MBTI Personality Predictor

A picture of our webpage
Our machine learning model

Inspiration

We were always interested in the Myers-Briggs Type Indicator, known as MBTI, that categorizes personalities into 16 different types. We wanted to explore the topics of machine learning and web app dev to build a project that would predict these types.

What it does

The user types in a short passage on our webpage; then the model will analyze it, and a list of personality traits will be returned, along with their predicted MBTI personality.

How we built it

For the machine learning portion, we built a TensorFlow Keras Sequential model that we trained on a dataset comprised of journal posts and MBTI types. For the web app development portion, the webpage was built primarily with JavaScript, styled with CSS and Handlebars, and hosted on Heroku.

Challenges we ran into

Although we were able to complete the web application and the machine learning model separately, we did not have enough time to integrate the two. Ideally, we would like to have the web application run the model and analyze the user's input, but we did not have enough time to explore this.

Accomplishments that we're proud of

We are proud of being able to accomplish the two parts of the project separately, with both being functional and presentable on their own. Building these entirely from scratch was difficult, but we were able to accomplish these and learn a lot along the way.

What we learned

We learned how to use a text classification model and how to construct a neural network. We also learned how to continuously re-fit and improve the model based on its performance. For the web app, we learned how to use Node, Materialize, and more in order to build the page and make it look polished.