Nakamoto-ctrl/sierra-leone-energy-transition

Sierra Leone’s energy transition confronts three concurrent deficits: infrastructural, finan- cial, and algorithmic. The first two are well-documented. The third—that AI planning tools, without explicit equity safeguards, systematically exclude the communities they are meant to serve—remains largely invisible in both research and policy. # AI, Big Data, and Sierra Leone's Energy Transition ### A Data-to-Decision Framework for Utility Reform and Distributed Electrification **Authors:** Jeremy Thoronka & Mohamed Alpha (2025) **Project:** *Cities: Smarter Futures* --- ## Overview This repository contains the complete analytical pipeline and manuscript source for the study on Sierra Leone's energy transition. The research introduces the **AI--Big Data Catalyst Model for Sierra Leone (AIBD-Cat-SL)**, a four-layer sociotechnical architecture designed to bridge the data-to-decision gap in LMIC energy systems. ### Key Contributions * **Double Blindspot Index (DBI):** Mapping the spatial co-location of energy deprivation and data invisibility (r = 0.972). * **Data Readiness Index (DRI):** A PCA-weighted assessment of district-level suitability for AI integration. * **Algorithmic Bias Audit:** The first LMIC demographic parity audit for electrification planning, demonstrating a 92% parity gap reduction through Layer 4 governance constraints. * **Monte Carlo Validation:** Robustness testing across 15,000+ simulations covering DBI stability, parity thresholds, and NTL revenue recovery. --- ## Repository Structure - `AIBD_Production_Notebook.ipynb`: The primary Jupyter Notebook containing the data processing, machine learning models (Random Forest, Isolation Forest), and Monte Carlo simulations. - `Cities_Smarter_Futures_Manuscript.tex`: The LaTeX source code for the 23-page manuscript. - `Results_and_Figures/`: A production-ready directory containing: - **14 High-Resolution Figures:** (.png) including 3 systemic framework diagrams. - **8 Analytical Datasets:** (.csv) including district-level DRI/DBI scores and MC results. - `references.bib`: Complete BibTeX database for the research. --- ## Getting Started ### Prerequisites The analysis is built using Python 3.12+ and requires the following libraries: ```bash pip install numpy pandas matplotlib seaborn scikit-learn geopandas scipy ``` ### Re …