Introspect started out as a data collection tool for a self-guided research project that attempted to find the cause of the decline in student mental health during the COVID lockdown.
It employed machine learning algorithms to find correlations between words used when describing one's day and the overall rating for that day. The algorithms were trained for a few weeks with responses from students at my high school. After that, they were able to predict the rating of a day accurately (about 87% of the time) based on the description of that day.
Using this data, I was able to find certain words and phrases that often correlated with negative/positive days. Here are a few of the results:
- Words such as "club" and "walk" that indicated socialization and outdoors activity were heavily correlated with higher rated days.
- On the other hand, words similar to "stuck" and "alone" correlated with lower rated days.
- For almost every student surveyed, weekends and holidays were rated lower than school days.
- Students that used words/phrases that indicated busy schedules generally rated their days much higher than other students.
- ...
After the conclusion of the research project, I modified Introspect to make it a Chrome/Firefox extension that allows users to rate, describe, and reflect on their daily lives. Here's how you can use Introspect:
- Rate, describe, and reflect on your daily life
- Gain insight into a period of your life with informative statistics
- View and edit any previous response
- Export your data as CSV
- Customize your experience with dark & light themes
Introspect requires NO permissions. It doesn't access your browser tabs, history, identity, etc.
Introspect collects nothing but your responses. All data is encrypted using the Advanced Encryption Standard (AES).
Contributions of any kind are welcome!
You can contribute in two ways:
- Open a new issue for a bug report or feature request in the GitHub repository
- Create a Pull Request that fixes/enhances some part of the extension
- Check out first contributions if you are new to contributing
- Follow the wiki page on building & viewing Introspect locally