Votes over the years
On the first page of the dashboard (switch between dashboards with the arrows below the dashboard), it is possible to see the voting behaviour of the Dutch over the years. The amounts shown on this page are all averages. When you select for example one year (the year 2006), you will see the amount of seats taken by a political party (of 150 seats). When selecting all the years you are able to see that (still active parties) CDA and PvdA are on average the parties with the highest amount of seats. When selecting the more recent years, it becomes apparent that the VVD is joining these parties. Along with the amount of seats changing, it is also visible that the trust of the Dutch is changing. In 2003 the Dutch gave their trust in political parties a 4.9 (of 10), while in 2006 their trust increased to a 6.2. Maybe the growing economy made the Dutch hopeful?
Votes by region
The second page of the dashboard shows the voting behaviour of the Dutch by region. It is known that right-wing party PVV (Geert Wilders) is doing well in the province Limburg. But are we able to see factors why that might be? Therefore we took some pointers of the campaign of the PVV and tried to quantify them: Crime (crime rates), health care (hospital intakes), and immigrants (% of immigrants of the total inhabitants). Of course these are not good quantifications, but we still had fun with it.
Along with the vote data per provinces, we have mapped these pointers of the campaign as well. When looking at the map of the Netherlands, we see that the province Limburg has the darkest colour. This means that the percentage of PVV-voters is highest. When hovering over the provinces, we see the average percentage of votes in that province (for Limburg this is 17,44%). On the right side of the map, we see the average amounts for crime, health care and immigrants. When the whole country is selected (all provinces have colour) than we see the average for the country. When one province is selected (the colour of other provinces on the map become transparent), we see the averages for that province.
We expect that in the province Limburg crime rates, hospital intakes and the percentage of immigrants is higher than the average in the Netherlands (to explain right-wing voting). When we click on the province Limburg, it becomes apparent that the percentage of crime is lower than the average of the Netherlands, but hospital intakes and the percentage of immigrants is higher than average. This gives a hint of why people in the province Limburg might vote right-wing party PVV.
To create these dashboards in Power BI we took different sources of data (vote data, crime data) and combined them to make these visualisations. This is the power of Power BI and makes it possible to see interrelations. The data used for this example is not qualitatively very good, but it shows how to combine different kinds of information to get more insights. As an example: for an ice cream company we can look over time how the weather correlates to the sales, or we can map neighbourhoods with lots of families and compare it to the sales. Both pages in the dashboard are not statistical models, but it gives us some useful insights. From here on it is possible to explore the data further, look for more correlations or even make predictions about the future. For now, we will wait and see what the elections of today bring us.