aiHackCovid
This is team 2 submission for the aiHackCovid hackathon by Bharath Raj, Rossella and Thanasis.
Our Goals:
- Explore the policies adopted by the different European countries, trying to highlight similarities and patterns among them
- Find trends and projections of new_deaths_per_million population in the Global scenario, EU continent and the Netherlands.
Our datasets:
- Oxford Covid-19 Government Response Tracker (OxCGRT) :
- COVID-19 Dataset by Our World in Data (OWID):
To get started:
- install the packages from the requirements.txt
Goal 1: Explore the policies adopted by the different European countries, trying to highlight similarities and patterns among them
We use the OxCGRT dataset for this.
- Visualize the data
- Compute Correlation Matrices
- Observe and Interpret the results
More detailed here
Goal 2: Find trends and projections of new_deaths_per_million population in the Global scenario, EU continent and the Netherlands
We use the OurWorldinData(OWID) dataset We gather the dataset over the entire world, create a dataset of our own by selecting feature vectors :
- Vaccination_ratio : It is given as a the ratio between the number of people who have been vaccinated(atleast 1 dose) and the total population (in 2020) before the pandemic
- Stringency_score : This is a score given by OWID based on the stringency measures taken in the location by closing of schools, public transport etc. This score is a measure between 0 to 100.
- positive rate : The share of COVID-19 tests that are positive given by OWID.
- current reproduction rate : This is the R_0 value, given by OWID.
Based on the above features we predict the new_deaths_per_million in our prediction models.
- We do the prediction models for Global scenario Global_Projection Model.ipynb
- We do the prediction models for Europe scenario EU_Projection Model.ipynb
- We do the prediction models for NL scenario NL_Projection Model.ipynb
We create a synthetic dataset for forecasting the new_deaths_per_million using the above models on all the three regions Creating_synthetic_dataForProjection.ipnyb
- We visualize the projections until March 1st 2022 on all the three models and observe their behaviour [Visualisation_and_inference.ipnyb]