Project Description

This project aims at extracting insights from app review analysis. You can find the project demo on YouTube, the source code in this link, and the test demo website here. The extracted insights include:

• Salient topics, e.g., issue topics with significantly lower ratings;
• Abnormal topics, e.g., issue topics that experience a rapid rise in volume during a time period;
• Correlations between two topics;
• Causal factors to rating or review quantity changes.

Note: The test demo website will be outdated in August, 2019. After that, users can download the source code and build the tool at localhost.

Survey Results

Question	Answer	Data
How do you think about the comprehensibility of the insight summary generated by INFAR?	Very easy to understand Easy to understand Not easy to understand for some parts Difficult to understand	4 (33%) 3 (25%) 5 (42%) 0 (0%)
Without INFAR, how do you think about the difficulty of analyzing user reviews?	Highly difficult Difficult Not very difficult Easy	5 (42%) 7 (58%) 0 (0%) 0 (0%)
How much time do you think INFAR can save for user review analysis?	Over 50% 30%~50% 10%~30% 0	6 (50%) 3 (25%) 3 (25%) 0 (0%)
How do you think about the information given by INFAR?	Not containing any unnecessary information Containing some unnecessary information Containing many unnecessary information	4 (33%) 8 (67%) 0 (0%)
How do you think about the usefulness of the word cloud part?	Highly useful Useful Possibly useless Very useless	7 (58%) 4 (33%) 1 (8%) 0 (0%)
How do you think about the usefulness of the extracted salient topics?	Highly useful Useful Possibly useless Very useless	8 (67%) 3 (25%) 1 (8%) 0 (0%)
How do you think about the usefulness of the extracted abnormal topics?	Highly useful Useful Possibly useless Very useless	6 (50%) 4 (33%) 2 (17%) 0 (0%)
How do you think about the usefulness of the extracted causal factors?	Highly useful Useful Possibly useless Very useless	4 (33%) 4 (33%) 4 (33%) 0 (0%)
How do you think about the usefulness of the extracted correlated topics?	Highly useful Useful Possibly useless Very useless	5 (42%) 5 (42%) 2 (17%) 0 (0%)
How do you think about the usefulness of the whole summary provided by INFAR?	Highly useful Useful Possibly useless Very useless	5 (42%) 6 (50%) 1 (8%) 0 (0%)