MethodologyFrom raw data to insight.
Overview of data sources, cleaning pipeline, analytical approach, and tools used in this project.
Methodology Roadmap
A structured workflow from raw conflict data to cleaned datasets, analytical insights, and interactive visualizations.
Data Sources
Collect datasets from documented public sources.
Data Pipeline
Organize raw, cleaned, and BI-ready files.
Clean & Transform
Standardize fields, handle missing values, and derive variables.
Analyze
Explore trends, demographics, and comparative KPIs.
Visualize
Build charts and interactive Tableau dashboards.
Extract Insights
Present findings with transparent limitations.
1. Data Sources
Russia – Ukraine
- Uppsala Conflict Data Program (UCDP) – Georeferenced Event Dataset
- Our World in Data – War & Peace (aggregated conflict deaths)
Israel – Palestine
- Tech for Palestine – Gaza Daily Casualties
- Tech for Palestine – West Bank Daily Casualties
- Tech for Palestine – Killed in Gaza (individual demographic records)
2. Data Pipeline
data/raw/data/processed/data/bi_ready/3. Data Cleaning & Transformation
- Standardized date formats across all datasets
- Converted death counts to numeric types and handled missing values
- Filtered UCDP data to Russia-Ukraine related events only (45,830 events retained)
- Created derived fields: daily killed, age groups, conflict labels, location
- Generated monthly aggregates for Russia-Ukraine data
- Built a combined daily master file for Israel-Palestine (Gaza + West Bank)
- Cleaned demographic records (sex standardization + age group binning)
4. Analytical Approach
- ✓Exploratory Data Analysis (EDA) using Pandas, Matplotlib, and Seaborn
- ✓Time-series analysis of daily and monthly casualty trends
- ✓Demographic breakdown by age group and sex (Gaza)
- ✓Comparative KPIs between the two conflicts
- ✓Preparation of clean, analysis-ready files for BI tools
- ✓Interactive dashboard created in Tableau Public
5. Tools & Technologies
Limitations & Notes
Casualty figures, especially during active conflict periods, may be incomplete or revised later by the original data providers. This project treats all numbers as reported figures from the cited sources. No political interpretation or judgment is applied — the focus remains strictly on data patterns, trends, and demographic distribution.