COVID Probability Checker

A Django And ML Based Covid 19 App for checking infection probablity

Comment

Inspiration

The COVID-19 virus has changed the lives of all. There is no cure, the only way we can keep ourselves safe is by taking precautions. So, I made an ML model (with Logistic regression) to predict the possibility of the infection of COVID-19 and a Covid Tracker App used to display statistics of different countries

What it does

it's a machine learning-based model to predict the possibility of the infection of COVID-19. so our model asks users about the important symptoms declared by the WHO in their guidelines and based on the response, it predicts the possibility and i have also created a Covid Tracker App used to display statistics of different countries by searching for a specific country .

How I built it

In this project, I used a logistic regression model to train the dataset available on the Kaggle. So logistic regression is used when the dependent variable(target) is categorical. For example to predict whether a person is suffering from a breathing problem or not we can just simply say, if a person is suffering from a breathing problem then we can assign a value 1 to it and if it's not then we can assign a value 0 to it, So After tuning all the parameters, when a user submits the form with all required information then that information is sent to the backend in the form of 0's and 1's and after that our trained ML model predicts the infection probability with the help of predict_proba function of Logistic Regression.

For the Covid Tracker App which is used to display statistics of different countries by searching for a specific country I used this API for getting all the data related to that country in json format

For the frontend I used HTML, CSS, bootstrap and javascript, and the backend was developed on Django.

Challenges I ran into

So basically this was my first ML project that I have worked on... 1st Challenge: At first I used linear regression for this problem but by using linear regression there is a need for setting up a threshold value based on which classification can be done, for example, if the actual class is malignant, predicted continuous value of .4 and the threshold value around something .5, then the data point will be classified as not malignant which can lead to serious consequence in real-time, after that point I came to know that linear regression is not suitable for classification problem so instead of linear regression I used logistic regression where value strictly ranges form 0 to 1.

2nd Challenge: training the model and fine-tuning the parameters as initially, the model was not predicting anything

note : but still there is a need for better data set to predict correctly

Accomplishments that I'm proud of

I at least tried to complete the project with different implementation and data sets and done everything on my own, which makes me feel proud as this was my first ML project

What I learned

1.I really learned a lot during this project about how to set up a Django app and work with ML modules at the same time and how to split data sets for the training set and test set in ML. 2.How to fetch data from API and de-structure it to display it on the front end 3.Learned how to tune the parameters and data sets

What's next for COVID Probability Checker

As of now the accuracy for this project is around 55% so to make it better I will try to increase the accuracy for this project to get more efficient results, to make it more accurate more fine-tuning is required and good and correct data sets are also required. So I will try to conquer this problems next!

Share this project:

Updates