Python Code for Forecasting Thailand Mobility During COVID-19

Aritath Siraphatwongkorn published Python code in 2026 for forecasting human mobility trends in Thailand during the COVID-19 pandemic. The code implements Facebook Prophet, ARIMA, and Feature Engineered XGBoost models to analyze Google Community Mobility Reports data across six location categories.

Use Cases

• Apply the provided Facebook Prophet code to forecast mobility trends for six location categories from Google's reports.
• Use the Feature Engineered XGBoost implementation to model relationships between mobility patterns and COVID-19 case numbers.
• Replicate the Granger Causality Test analysis to examine causality between mobility and case data across different lockdown phases.

Strengths

• Code implements three distinct forecasting models (Prophet, ARIMA, Feature Engineered XGBoost) for comparison.
• Focuses on six specific location categories from Google's mobility reports.
• Released under a permissive CC BY 4.0 license for reuse.

Limitations

• The dataset is a 13.0 KB text file containing code, not the underlying mobility or case data.
• No sample data or column descriptions are provided to understand the required input format.
• The code's performance is tied to the specific Thailand COVID-19 context and may not generalize directly.

Provenance

Source

figshare

Collection Method

Python code for implementing forecasting models on Google Community Mobility Reports data.

Time Range

COVID-19 pandemic period in Thailand.

Freshness

null

Geography

Thailand