Medical resources have always been hard to access for patients. And this is crucial for patients who need it the most. That is why we believe that this problem is interesting to tackle and needs to be solved as soon as possible. There has to be a better way to tackle this problem.
What it does
From the six months of individual observations we were given, our website gives access to every patient to showcase the correlations between the features and the predicted problems. This provides a more interactive way to connect between health providers and the patients.
How we built it
Using Tactio API, we were able to successfully (after many tries) obtain the observations needed to determine the disease of each patient.
Through online extensive research about the four conditions, a kaggle dataset, which we had to clean, with similar features but also labels as well as the given equations, we were able to set a definite threshold to determine the condition of each patient.
Challenges we ran into
Parsing through the data correctly was definitely a challenge! In fact, the approach we used to parse the data did not take into consideration all of the specific structures of the json file. Hence, after we set our thresholds, we had realized that we were missing parts of the data. Therefore, we had to go back on our tracks and fix the problem.
Accomplishments that we're proud of
We are proud to have learned so much from this Hackathon and from obtaining so much information even in the health field. We are also very proud to have finished this project in such a small amount of time.
What we learned
We have learned a lot about how to make post requests and get requests to the HTTP, as well as how to extract the data from the JSON file.
We also learned how to clean data given a huge dataset with different formatting for each feature as well as how to identify differences as well as making this difference interactive for everyone.
What's next for Patients Connected
We were very short on time and would have liked to add a mobile version of our website, we understand that this would have connected with the patients in a more modern way due to the accessibility of mobile devices.
At last, we would like to give credits to:
Itelina's Diabetes Prediction dataset: https://github.com/Itelina/DiabetesPrediction.git As well as our friend ChangHeng Mo who helped us tremendously through this process.