LibQUAL-Sentiment ANALYSIS WITH AlchemyAPI
Perform sentiment analysis on LibQUAL+ comments using AlchemyAPI
Open-ended respondent comments are a valuable complement to the regular quantitative findings provided by most surveys. While reading every comment provides insight for the investigator, more systematic analysis and summarization of extensive qualitative feedback is a time-consuming task. Recent advances in sentiment analysis, harnessing natural language processing, text analysis, computational linguistics and machine deep learning have opened a rich seam of possibilities for the automation of a range of hitherto, human-only tasks.
The AlchemyAPI conveniently exposes the necessary computing power, algorithms and data via well-defined interfaces to allow us explore these possibilities.
This project is a recipe to:
- Parse the textfile output of a set of LibQUAL+ comments (multiple files over multiple years are accommodated)
- Submit each comment to the AlchemyAPI for both document and keyword level sentiment analysis
- Store the LibQUAL+ data together with per comment, and per keyword sentiment scores in a SQlite database.
- Additionally store Sentiment polarity and keyword relevance.
- Provide a series of SQL statements and custom queries
- Visualize the data
Slideshare.net overview: http://www.slideshare.net/conulconference/peter-corrigan-wed14301515
comments.pl This program will perform document level sentiment analysis on your comments and store your results in Sqlite3 database for analysis. Dependencies:
- The SQlite database availalbe from https://www.sqlite.org/
- The SQLite DBI driver DBD::SQLite - install using cpan DBD::SQLite
- The AlchemyAPI Perl SDK available from AlchemyAPI.com (You'll need your own api key, also available from AlchemyAPI.com to run this)