Data Science and Google Mobility Reporting for Fuel Demand Modeling
Description
The recent global pandemic situation prompted the generation of open
mobility reports, an initiative created by Google in support of health policies
associated with COVID-19. The significant change in the flow of vehicles during
the pandemic situation and the variation in fuel consumption associated with
transportation and various productive activities required the creation of new
predictive models related to an unusual data set (e.g., geolocation of drivers).
Proper manipulation and analysis of these data provide a forecast that improves
the traditional forecast of fuel demand associated with actual consumption. In
this paper we analyze the databases of fuel sales (gasoline and diesel), available
and open in official websites and information of the company YPF. The results
show a positive correlation between the variables related to the demand for these
fuels and the Google mobility records, with some particularities. The programming language used for the development of the visualization code, geostatistics, predictive calculation and research reports is "R".