covid19-prediction

Models to predict covid-19 cases in India and USA

Report title: Covid-19 Forecasting with Vaccinations as a factor: the case of India and USA

Team Details

Team No: 60
Team Name: Vanadium
Team Members:

• Vishruth Veerendranath (PES1UG19CS577)
• Vibha Masti (PES1UG19CS565)
• Harshith Mohan Kumar (PES1UG19CS276)

Data sources

The data were sourced from the following sources:

1. Daily state-wise COVID-19 cases for India: COVID-19 Data Repository by the Center for Systems Science and Engineering (CSSE) at Johns Hopkins University

2. Daily state-wise COVID-19 vaccinations for India: COVID-19 India API - cowin_vaccine_data_statewise

3. Daily state-wise COVID-19 cases for USA: todo

4. Daily state-wise COVID-19 vaccination for USA: Data on COVID-19 (coronavirus) vaccinations by Our World in Data

Data cleaning

The data were cleaned in our eda repository and are stored under cleaned_datasets/.

The raw sourced data are store in raw_datasets/.

Model Comparision

Models Compared:

• ARIMA
• GARCH (ARIMA+GARCH)
• LSTM
• Stacked LSTM
• VAR
• VARMA

All the models are present in Python notebooks (.ipynb files) under the /models folder.
The following subfolders are present under /models:

• /lstm
• /uni_timeseries (ARIMA, ARIMA+GARCH, GARCH)
• /mult_timeseries (VAR, VARMA).

Docker

Development

To develop/maintain code use the following steps to setup your environment.

Docker Container

1. To build the docker dev image run the following command

docker-compose up

2. Next use the following command to start up the dev docker container.

docker run --gpus all -it --rm -p 8888:8888 -p 6006:6006 -v $PWD:/covid19-prediction covid19-prediction_dev

Jupyter Notebook

Once the container is up and running use the following code to launch jupyter notebooks.

jupyter notebook --ip 0.0.0.0 --no-browser --allow-root

Debugging Errors

1. error checking context: 'can't stat 'error checking context: 'can't stat '...error checking context: 'can't stat'

Solution

ls -a
sudo rm -r .Trash-0/
docker-compose up