Use XGBoost Algorithm to Predict the GDP of Chinese prefecture-level Cities using Nighttime Light Data
Date:
Description:
This project is primarily designed to gather research data for my working paper titled “Cheat and Get Promoted: The Political GDP Manipulation Cycle in China.” In our study, we emulate the methodology of Chen et al. (2021) in their paper “Chasing or cheating? Theory and evidence on China’s GDP manipulation” by using nighttime light data. We employ the XGBoost algorithm to calibrate the annual GDP data for China’s prefecture-level cities. Each file in the repository is detailed as follows:
- merged.csv: Contains the nightlight data alongside the official GDP figures for China.
- xgboost_new.ipynb: Utilizes XGBoost and LightGBM to predict the actual GDP in China.
Repositries:
Related with:
“Cheat and Get Promoted: The Political GDP Manipulation Cycle in China”