Inspiration
i have been feeling like people miscommunicate a lot recently, and part of the reason is because our shared understanding of words might be fundamentally different. also, i have analyzed personalities of people around me too much this semester; hence, a very small hackathon project that combines both...
What it does
this was a pretty open-ended exploratory style data science project: i used a mbti dataset from kaggle which had written text associated to the user's mbti to train word embeddings (and display them for some of the most common words). this in theory can show which words are viewed as more or less similar for different personality types
How we built it
much inspiration was taken from recent work done in Dr. Jessy Li's LIN 371 class! uses python packages like nltk and pandas, in jupyter notebook
Challenges we ran into
limitations of dataset: it was not a very even distribution of personality types, so i just ended up looking at NF types, which in hindsight probably was not the most distinguishing factor i could have chosen... also, this dataset only took comments from personalitycafe forum, which makes the content quite limited; i was not able to get access to another reddit dataset that looked interesting in time...
Accomplishments that we're proud of
actually being concentrated and grinding out this thing in like 4-5 hours...!
What we learned
i got more practice with the aforementioned libraries, and learned that ideation is pretty hard... and learned that i shouldn't be too stressed because everyone's skill level differs a lot
What's next for mbti word similarity analysis
expand on different datasets, implement into an app that analyzes your potential misunderstandings with other people through journal entries or written responses to questions
Log in or sign up for Devpost to join the conversation.