also, interesting fact according to statistics is that Splitsko-dalmatinska county turned out to be the county with the highest number of new voters in the second round.
If we look at the changes of valid ballots for each polling station, we can see that on most of them we have an increase in the number of valid votes. More people voted in the second round on almost all polling stations in Croatia.
However, if we take a look at the lateral dispersion of the polling station that had a decrease in the number of valid votes, we can see that some parts of Slavonia and Dalmatia stand out. On those polling stations, fewer people made valid votes because their candidate lost in the first round. If we recall image with the location of Miroslav Škoro voters, the positions match. As we know Miroslav Škoro asked his voters not to vote or to annul their ballots in the second round after he lost in the first round. Maybe these were the polling stations that responded positively to his request.
The most interesting part is Split, where we can see the highest positive and negative changes in the number of valid ballots. Over 500 new people voted on some polling stations, while on others more than 500 people decided not to vote in the second round. This graph highlights the changeable mood of people living in Split. Blue color represents an increase, while the red color represents the decrease in the number of valid ballots.
In the second round, Zoran Milanović won 50% of all votes, while Kolinda Grabar-Kitarović got 45.6%, and 4.4% of total votes were considered invalid.
Let's now review how candidates performed in each county. We can see that Zoran Milanović won in Zagreb, Istarska, Zagrebačka, Varaždinska, Međimurska, Primorsko-goranska, Koprivničko-križevačka and Krapinsko-Zagorska. Kolinda Grabar-Kitarović won in all others.
Purple color on the graphs below represents counties where each candidate won, while orange represent counties where he/she lost.
We can see that voting centers for Zoran Milanović were Zagreb, Varaždin, Rijeka, whole Istria and Osijek. While for Kolinda Grabar-Kitarović centers were Zadar, Split, Zagreb, diaspora and Osijek.
We can assume that everyone who voted for Zoran Milanović in the first round voted the same in the second round as well. Based on that, we can analyze just the change in the number of voters from the first to the second round and call them Zoran Milanović’s new voters. We will use the same assumption for the new voters for Kolinda Grabar-Kitarović in the second round.
In the second round, Zoran Milanović got most of the new votes in Zagreb, Splitsko-dalmatinska, Zagrebačka and Primorsko-goranska.
In the second round, Kolinda Grabar-Kitarović got most of the new votes in Zagreb, Splitsko-dalmatinska, Osječko-baranjska and Zagrebačka.
Let's compare the number of new votes in the second round for Zoran Milanović and Kolinda Grabar-Kitarović with the number of votes for Miroslav Škoro and others in the first round. We will also plot new people who didn't vote in the first round to have balance in the number of voters. This graph will show us whom did the voters chose in the second round after they lost their candidate in the first round.
Again, we can see that voters from several Slavonian counties (like Osječko-baranjska and Požeško-slavonska) preferred to boycott the second round since their candidate didn't pass. We can see that voters didn't just shift from Miroslav Škoro to Kolinda Grabar-Kitarović, some of them did, but not as many as expected. That means some people boycotted, some selected Kolinda Grabar-Kitarović but many selected Zoran Milanović in the second round.
If we compare new votes for Zoran Milanović in the second round and number of votes for the other candidates in the first round (without votes for Zoran Milanović, Kolinda Grabar-Kitarovič, and Miroslav Škoro) plus all the new voters that didn't vote in the first round, we can see that the number of votes for Zoran Milanović in some counties was higher. This means that probably part of the Miroslav Škoro voters moved to Zoran Milanović in the second round.
The counties with the biggest differences were Osječko-baranjska (around 15 thousand people), Zagreb (15 thousand people), Vukovarsko-srijemska (7-8 thousand people), Brodsko-posavska (7-8 thousand people), Zagrebačka (7-8 thousand people), Virovitičko-podravska (around 5 thousand people). This means that around 62 000 people changed their votes from Miroslav Škoro to Zoran Milanović. Which is roughly around the number of people that led to close victory of Zoran Milanović over Kolinda Grabar-Kitarović.
We got this number with assumption that all the new voters in the second round and everyone who voted for the others in the first round voted for Zoran Milanović. This is certainly not true, which means that even larger proportion of people who voted for the Miroslav Škoro in the first round voted for Zoran Milanović in the second.
If we compare the number of new votes for Kolinda Grabar-Kitarović in the second round and number of votes for Miroslav Škoro in the first round, again we can see that part of Miroslav Škoro’s votes transferred to Zoran Milanović. As the number of votes for Miroslav Škoro in individual counties was higher than the number of votes for Kolinda Grabar-Kitarović, we can conclude that Kolinda Grabar-Kitarović lost her votes in those counties.
Let’s now compare the total number of voters in the first round with the number of voters in the second round. We assumed that everyone who voted for other candidates in the first round plus the new voters who didn't vote in the first round have voted for Zoran Milanović in the second round. This gives us 72 600 people who voted for Miroslav Škoro in the first round voted for Zoran Milanović in the second round.
If we take a look just at some counties in Slavonia, for which we know part of the people boycotted elections, we can see that around 38 000 people that voted for Miroslav Škoro in the first round voted for Zoran Milanović in the second round. Most of them did vote for Kolinda Grabar Kitarović, some didn't vote but many of them changed their votes to Zoran Milanović