In 2019, a total of 463 million people are estimated to be living with diabetes, representing 9.3% of the global adult population (20–79 years). This number is expected to increase to 578 million (10.2%) in 2030 and 700 million (10.9%) in 2045.For getting a quick intuition about whether he/she has Diabetes, we have developed a website by which person will get a quick guess about whether he/she is non-diabetic, pre-diabetic, or diabetic after which he/she can visit the nearest pathology lab.
What it does
It basically gives a probability of whether the person is non-diabetic, pre-diabetic, or diabetic. Hence he/she can take necessary actions relevant to it.
How we built it
First of all, we built a machine learning model which gave the predictions using deep learning. In that model, we consider a total number of 21 parameters that are medically standard and relevant for our predictions. After that, we created webpages using React.js by which we got the inputs from the user. Consequently, we use Flask Framework for supporting the backend. In the Flask Framework, we used the .h5 file to provide access to the user input to our ml model so that it can predict the result and output be displayed accordingly. Eventually, we deployed our web app on Heroku.
Challenges we ran into
The first challenge we faced is to select a machine learning algorithm that would give us the highest accuracy within a practical time duration. Also we faced problems of converting JSON into different datatypes. We also faced some challenges during the deployment of the web app.
Accomplishments that we're proud of
We are very proud that we have finished this project in a very short period of time. Pred-Diabetes is able to give 85% accurate results whether the person is non-diabetic, pre-diabetic, or diabetic.
What we learned
Through this very project, we got a chance to know about the Diabetes epidemic and its depth. Also by making the website for this project we polished our skills in web development. Also, we gained the knowledge of making a good predictions using machine learning models. Besides, we learned about brand-new technologies and were glad to use them.
What's next for Pred-Diabetes
We will most likely maintain this website and will make the accuracy better for the predictions. Also in the future, we aim to make this website predict the possibilities of various other diseases and epidemics which also need an equal amount of social awareness.
Log in or sign up for Devpost to join the conversation.