Graph comment sentiment data #62
Description
Once I have a good sentiment analysis framework in place, it would be nice to represent this data visually.
One idea is to graph the average comment sentiment by hour. The y-axis could range from -1 - 1, where -1 is the most negative and 1 is the most positive. The x-axis could range from 0 - N, where N is the number of hours the video has been live. The graph would then visually show how the average comment sentiment has changed hour by hour since the video was posted.
It would be interesting to see if there are any notable spikes in sentiment and where they might occur. I suspect I'll see spikes in sentiment close to when the video goes live, since YouTube subscribers generally get notified when the author posts. But if sentiment spikes occur later on, it could indicate some outside influence (i.e. negative feedback because of a scandal, coordinated traffic such as from an outside Internet forum, or even a bot campaign).
Average sentiment may also be represented as a separate slider, showing the video 'leaning' positive or negative by some percentage.