I am starting to experiment with audio data (recorded calls from call centers) beyond text analytics and curious about what others are doing in this area. Has anyone done any work with deriving variables from MP3 files and used them as independent variables in models? There are very simple variables, i.e. length of the call, but I am curious if anyone in this forum has done work creating variables from soundwaves, etc.?
In my mind this would be similar to music matching algorithms like Shazam. Is anyone doing this today, and what tools are you using to quantify audio files?
Take a look at http://clarify.io/
I'm not affiliated, just seen them present at a conference, looks like an interesting solution for audio
Thanks Ali. I just took a look and it is an interesting solution. I will spend more time on the website this evening.
I know there are a lot of companies who are working with recorded calls. Let me broaden the question a bit. How are you leveraging recorded call data to improve your business? Approaches, tools, etc.. Thanks for contributing.