The Parallel Ranking Use Case

For this use case I’ll use statistical data from the EU regarding airline operations in 2012. For each country, I have several measures available :

  • Number of flights departing from a given country.
  • Number of seats available on flights departing from a given country.
  • Number of passengers boarded on flights departing from a given country.

To simplify this example, we’ll use only the 10 countries with a higher number of departures in 2010. The challenge here is to create a top 10 ranking using this countries and to be able to compare how a country who is the number one in a given measure (let’s say number of flights) is ranked along the number of boarded passengers. This is the final result :

multiple_top_10

A┬ásimple conclusion from this chart : France is the third country with most flights in 2012. But when we look at the number of seats available and the number of passengers, it ranks the fourth place. Spain has the opposite behavior. One explanation for this may be the size of the aircrafts departing from one country and the other, but for now we’re not going to focus on that “why” analysis.

To achieve the above results, you’ll need to follow this steps :

  • Start by creating a bar chart for each measure. Since we’re creating a top approach, order each bar by descending Y-Value. Each chart should have distinct expression (in this example, something like Sum(Flights), Sum(Seats) and Sum(Passengers)), but it should share the same dimension (Country in our example).
  • When you click in one of the bars, the default behavior is to filter the existing data to show only that country. That is not what we need, so we need to change our expression in all the charts to something like this :

  • This set analysis will make the chart┬ánot responsive to the country selection. When you click on UK, UK will be filtered, but the chart will not respond to show only that country. This is ok, but we still need to know what country is selected, and to see that selection on every chart. For that, we’ll need to tweak the background color with this expression :

  • It’s easier to explain that this works, why it works is more complicated. My understanding is that the Country=Country expression will be evaluated looking at the first Country as the dimension on the chart and the second Country as the document field value. Therefore, when the dimension value is the one that is selected, the bar will be painted red.

Hope you’ve enjoyed this one.

Leave a Reply

Your email address will not be published. Required fields are marked *