The Nitty Gritty on #netDE: a twitter analysis

The Nitty Gritty on #netDE: a twitter analysis

netDE

You may have noticed that I skipped my weekly whiteboard visualization last week. This was because I became fixated on another idea and had to see it through. When it comes to data science, I am undeniably distracted by new possibilities. I promise you’ll enjoy another Whiteboard Data Viz on Wednesday, but until then…take a peek at what super glued me to the screen all week.

At the Social Larsons we are extremely data driven. We monitor our social media and blog analytics constantly, looking for ways to improve our performance. So when thinking of ways to tie my passion for data to community advocacy, my mind went to the popular hashtag #netDE. Have you used it? If you used it on Twitter over the past week I probably pulled it, and your post contributed to the story this data tells.

I pulled 3,017 tweets that contained #netDE. The majority [55%] were original tweets. The chart to the right shows the distribution of Tweets and Retweets.

TorT

Twitter allows you to pull tweets from the past seven days. Therefore, I gathered tweets from Sunday May 8th through Saturday May 14. I didn’t pull the Tweets from today because at 9am when I pulled the data, the number of tweets published was to small. The next chart shows how many tweets [with mentions of #netDE] there were per day. The red line shows the average number of mentions per day for the past six days. On average there are 502 tweets with #netDE in them per day. On Tuesday the 10th there was a large spike in this occurrence.

days

Of course, Lynsey and I are most interested in identifying the best time to post our tweets for optimal reach [specifically when including hashtag #netDE]. The chart below shows the average number of retweets and favorites by the hour of the day of tweets including #netDE. I found it interesting how many people are liking tweets that were posted in the early morning.

RandF

We also wanted to know what people were talking about when they used this popular hashtag. The word cloud below shows any word that was used more than 15 times. Not surprising is that #netDE was number one. You’ll notice that people are using #netDE and #delaware in conjunction. When I originally pulled this data and formed the word cloud, Lynsey was concerned about the negative trend that appears in the illustration. It is important to note that data science is a powerful tool for enacting change. Data itself cannot [and should not] be altered to the desires of its owner. What you see below is a real time, current events, birds eye view of what our state is challenged with recently. Proper data analysis feeds you the good, the bad and the ugly. It’s up to us to use the results to impact our future for the better.

wc

Lastly, I assembled a sentiment analysis. I like ending on this as it sheds a positive light on the relationship between the hashtag #netDE and the context in which it was tweeted. A sentiment analysis pulls the words I mentioned above, and assigns them with a positive or negative connotation. As you can see in the graph below, the overall sentiment when using #netDE is positive. This was a great lesson for Lynsey [and maybe for you?] that not everything is as it appears at first glance. The sentiment analysis is my new favorite data tool to use for our social media improvement. Imagine pulling sentiments for your small business based on reviews or mentions…there’s a lot of power there.

sentiment.png

There is so much we can do with this data set. We are excited to dig in a little more. I have not included any of the real nerdy stuff (My R code), but you can check it out on my github page. The data that we pulled is posted as a .csv if you want to play around with it!

Now…back to drawing visualizations on our whiteboard.


Data Dan Larson is a data enthusiast who loves problem solving. He looks for ways to utilize data to optimize the world around. He’s also a papa who loves relating his findings back to life as a modern parent.

Comments are closed.