Imagine the following scenario:
You're a restaurant branch, you run S3 on reviews you got from your customers, and you get a handful of interesting axes.
You would most likely want to query documents based on values on this axis to see for instance which reviews are negative in valence and talk about the food.
A good solution to this would be an interactive little app, where you can add or remove sliders over different topic axes, and the app would show documents that rank closest to the set values on the given axes.
All axis not added would be ignored.
I think this would be immensely useful so we should totally implement it.
Imagine the following scenario: You're a restaurant branch, you run S3 on reviews you got from your customers, and you get a handful of interesting axes. You would most likely want to query documents based on values on this axis to see for instance which reviews are negative in valence and talk about the food.
A good solution to this would be an interactive little app, where you can add or remove sliders over different topic axes, and the app would show documents that rank closest to the set values on the given axes. All axis not added would be ignored.
I think this would be immensely useful so we should totally implement it.