Thailand COVID-19 Mobility Forecasting Code and Analysis

Python code implements a rolling-origin forecasting approach for Google Community Mobility Reports data in Thailand. The work applies Facebook Prophet, ARIMA, and Feature Engineered XGBoost models to forecast mobility trends across six location categories during the COVID-19 pandemic. The study includes a Granger Causality Test to examine relationships between mobility patterns and COVID-19 case numbers.

Use Cases

• Implement the rolling-origin forecasting methodology described in the study to evaluate time series models on mobility data.
• Replicate the Granger Causality Test analysis to assess relationships between mobility trends and COVID-19 case numbers.
• Apply the provided Feature Engineered XGBoost, Facebook Prophet, and ARIMA models to forecast trends for the six Google mobility location categories.
• Use the code as a template for forecasting other public health or mobility time series data with a rolling-origin validation scheme.

Strengths

• Code is associated with a published study that evaluated three distinct forecasting models (Prophet, ARIMA, XGBoost).
• Analysis is geographically focused on Thailand, providing a specific case study of pandemic mobility.
• Work incorporates a causal inference method (Granger Causality Test) alongside forecasting.
• Released under a permissive CC BY 4.0 license for reuse and modification.

Limitations

• The dataset is a 27.2 KB text file containing code, not the underlying mobility or case data, which must be sourced separately.
• The scope is limited to the methodology and analysis for a single country (Thailand) during a specific period.
• Code functionality and dependencies are not described in the provided input details.

Provenance

Source

figshare, authored by Aritath Siraphatwongkorn.

Collection Method

Python code implementing a rolling-origin forecasting study on Google Community Mobility Reports.

Time Range

Covers the COVID-19 pandemic period in Thailand.

Freshness

Last updated in March 2026, indicating recent maintenance.

Geography

Thailand.