This research develops a comprehensive data-driven framework to enhance onion marketing decisions for farmers in Maharashtra, focusing on profit maximization through improved price forecasting, profitable market allocation, fraud detection and transportation logistics. The study combines time-series analysis, statistical modelling, operations research techniques, data visualization and building recommendation engine to address real-world challenges faced by farmers. The study uses AGMARKNET onion market data, containing arrival quantity and price information across multiple APMC markets in Maharashtra. The study employs a mixed-methods approach: quantitative analysis of over 72, 000 historical records from multiple APMC markets using linear regression for price sensitivity, SARIMAX for forecasting, and linear programming for optimization; complemented by a descriptive survey of 150 farmers (100 from Ahilyanagar/Ahmednagar district and 50 from Satara district) to capture real-world farmer's issues. achieving an average MAPE of 17. 73%, MAE of 188. 51 /quintal, and RMSE of 212. 95/quintal across 84 markets, evaluated via 4-week rolling back testing. Data visualization (bar plots, moving average charts) was employed to visualize the issues faced by farmers. Successful estimation of price sensitivity (a) in 101 markets revealing a generally negative relationship between arrivals and prices. Transportation costs were modelled using a tyre-based cost table, comparing Full Load Transport (FLT: cost per km) and Partial Load Transport (PLT: flat cost per ton). For each allocated market, the minimum-cost option was selected. Additionally, a transportation problem was solved using Operations Research techniques: the initial basic feasible solution was obtained via the Vogel's Approximation Method (VAM), followed by optimal solution through the Modified Distribution (MODI) method. This analysis calculated and proved that there is the net profit advantage of selling produce to distant high-price markets (e. g. , Pune and Mumbai) compared to local markets (e. g. , Ahilyanagar). The Critical Path Method (CPM) and Program Evaluation and Review Technique (PERT) is applied to model the onion supply chain timeline identifying critical activities and estimating project duration under uncertainty. Overall, all of the Descriptive Analytics, Diagnostic Analytics, Predictive Analytics and Prescriptive Analytics is done by using Python. Limitations include reliance on historical data without real-time external factors (weather, policy) and assumptions in capacity and cost models. Agriculture is highly dependent on market prices, transportation costs, and demand fluctuations across different regions. Farmers often sell their produce in nearby markets without analysing whether distant markets could provide better returns. This research proposes a data-driven framework to forecast agricultural commodity prices and optimize selling locations to maximize farmer profit.
Kulkarni et al. (Tue,) studied this question.