PopCatstinate

the GUI

💡 Inspiration

Uhh, how can we combine cute cats and trying to survive McGill's exams? Pop Cat + Procrastination = PopCatstinate!

We thought that looking at cute stuff would improve performance in focus-heavy tasks, such as studying for exams. Seriously though, there's real legit research on this topic :O very very scientific paper yes

🐾 What it does

Pop Cat will try to predict whether you are procrastinating or not by analyzing your typical activities during specific times of the day (we know you're scrolling on TikTok until 3 a.m. lol).

Pop Cat will also give you helpful analytics about the websites you procrastinate the most on, and provide suggestions on how to improve your studying habits! Basically, it compares your past data to your current data to see whether you improved! Taking advice from a cute cat can be useful sometimes you know...

⚙️ How we built it

The data interpretation and analysis is done in Python and the GUI is done in JS, CSS and html. The server hosting is done with flask, deployed on a Microsoft Azure server.

We first retrieve the user history with Javascript, which is sent to our Azure server ran with flask to our backend.

We implemented word2vec, a natural language processing (NLP) algorithm to extract the meaning of the websites in the user's browsing history and a procrastination constant for each website, basically how likely you are procrastinating if you are on them.

We then used k-Nearest Neighbors (kNN), a machine learning algorithm to make predictions about the level of productivity based on the procrastination constants of the different visited websites at the different times. This algorithm gives us a prediction of the productivity of the user on a given hour or day, depending on his past history.

Finally, we organize the data from our previous algorithms and send them back to the user on their Chrome Extension. We then display our summary of user behaviour in a neat graph in our extension (visuals in html and CSS).

❓ Challenges we ran into

Trying to assign "productivity values" to commonly visited websites. There are way too many websites that people normally use and it's impossible to classify them all. Training the model was also difficult because there were a lot of attributes to take into account for each website. SO MANY issues with the server.

🏆 Accomplishments that we're proud of

The fact that Pop Cat and Procrastinate could make such a good pun was completely a coincidence. Very proud of that one heh ¬‿¬

🧠 What we learned

That drinking hot chocolate at 3 a.m. is the ultimate life hack to staying up for more than 24 hours. Very useful information for future applications :D

We learned how to separate tasks efficiently to maximise individual productivity.