The world cup 2010 was a great event - although sometimes quite unexpected things had happen. For example, the results of France were certainly worse than most people had expected.

As many of you might know, we have set up a service called "Mannschaft der Herzen" ("Team of Hearts") which monitored the sentiment in news texts, blog or forum posts and aggregated the results. We connected RapidAnalytics, the server behind RapidMiner, to more than 1000 online channels and analyzed thousands of texts per minute. One of the most surprising results: We predicted Spain as the winner of the World Cup even before the event had even started.

Look at the following screenshot - it was taken on June 4th,  one week before the tournament had started:

The bars show the amount of positive document segments compared to the amount of negative or neutral texts. The higher this ratio the higher we considered the winning chance ("Siegchance") of the team. Spain is the green bar on the right and you can see that the positive sentiment ratio is clearly highest for them.

It was also really interesting to monitor the change in sentiment over time. During the World Cup, the positive comments about the German team got more and more. Finally, Germany even had the highest winning chance but was again replaced by Spain during the finals.

But also the extracted topics turned out to be quite useful. It was easy to detect the fact that the English captain Ferdinand was injured which had a great impact on the sentiment analysis. The following screenshots show the sentiment change over time and the extracted topics as tag cloud displaying the sentiment of the topic by color and the importance by tag size:

The analysis was performed for German texts but it is of course also possible in any other language that is supported by RapidMiner (which is basically any language but sums up to about  30 if you need linguistic approaches).

Usually, we perform Sentiment Analysis for our customers to monitor their brands or products or their competitors. But it was really fun to analyze soccer texts for a change.

Hope you liked it. The web site showing also the cool and interactive flash charts of RapidAnalytics is still available on 

http://www.mannschaft-der-herzen.de

Cheers,
Ingo