Farm Market: A Review of Digital Agricultural Marketing, Crop-Price Prediction and Transparent Offer Scoring with an Implemented Farmer Decision-Support Prototype

Small farmers frequently make selling decisions without reliable information about market prices, buyer dependability or transport costs. Digital marketplaces and machine-learning models may help address this information gap. This review examines 14 published studies from 2007–2025 and one government-commissioned report covering mobile-phone market information, India’s electronic National Agriculture Market (e-NAM), and statistical, machine-learning and deep-learning approaches to crop-price forecasting. Forecasting results vary with the commodity, prediction horizon, dataset and evaluation method; in some cases, simple baselines perform competitively. A single score therefore cannot establish whether a model suits a particular mandi. The paper also analyses the supplied FarmMarket project: a React/TypeScript interface with farmer, dealer, transporter and administrator workflows, a five-factor dealer-offer score, a net-return calculation, a rule-based crop-price estimator, a PHP/MySQL API and a scikit-learn ridge-regression script. The code contains these components, but the front end currently stores data in browser localStorage and is not connected to the API. The front end and back end also use different scoring formulas, while the price model relies on a 21-row demonstration dataset. The quoted accuracy figures cannot therefore be independently verified, and neither predictive performance nor practical benefit has yet been established.