Data Analytics & Machine Learning

Understandingfood pricesthrough data.

A data-driven exploration of agricultural market prices, seasonal behavior, crop-level trends, and machine learning forecasts across more than a decade of historical observations.

AGMARKNET•Python•Pandas•Machine Learning
484,504
Market observations
01
2015–2025
Historical coverage
02
9
Forecasting crops
03
1.99%
Forecast MAPE
04

The project

From raw market recordsto meaningful insights.

This project follows agricultural market prices from raw observations through cleaning, exploratory analysis, feature engineering, and time-aware machine learning evaluation to understand historical behavior and forecast future prices.

01DATA ANALYSIS

Market Intelligence

Analyze long-term food price movements across crops, markets, states, seasons, and historical periods.

Explore→
02PATTERN DISCOVERY

Seasonal Patterns

Explore recurring monthly behavior, price volatility, crop-level trends, and changing agricultural market conditions.

Explore→
03MACHINE LEARNING

Price Forecasting

Use engineered time-series features and machine learning to forecast the following month's crop prices.

Explore→
01→
484,504
Raw market observations
02→
1,210
Monthly price observations
03→
58
Forecasting features
04→
243
Test observations

Data analysis

What does themarket data reveal?

The analysis examines crop-level price distributions, long-term movements, volatility, seasonality, market coverage, and unusual price observations across the available historical period.

Highest median price01
Lentil

Among the analyzed crops, cotton records the highest median monthly market price in the project's crop-level analysis.

Highest volatility02
Onion

Onion shows the highest measured price volatility in the project's crop-level analysis.

Machine learning

Forecastingthe next month.

Multiple forecasting approaches were evaluated using a chronological train-test split. The selected model was evaluated using MAE, RMSE, MAPE, SMAPE, and R².

Understand the methodology→
Forecast engine
Extra Trees
Evaluated
RMSE
₹134.46
MAE
₹78.74
MAPE
1.99%
R²
0.9941
Test
243
Features
58
Selection basisLowest test RMSE among evaluated models
Model performanceR² 0.9941

Methodology

A completeanalytical pipeline.

From raw records to validated forecasts, every stage of the analytical workflow contributes to the final model.

01↗

Data ingestion

Market price records

02↗

Data cleaning

Validation & anomaly handling

03↗

Feature engineering

Temporal & statistical features

04↗

Model evaluation

Chronological validation

Interactive dashboards

Explore theproject visually.

Explore the published Tableau and Power BI dashboards used to present the project's historical price analysis, seasonal patterns, and forecasting results.

Tableau dashboard

Homepage Visual Analytics

Open Tableau
Live Tableau report
100%
Scroll and touch-pan enabledResponsive default zoom on mobile · 100% on desktop
Power BI dashboard

Homepage Visual Analytics

Open Power BI
Live Power BI report
Scroll and touch-pan enabledResponsive report container