> We're running sentiment analysis to identify the emotional tone behind the mentions. Handling multiple mentions in a single sentence and filtering out things like questions that shouldn't count as an opinion requires some effort. In addition, a minimum sample size is required to get statistically relevant results.
That's great that you are taking that into account, but I'm not totally sure how to interpret the value bars. So I presume a high value means lots of positive engagement. But how does a lot of negative engagement show up? I think it's valuable to know when there is a lot of chatter about something and it's mostly negative. Or when it's controversial.
Maybe something like a statistical distribution graph or weighted color gradient, so you could tell at a glance the density/quality/depth of discussion, and also the distribution of sentiment.
That's great that you are taking that into account, but I'm not totally sure how to interpret the value bars. So I presume a high value means lots of positive engagement. But how does a lot of negative engagement show up? I think it's valuable to know when there is a lot of chatter about something and it's mostly negative. Or when it's controversial.
Maybe something like a statistical distribution graph or weighted color gradient, so you could tell at a glance the density/quality/depth of discussion, and also the distribution of sentiment.