apptrace blog

adeven gets to the heart of 24 million AppStore reviews

Just launched, apptrace’s sentiment analysis reveals which apps in the App Store are the most liked, the most addictive and which ones are the biggest crashers.

apptrace takes 24 million English based Apple AppStore reviews and gives a reviewer centred analysis showing what characteristics the app has and what type of reviews the app has in the AppStore. Each app with enough reviews is then given a rating for positive feedback, addictiveness and crash. Changes in sentiment over time are graphed for each individual app, giving a new dimension to understanding app performance. Just click on the sentiment analysis tab of an individual app to see the graphs.

So what magic is weaved and how does it work?

apptrace breaks each app review down into the individual sentences and then our algorithm looks at grammatical patterns such as the key verb and the language around it. Our dictionary includes verb conjugations and tenses, and a range of language patterns, allowing it to be flexible while still the crunching data from every single review.

Ratings for each individual sentence in the review are combined and weighted to give an overall message score, which is tracked over time.

The best bit?

Yes, it is all still free (and will remain forever free)! Our aim is to give developers, just like the tech guys who work here at adeven, the possibility to gain insights into how their apps are performing and how their competitors are doing.

Read what Darrell Etherington says on Tech Crunch - Adeven’s new App Store sentiment analysis tool helps developers draw useful conclusions from app reviews

http://techcrunch.com/2012/12/04/adevens-new-app-store-sentiment-analysis-tool-helps-developers-draw-useful-conclusions-from-app-reviews

Get up to speed with Matt Henkes’ article on Developer Tech - Apptrace takes the pain out of app review insight with sentiment analysis

http://www.developer-tech.com/news/2012/dec/06/apptrace-takes-pain-out-app-review-insight-sentiment-analysis

Sentiments

Published

January 15, 2013 at 3.33 p.m.