# NMIS Agricultural Market Price & Shock Detection Pipeline
An end-to-end data analysis and exploratory research pipeline built to process, clean, and analyze agricultural commodity market data from the **National Market Information System (NMIS)**.
This repository handles structural data cleaning, multi-tier price parsing (Retail, Wholesale, Producer), commodity-specific outlier handling, and prepares the dataset for time-series price shock detection and market volatility modeling.
---
## Repository Structure
```text
├── data/
│ ├── raw/ # Raw NMIS CSV data files
│ └── processed/ # Cleaned and standardized datasets
├── notebooks/
│ └── nmis_eda.ipynb # Exploratory Data Analysis & statistical workflows
├── src/
│ ├── __init__.py
│ ├── data_cleaning.py # In-place header mapping, data type casting, and zero-handling
│ └── visualization.py # Plotting utilities for market price trends
├── .gitignore # Ignored local files (data, bytecode, checkpoints)
├── README.md # Project documentation
└── requirements.txt # Project dependencies
```
The Data is collected by the Ethiopian Agricultural Transformation Institute(ATI), digital market team.