Winter Olympics Analytics with Bottlenose

  2 m, 24 s

The Winter Olympics have dominated media this month. Everything from Sochi to Shaun White has been endlessly picked over by news outlets and social media. Using Bottlenose, a trend intelligence platform, we set up a Stream to monitor the Olympics so that we could keep up with the conversation.

Bottlenose takes unstructured social data from Datasift, and builds out a conversational map. Its core technology – StreamSense™ – continuously data-mines to detect an unlimited range of entities and relationships as they emerge and develop, including topics, phrases, opinions, links, people, organizations, events, and the connections between them.

Datasift provides a highly flexible, reliable service for aggregating and filtering social media streams, content which is then fed to Bottlenose for processing.

Bottlenose tracks 116 metrics per entity. Salience powers Bottlenose’s sentiment measurements, one of their most important metrics. 

The sheer volume of tweets and other social media mentions means that the numbers represented in the following graphs are a sample size for the purpose of analytics. 

To start us off, here’s a look at the geographical distribution of tweets. The United States dominates the field, with half the total volume of Olympics related tweets. The second largest contributor is Russia, which is unsurprising. What is surprising is how neutral the Russians seem to be when talking about the Olympics. Canada, whose volume in our sample is approximately 2,000 than Russia, has positive and negative sentiment scores far higher than that of Russia. Russia’s positive to negative sentiment ratio is also much higher, but as the host country for the Winter Olympics, that doesn’t seem too surprising.

Here’s a breakdown of tweet volume by state. According to this map, we can safely say that the state that cares least about the Winter Olympics is Mississippi. What’s impressive is that Mississippi, with twice the population of Montana, and four times that of Wyoming, beat both states for this dubious honor.


And while we’re on demographics, it appears that women talked about the Winter Olympics far more than men did.


We also have positive and negative sentiment volumes by topic. Many of the two overlap, but a few stick out. Shaun White, Olympic gold medalist and American snowboarder, is exclusively negative, probably due to his high profile failure to medal at this years Olympics. 



Finally, the graph I find most interesting, the psychological profile – which breaks down social media into more specific emotions. The top four emotions by volume were Social Behavior, Expressive Behavior, Aggression, and Glory, with Glory taking a late lead as the Olympics wound to a close. 


The Winter Olympics generated a lot of chatter. With Bottlenose we can thoroughly examine much more than simply what was being talked about. It’s a powerful tool with a power engine at its heart. Overall, I think we can come to at least one conclusion: Stay frosty, Mississippi.

Categories: Analysis, Social Media