Human mobility is thought to provide the best ground for proliferation of Covid-19, and as a result the focus of the project was understanding the best channel for virus spread.
The project used datasets that are relevant to human activity such as:
- Infrastructure Density
- Population Density
- Road Traffic Flow
- Air Traffic
-GDP Per Capita
as well as Covid-19 history concerning the United Kingdom.
The drive to innovate with the tools at our disposal pushed us to join efforts into tackling the Space Apps Covid 19 Challenge.
Starting with the data sources introduced through the Virtual BootCamp, e.g. Nasa GIBS, Planet Earth, we chose, as first step, sifting through this material to find what is most relevant to human activity out of all Satellite Imagery Layers available. We then proceeded to look for publicly available datasets that support the quantification of human mobility in the United Kingdom. This was followed by analysis of resulting datasets, focusing on evaluating the impact of each mobility metric on the evolution of the number of Covid 19 cases.
The tool of choice for all team members was Python. In order to do data manipulation pandas was employed along with seaborn , matplotlib and geopandas for visualization. Part of our toolkit was also numpy, scipy for mathematical manipulations, OpenCV and PIL for manipulations on image data and requests to interact with the Rest APIs.
The biggest challenge was matching the tiling of Satellite Imagey to the latitude and longitude coordinates so that they become compatible with the data format used everywhere else.
The greatest achievement was in finding the features that are most relevant to Covid-19 spread in the UK, which provide a foundation for advising policy makers.