Contrasts the data from multiple nations to obtain the best way to respond against the pandemic. Taking what each nation is doing best and learning from mistakes provides a set of measures to combat COVID-19. The idea behind this is cooperation. Since some nations are excelling at their response while others are failing, we are trying to compile the best measures from each country and learn from their worst to create a framework of factors that increase coronavirus contagion as well as general parameters that should benefit most nations facing the pandemic.
My team is really passionate about the idea of using data for good. We came to the conclusion that we could have an appropriate impact on this challenge. We knew that we were going to due code and AI to address this issue, but we debated upon specifics like how should we visualize our data and what AI models will we use. We decided to used satellite imagery as well as datasets from multiple US universities to create this project. we used python to code this along with a ton of libraries that are not worth listing here. We came upon the issue that there was too much data for our 2015 laptops to process, so we used some free cloud computing to train some of the algorithms. Much of our predictions are done with Support Vector Regression (SVR) and Artificial Neural Networks (ANN). One problem that we faced was that we used Convolutional Neural Networks to analyze the satellite imagery but this was not able to provide the results of interpreting the images, so we switched to the You Only Look Once algorithm also known as (YOLO) for our computer vision requirements. We programmed all of this in the Spyder text Editor distributed by Anaconda.
https://docs.google.com/presentation/d/1tVbnbcvRaNascGU2-e_84gZ3Pdkc0nxLz0vAmEY5KfA/edit?usp=sharing
Novel Corona Virus 2019 Dataset /dashboard by John Hopkins University - https://www.kaggle.com/sudalairajkumar/novel-corona-virus-2019-dataset/data
Satellite imagery from Earthdata-https://earthdata.nasa.gov/