E-ugine/Retail-Inventory

A data product that generates outlet-level FMCG demand intelligence for informal retail markets in Kenya. Built on publicly available data with a synthetic transaction layer, designed to demonstrate value to FMCG brand owners and distributors. # Retail Intelligence A data pipeline and analytics demo that currently(aiming to expand to other towns) generates outlet-level FMCG demand intelligence for informal retail markets in Kisumu, Kenya. Built on publicly available data with a synthetic transaction layer. --- ## What this is FMCG brand owners in Kenya can see distributor sell-in numbers, what left the warehouse, but not sell-out numbers, what left the duka shelf. This project infers that gap using spatial data, demographic consumption weights, and a simulated transaction layer, then surfaces the results in an interactive dashboard. The output is three intelligence modules: - **Sales velocity** — which product categories move fastest in which neighbourhoods - **Stockout probability** — which outlets are likely out of stock right now based on last delivery and depletion rate - **Coverage gaps** — which outlets are being underserved by distributor visit frequency The demo runs on synthetic data modelled on real Kisumu retail geography. The pipeline is designed so that real distributor invoice data can replace the synthetic layer without changing anything downstream. --- ## What this is not This is not a production system. There is no authentication, no multi-tenancy, no real-time ingestion. It is a functional demo built to validate a market hypothesis and open conversations with potential data partners. --- ## Stack | Layer | Technology | |---|---| | Spatial database | PostgreSQL 17 + PostGIS | | Data processing | Python 3.12, Pandas, GeoPandas | | Simulation | Custom discrete event simulator | | Dashboard | Streamlit, Plotly, Folium | | Hosted database | Supabase | | Deployment | Streamlit Cloud | --- ## Prerequisites - Python 3.10+ - PostgreSQL 17 with PostGIS extension - Git If you are on Mac with Homebrew: ```bash brew install postgresql@17 postgis brew services start postgresql@17 ``` --- ## Local setup **1. Clone the repository** (My bad on the clash of names. Should have been int …