I'm trying to cover the tickets in our internal support inbox and analyze the sentiments of initial messages to get a general score.
We use Help Scout, and at first their reporting system wouldn't export the body messages because the nature of the threaded conversations. I've written a Ruby script with this gem (https://github.com/hramos/helpscout) to export our conversations. Moving forward I'll run the script at the end of each month and append the deltas to a running sheet.
My question is, now that I have the data, what is the best way to complete sentiment analysis on our support tickets? I want to gauge a better idea of where users have a problem and improve the help resources for those particular topics. Benjamin Brandall analyzed tweets to find how the top Saas companies were using their support twitter channels. I'm looking for some articles to read up on, as I'm more versed to writing my own script to complete the sentiment over a service like Aylien's Text API (cost purposes).
We use Help Scout, and at first their reporting system wouldn't export the body messages because the nature of the threaded conversations. I've written a Ruby script with this gem (https://github.com/hramos/helpscout) to export our conversations. Moving forward I'll run the script at the end of each month and append the deltas to a running sheet.
My question is, now that I have the data, what is the best way to complete sentiment analysis on our support tickets? I want to gauge a better idea of where users have a problem and improve the help resources for those particular topics. Benjamin Brandall analyzed tweets to find how the top Saas companies were using their support twitter channels. I'm looking for some articles to read up on, as I'm more versed to writing my own script to complete the sentiment over a service like Aylien's Text API (cost purposes).