Description
This talk focuses on the methodology and intent of studying feedback form data using Python tools and libraries for natural language processing and machine learning analysis. I will discuss potential trouble areas of starting such a project from scratch from a developer’s perspective. This will include the type of analysis that might be helpful in discerning user experience and what analysis that you run, but might end up tossing out at the end due to lack of insight on your data. For instance, what is the value of running a K-Means cluster analysis and does it offer substantial actionable insights for textual content? And how much data do you need to pre-label for a training set for a Naïve Bayes Classification in order for it to be accurate? My intent is that people will walk away learning the basics of textual analysis and become motivated to help their users succeed in whatever tasks they are trying to accomplish through finding potential points of friction and even issues that spring up from changes in the design. This talk will be for developers or marketers who don’t have a lot or any experience in data analysis or machine learning.