4 ms·
Interesting, but after you said "we processed 7+ million app reviews" I was really expecting deeper insights. This is a super exciting data set and I hope you
by MasterScrat 10y ago
Interesting, but after you said "we processed 7+ million app reviews" I was really expecting deeper insights.
This is a super exciting data set and I hope you can get more out of it: is it actually worth it to prioritize the features people are asking for? how long does it take to recover from a drop, can this be modeled somehow? how come some apps receive inconsistent emotion? etc
- martensd 10y agoThanks for your comment, MasterScrat. These are good points, actually our results are preliminary. As we consider this topic to be helpful for developers, we aim to further analyze the data. I would be very happy to receive more replies like yours to come up with new ideas and different perspectives on how to look at the data, and to prioritize which questions to answer.
- jhurliman 10y ago- I assume sentiment and star ratings are strongly correlated, but are there cases where these deviate? - Is there any seasonality to review volume and/or sentiment (day of week, time of month/year/etc) that holds across many apps? - Are there interesting differences in sentiment mean+variance by app category? For example, are utility app reviews more polarized than games?
- MasterScrat 10y ago> - I assume sentiment and star ratings are strongly correlated, but are there cases where these deviate? > - Are there interesting differences in sentiment mean+variance by app category? For example, are utility app reviews more polarized than games? These are answered in the white paper: https://arxiv.org/pdf/1703.02256.pdf https://arxiv.org/pdf/1703.02256.pdf