In a recent post, we demonstrated how Power View can produce compelling visualizations of Twitter conversations. Several readers asked about the possibility of tracking sentiment in Power View. The answer is unequivocally YES; you can, in fact, produce compelling visualizations that reveal the public stance on a product or brand. Whether you want to know how the Tweeting public views your newest product or find out the success of a conference, Power View can provide a good indication of public opinion.

Let’s look at an example. We revisited the Twitter dataset to measure the subjective stance of a single event: the men’s gold medal basketball game between the United States and Spain.

Modeling Public Mood and Emotion: an Imperfect Science

Since we captured unstructured Twitter data, we were able to re-map our original data model to include sentiment. Using the Twitter Sentiment 140 API, we tagged each Tweet as Positive, Neutral or Negative. Before we proceed, it’s important to note that sentiment tracking tools are an imperfect science, typically requiring human intervention to filter false positives. For the purpose of this demonstration, we skipped the tedious manual labor step, so please accept the results with a few grains of salt, now on with the show.

The Setup

Check out our previous post for details about how we setup the databases, data feeds, tables, and reports. Remapping the data simply required importing a revised set of keywords, sentiment data, country filters, and related categories.

Sample of tables imported from SQL Server 2012 database into Power View to begin building interactive reports

Tweet Volume by Day

For our first visualization, we took the pulse on the general mood of Tweeters over the course of the games, as well as the breakdown of Tweets by date and day of the week. No real surprises here: most Tweets were neutral—indicating a high volume of amplified news, event updates, and general Summer Games chatter.

An example of a report analyzing the sentiment and volume of tweets related to the London Games by date and weekday

We can filter this view for additional insight, including sentiment and volume by country, language, time frame, or a range of keywords—specific sports, events, or participants (as we’ll demonstrate below).

Twitter Volume by Events

Let’s focus on a single event—one with a high volume of Tweets that’ll provide a rich dataset to drill into. We sampled the overall volume of events to determine the event with the highest conversation rate.

Basketball and Soccer received the most mentions on Twitter during the London Games


Basketball and Soccer are the clear winners in a photo finish. We selected basketball for our demo since the term is universally used—otherwise, we’d need to track and compile both ‘soccer’ and ‘football’—definitely doable, but it’d take a few extra steps.

A Courtside Seat to Twitter Chatter

On August 11th 2012, the U.S. and Spanish national teams competed for gold at the London Games. Digging deeper using keyword and sentiment analysis, we uncovered the most talked-about players during the men’s gold medal match.

We created a column chart to drill into rows of our data table, specifically sentiment value of all players on the U.S. and Spanish national teams in the time frame during and immediately following the gold medal match.

Kobe Bryant, Marc Gasol and LeBron James received the most mentions on Twitter during the men’s gold medal basketball game

If the all-star team was determined by Twitter volume and sentiment, the starting line-up would include L.A. Lakers shooting guard Kobe Bryant (U.S. national team), the dominant force on the Twitter court; Memphis Grizzlies power forward Marc Gasol (Spain), Miami Heat small forward LeBron James (U.S.), L.A. Lakers power forward Pau Gasol (Spain), and Oklahoma City Thunder small forward Kevin Durant (U.S.). Based on our data, the line-up would hear an earful of cheers and jeers, with only Durant escaping negative chatter (and Anthony Davis somehow missing out on positive feedback).

Tracking Sentiment in the Real World

This demonstration gives you a sense of the possibilities of visualizing Twitter trends and sentiment using Power View and it’s just a taste of the real-world applications. Obvious uses include tracking customer sentiment for your products, brands and activities, like trade events. Or analyzing customer emails to track key support issues. By tracking feeds from real-time data, your company can reduce response times and isolate potential issues before they become problems. No promises, however, that Power View can change the outcome of your gold medal match.

Visualizing Global Positive and Negative Sentiment from Captured Tweets of the Summer Games by Country

Experience Power View Today!

To play with and learn more about Power View be sure to check out the interactive demos on the Microsoft BI website. To experience Power View as part of the Excel 2013 Preview go to the Office 2013 Preview website and select the “For Enterprise” sign-up.