Analyzing Tweets From the Biggest Storm in California in Decades

It was a rainy Friday afternoon when the email came from our friends at CrowdFlower. “Hey, we collected and analyzed 10,000 Tweets about the big story that just hit. Can you give us some visualizations?” We said “No problem!” and a few minutes later our California Storm Tweets Analysis Silk] was born.

As Silks go, it’s a simple one — only a few visualizations, not a lot of data on the pages. But what CrowdFlower liked was our ability to turn around a live interactive visualization in minutes. So we helped them put together a blog post and visualize the data.

Here’s a bit of context. On December 10, a huge storm hit just North of the Bay Area. This “Pineapple Express” weather event flooded homes and businesses, closed schools, plunged over 100,000 Bay Area residences into darkness, raised waves of nearly 30 feet, and delivered pounding rainfall of over 6 inches in 24 hours for some locations.

CrowdFlower took the opportunity to harvest sentiment from Twitter and find out the most prevalent sentiments expressed about the #BayAreaStorm. To do this categorization CrowdFlower used it’s distributed workforce of remote workers, assigning them small online tasks to analyze the sentiment of tweets the CrowdFlower data science team had collected. The crowd was asked to categorize the sentiment in the tweet (e.g. power outage, humor, etc) as well as to capture the number of retweets. (Ed Note: This is a great use case for CrowdFlower: a number of large advertising agencies and PR agencies use CrowdFlower specifically for this purpose.)

The Silk Data Journalism team took the data, scrubbed it for dupes, and reduced it down to 9300 relevant tweets and then published the data on Silk for quick visual analysis and to build visualizations.

The findings? In typical laid back NorCal fashion, Bay Area residents were more amused by the storm, than they were concerned about road closures or power outages. (See the above donut chart for the precise breakdown).

The most popular tweet came from the social media team at NetFlix: “@netflix: Netflix Recommends: Working From Home tomorrow. #NetflixWeather #BayAreaStorm” got a whopping 514 retweets (as of this posting, it was up to 589. Very clever. Check out the Silk visualization above and sift through the text of the tweets by hovering over the lines on the bar chart to get a taste of what tweets garnered the most interest. If you have ideas for interesting social media visualizations, please find us on Twitter @silkdotco and let us know what you think of this blog post, too.