Read our paper in Nature Energy Machine learning model to project the impact of COVID-19 on US motor gasoline demand
During this COVID-19 pandemic, we have seen a significant drop in mobility and fuel demand. We take a machine learning approach to understand the impact of COVID-19 statistics and states policy on personal mobility. We also build a model to connect the mobility changes to fuel demand changes. We then take the projected COVID-19 infection case numbers from the epidemiology model as inputs to predict future trends of mobility and fuel demand.
This repo hosts the source code and raw projection of the Pandemic Oil Demand Analysis (PODA) model.
-
PODA Model source code:
/PODA_Model_Code
-
Raw projection data:
/fuel_demand_projections
-
Projection visulization:
/visulization_plotly
You can find interactive visulizations of the projection and model details at covid19-mobility.com and teem.ornl.gov.
- Shiqi Ou (Energy and Transportation Science Division, Oak Ridge National Laboratory)
- Xin He (Aramco Services Company)
- Weiqi Ji (Massachusetts Institute of Technology)
- Wei Chen (Michigan Department of Transportation)
- Lang Sui (Aramco Services Company)
- Yu Gan (Energy Systems Division, Argonne National Laboratory)
- Zifeng Lu (Energy Systems Division, Argonne National Laboratory)
- Zhenhong Lin (Energy and Transportation Science Division, Oak Ridge National Laboratory)
- Sili Deng (Massachusetts Institute of Technology)
- Steven Przesmitzki (Aramco Services Company)
- Jessey Bouchard (Aramco Services Company)
The authors are solely responsible for the views expressed in this study.
We would like to thank the helps from
- Youyang Gu on the COVID-19 pandemic model
- Shihao Wen on the data visulization of project website