High-Level Project Summary
With the help of TensorFlow and Keras, we created our model to describe how important is to get vaccinated and what are the benefits of getting vaccinated. In this model, we used NumPy, pandas, Min_max scaler, LSTM, Dense and Plotly, and Matplotlib for the plotting of data.The model we trained used data from external sources like Kaggle.The state-wise covid testing dataset was taken from Kaggle which was analyzed and active cases details of all states were separated for making a state-specific model rather than a country-specific model. That model connect through android app can help user to prevent hight covid risk locations.
Link to Project "Demo"
Link to Final Project
Detailed Project Description
In this project, we created an android application that predicts the risk of COVID-19 at a particular location. This app connects to the ML model API that takes location as a parameter and calculates risk according to the input.
We used Space Agency Data to visualize the various factors affecting the spread of the COVID-19 virus in a particular location. The resources provided by Nasa help us in better understanding the problem statement. However, the model we trained used data from external sources like Kaggle.
The state-wise COVID testing dataset was taken from Kaggle which was analyzed and active cases details of all states were separated for making a state-specific model rather than a country-specific model. A prophet model trained predicted for the next 60 days the value of the active cases and returned the risk level outside as low, moderate, or high.
Other factors like humidity, pressure, and temp can be taken into account to predict the risk of COVID-19.
We weren't able to deploy those models in API due to time restrictions. But we created a basic prototype that can be further improved by connecting the app to the ML model via the use of an API.
This app can be used as a constant indicator for the risk level that checks risk level regularly. This app may be able to prevent some COVID-19.
We used Flutter Framework for developing an android app that works on Dart Language. We also used TensorFlow and Keras and other various libraries in training our model.
Space Agency Data
We used Space Agency Data to visualize the various factors affecting the spread of the COVID-19 virus in a particular location. The resources provided by Nasa help us in better understanding the problem statement. However, the model we trained used data from external sources like Kaggle.
Hackathon Journey
Our experience in the Space Apps challenge was immensely pleasurable, enjoyable, knowledgeable, and educational at the same time. We learned a lot of things in the challenge, we concreted our understanding of machine learning models and implementing those models into a user-friendly application. Our biggest inspiration to have chosen this challenge is the deadly situation of covid and to contribute our part to the society through our previous and newly gained knowledge. Our major approach to this project was to spend a good amount of time in analysis and preparation, since a well-prepared start is half the work, and then use our understanding to solve the problems that rose, in a simple yet effective and efficient manner. All the members of our team worked hard, cooperated well, and thus we didn't face any major setbacks, only certain minor discrepancies that the team worked really well to solve together.
But we couldn't deploy our model to connect with our android application due to lack of time. We would like to thank each other in the team and most importantly to Nasa Space Apps for providing us with this great opportunity.
References
Study Covid-19 effect on pollution : https://www.kaggle.com/parulpandey/breathe-india-covid-19-effect-on-pollution
COVID-19 state-wise data : https://www.kaggle.com/sudalairajkumar/covid19-in-india
Flutter : https://github.com/flutter/flutter
Tags
#AndroidApp #MachineLearning