sentiment-graph

Exactly what did people think about the Scotland Decides TV debate?

Should Scotland go independent? We can’t answer that question, but we can help answer this: exactly what did people on Twitter think about the TV referendum?

People love to tweet while they’re watching TV. The recent televised debate between Alistair Darling and Alex Salmond prompted over 250,000 tweets with the hastag #bbcindyref. And that in turn prompted us to ask – wouldn’t it be great to know exactly which moment of the TV debate each tweet was referring to? And then, even better – what if we could say whether those tweets were positive or negative, to get a sense of who was winning public opinion at any given moment?

The first task was to work out whether or not all 250,000 tweets were really worth using. We filtered out tweets that occurred before the TV debate started and after it ended, and also removed retweets and spam tweets. That left 34,590 tweets. Next we put a sentiment analysis on each tweet, to figure out whether the tweet was positive, negative or neutral. Then we mapped each tweet onto a video of the TV debate, and added some nice gadgets, to let you drill-down into the detail of each moment and change the precision of the trends.

The result: a fully-functioning (if slightly hacked together) Reframed video twitter sentiment analysis!

sentiment-analysis

Ok so what are we looking at here?

Sentiment graph sentiment-graph
The first thing to look at is the sentiment graph. The TV debate took 89 mins, 03 seconds, so what we’ve done is split that time into 100 equal bars, each being around 53 seconds. Each bar is a summary of all the tweets that took place in that 1% of time.

The colour of the bar shows the average sentiment of tweets during that time: bright green is very positive, bright red is very negative, and everything else is inbetween. The size of the bar indicates the volume of tweets with the hashtag #bbcindyref at that time.

Click on the graph graphclick
Click on a bar, and on the left of the screen you’ll see a list of all the actual tweets which were sent during that 53-second segment of the TV debate. On the right of the screen you’ll see the most frequently used keywords used in tweets. Green words are positive, red are negative, and blue is every keyword regardless of sentiment.

Best of all, look at the video above the graph – it’s ready to watch at exactly the right time for the Twitter chat you’re reading. So you can watch the video and know exactly what people were saying – and how they were feeling – at that point during the live broadcast.

Try it for yourself
If you’d like to find our more about our sentiment tool, have a demo or talk to us about doing something similar on your video footage, get in touch: hello@reframed.tv.

Weekly curated livestream content directly to your inbox

Including all the top livestreams you may have missed, plus you can stay up to date with all the latest at Reframed.

We will never spam you or share your details with any third parties.