AshleyNyaboke/Kenya-Carbon-Fintech-Analysis

Kenya carbon fintech data model & ETL pipeline tracking results-based mobile micro-payouts vs. ecological impact. Features custom Power Query M-code ingestion and clean dimensional star-schema modeling in Power BI # Kenya Carbon FinTech & Agroforestry Analytics An analytics engineering and data modeling project focused on Results-Based Financing (RBF) for smallholder agroforestry initiatives in Kenya. This project processes transactional mobile money (M-Pesa) micro-payouts distributed to farmers, mapping financial deployment against ecological impact metrics like tree survival rates and carbon sequestration. ## Project Overview & Domain Context In climate finance and international development, sustainability projects are moving toward **Results-Based Financing**. Instead of paying strictly for upfront tree planting, local NGOs incentivize long-term canopy growth by tracking tree survival. This repository models a dataset of carbon credit payouts, capturing two distinct phases of environmental stewardship: 1. **Initial Planting Payouts:** Capital distributed when new seedlings are actively put into the ground. 2. **Maintenance & Survival Payouts:** Recurring micro-incentives distributed to farmers for keeping existing trees alive (where `Trees_Planted = 0` but an M-Pesa disbursement is logged). --- ## Data Architecture & Pipeline ### 1. ETL & Data Pipeline (`/scripts/etl_pipeline.m`) Built completely using **Power Query M-Language**, the ingestion pipeline handles data cleaning, transformation, and optimization at source: * **Business Logic Mapping:** Implemented conditional logical parsing to explicitly split transaction cohorts into `Initial Planting Payout` and `Maintenance Payout` based on survival milestone rules. * **Type Enforcement:** Explicitly declared strict data types across columns to maximize VertiPaq storage compression and enforce referential integrity. * **Dynamic Dimension Generation:** Programmed a dynamic temporal calendar master dimension (`Dim_Calendar`) using pure M-code to allow for future time-intelligence calculations (YoY growth, moving averages). ### 2. Star Schema Data Model The project architecture strictly adheres to a **Star Schema** di …