Resources | Crux

How Dr. Rachael Tatman Thinks About External Sentiment Analysis Data

Written by Chris Handy | Nov 22, 2022

An Interview with Dr. Rachael Tatman, Language Technology Educator

Recently, I had the opportunity to sit down with my previous colleague, Dr. Rachael Tatman, as part of Crux’s new educational content collection. We chatted about her experience with public data sources in the NLP and data science space, and she shared a great metaphor for sentiment analysis. 

 

Sentiment analysis is widely popular in the external data industry right now, both as a piece of ESG data and with the ever-popular social media platforms. Its use cases vary widely from understanding customers’ feelings about a brand to creating targeted ads based on the feelings a company derives from a user’s Tweets. 

Dr. Tatman doesn’t put sentiment analysis on a data pedestal–quite the opposite. She argues that it’s likely never the best option when it comes to external data sources and that it’s akin to wearing a mood ring. “I think it can be part of a process that works, but I think that it's like a mood ring, right? I don't know if any of you have ever used a mood ring, but it sure does change color, and that can be fun to watch, but it doesn't necessarily actually reflect the mood of the person who is holding the mood ring.” 

And she’s right–mood rings can change color based on body temperature, blood pressure, or even the weather, and none of those data points are controlled by the wearer. It may be helpful in addition to another dataset to provide more color (pun intended), but on its own isn’t a fully reliable source for the sentiment. And that doesn’t account for the complexity of language, either. Twitter specifically is widely known amongst users for its dry sarcasm and mountain of memes, and technology isn’t up to par with a trending sense of humor. 

Dr. Tatman also gave a solid piece of advice to anyone looking to dabble in data science–learn Sequel. Anyone touching data for any kind of project, education, or professional task will need to manipulate it, and SQL is an unavoidable–but critical–tool for data usage. 

To hear more about Dr. Tatman’s educational work, check out her biweekly live streams on YouTube, follow her on Twitter, or check out her website.

To watch or listen to the full interview, watch the video on the XDU YouTube channel. 

Have an idea for an expert interview or interactive session? Email scout.scholes@cruxdata.com with your thoughts.