Data Prep & Exploration

Data Cleaning & Preparation

The data cleaning process was straightforward due to the high quality of the source datasets. The primary tasks involved removing rows with missing values (NaNs), specifically where military data was unreported for early transition years. We also mapped SIPRI's country names to World Bank ISO-3 codes to enable merging and engineered a "Region" column to group the 15 republics for better visualization. No complex normalization was needed for this initial exploratory phase.

Data Sources:

• SIPRI Military Expenditure Database 

  • 1949-2024 military spending for USSR and all 15 republics post-1991

  • Direct link : https://www.sipri.org/sites/default/files/SIPRI-Milex-data-1949-2024_2.xlsx 

• World Bank API via WBGAPI Python package (API Source) 

  • GDP, GDP per capita, population for each former Soviet republic

  • Direct link :  https://pypi.org/project/wbgapi/

Data Before Cleaning:

Data after cleaning: 

Merging Datasets:

Cleaning after merge:

Link to code used: 

https://github.com/rileythejones/-CUBoulder-DS-CSCI-5612-Project/blob/main/ProjectModuleOne%20(1).ipynb