There are two files positive10k and negative10k that contains positive and negative reviews of some of the top apps of android app store. ID of the apps whose reviews are taken is in the file appsid.
This data set can be used to make projects that uses supervised machine learning algorithms and trained the algorithms with this data. Sentiment analysis algorithms requires data to train and then test the algorithm.
This project also conatins a simple python script for crawling the play store apps and string the reviews. So, anyone who wants to build their own dataset with their custom requirements can modify the appsid file and Androidapp_reviewscrawler.py files. To reduce the number of requests this crawler make only one request to get the reviews for each app(40 reviews).
For ex:
- Dataset for Communication apps only.
- More reviews etc ...