(Photo by Flickr user Steve Harris, used under a Creative Commons license)
For all the yelling Frank and Estelle Costanza did on Seinfeld, the ever-present 1990s sitcom, they weren’t the angriest characters on the show, according to an algorithmic analysis from CompassRed’s Data Lab project.
Data scientist Daniel Larson had an algorithm scour through the 53,629 lines of text that make up the show’s 177 original episodes and sort them by sentiment (similar to this analysis of President Donald Trump’s tweets.)
The results from the analysis are broken down in a blogpost penned by Larson and Technical.ly Delaware alum Joey Davidson. A trio of interactive infographics lets users explore the data by character, match up season against season or see which of the main four characters had more lines through the years.
“The algorithm in this analysis moved through every spoken line in the show and scored the words used in a host of categories,” the post reads. “Each line might include words that the algorithm deemed indicative of fear, joy, trust, surprise or sadness. The line would also have an overall positive or negative emotional score.”
The project leveraged something called the NRC emotion lexicon (basically a list of words with associated emotions and sentiments) and an R package called tidytext to quantify the sentiment in each line of dialogue.
Yada, yada, yada: So who was the angrier character in the show’s nine-season run? None other than outrageous, egregious, preposterous attorney Jackie Chiles.
There might be a business application for the lil experiment, the duo said.
“Think of a data science team that moves through a dataset like, say, tweets about a given business or product with a similar algorithm in hand,” Larson and Davidson wrote. “Gleaning customer emotion from the huge dataset they create by using social media on a massive scale would offer intangible value to the decisions and movements of business.”
The 5 stages of the data scientist
The Philly revenue department’s data-driven strategy to upend city bureaucracy
Why diversity in artificial intelligence development matters
Why working with the University City Science Center was a game changer for 4 Philly startups
Is it legal? A quick look at 5 disruptive technologies — and the laws that make or break them
Data Jawn speakers will look at city budget, opioid crisis through the lens of data
Looks like Looker is stopping in Philly for its JOIN 2019 tour. Here’s why it matters
Take a peek at the opportunities popping up at PromptWorks
Sign-up for daily news updates from Technical.ly Philadelphia