Inspiration
We were inspired to determine what could happen if you find yourself suffering from a variety of diseases and how to find the correlations between those diseases for possible correlative links.
What it does
Our model is able to take in a variety of possible diseases, illnesses, conditions, or diagnoses and show the most relevant other conditions as well as where that condition is most prevalent.
How I built it
We built upon a very thorough github repo and worked through lots of EDA to find the correct statistics to take. Using those statistics of Z-score and R-value we were able to find correlations between conditions based on a regression performed between sets of z-scores between conditions.
Challenges I ran into
We had to backtrack quite a lot during our exploratory data analysis because we kept running into roadblocks for what we wanted to analyze. Being able to sit down and work on the model was actually the easier part of undertaking this project. Most of our time spent on this project revolved around thinking of what to do next and how to interpret what information we found.
Accomplishments that I'm proud of
We had a very clean and well kept codebase that interacted in a clean manner between jupyter notebooks and source code, allowing for easy importing of various robust functions to allow for quick developing across all of our computers.
What I learned
We learned how to critically think about what we needed to do, what kind of features would be helpful, and how to make a roadmap given a certain set of information and desires.
What's next for Correlation Between Prevalent Diseases; Alcoholism Focus
If we had more time to improve our model, we'd like to merge in additional data sources found in the given Science track dataset. We'd also like to do an even more thorough analysis of the data dictionary found on the data source webpage.
Built With
- infogram
- matplotlib
- numpy
- pandas
- plotly
- python
- scikit-learn
- seaborne
Log in or sign up for Devpost to join the conversation.