The notebook host the work done towards the paper: Bridging Policy, Infrastructure, and Innovation: A Causal and Predictive Analysis of Electric Vehicle Integration Across Africa, China, and the EU.
# ev-adoptability-analysis
**Causal and predictive analysis of electric vehicle integration across Africa, China, and the EU.**
Quantitative codebase supporting the paper:
> Nsan, N., Obi, C., & Etuk, E. (2025). Bridging policy, infrastructure, and innovation: a causal and predictive analysis of electric vehicle integration across Africa, China, and the EU. Sustainability, 17(12), 5449.*
---
## Overview
This repository contains the full analytical pipeline for a multi-method study of electric vehicle (EV) adoption feasibility across three distinct regulatory and infrastructure environments: Sub-Saharan Africa, China, and the European Union.
The study combines **machine learning prediction**, **econometric causal inference**, and **time series analysis** to answer two core questions:
1. **What drives EV adoption feasibility?** — Which policy, infrastructure, and energy mix variables most strongly predict EV market share?
2. **Does causality hold?** — Are those relationships causal (panel fixed effects) or merely associative?
---
## Methodology
The analysis proceeds in five stages:
### 1. Feature Engineering & Composite Feasibility Score
Raw panel data is loaded for each entity (country/region) across multiple years. Two derived features are constructed:
**Charger Station Density (CSD):**
```
CSD = Charging Infrastructures / Population Density (per km²) × 100,000
```
This normalises charger counts for population and geographic spread — a more meaningful metric than raw infrastructure counts.
**Subsidy encoding:**
The categorical `Subsidies and Tax Exemption` field is binary-encoded (Yes = 1, No = 0) for compatibility with regression models.
**PCA-based Feasibility Score:**
A composite EV readiness index is constructed from three indicators:
- `EVAR (%)` — EV adoption rate
- `CSD` — charger density (derived above)
- `RES (%)` — renewable energy share in the electricity mix
These are standardised and compressed to a single principal component via PCA, t …