The problem
Farmers commonly sell at harvest when prices dip, then watch prices recover weeks later. Price information is scattered and rarely turned into a clear decision.
What to build
Create a service that forecasts short-term wholesale prices for a chosen crop and market, and recommends whether to sell now or store, accounting for storage cost and spoilage risk.
Success looks like
A backtest showing the advice would have beaten sell-at-harvest on at least 60% of past seasons, with every recommendation accompanied by a plain-language reason.
Suggested technologies
- Time-series forecasting
- Public market data
- Feature engineering
- Dashboards
- SMS delivery
An original, independent statement. It is not an official Smart India Hackathon problem statement, and Arventiq Labs is not affiliated with or endorsed by that programme. The scenario is illustrative and does not describe any specific organisation. Free to use with credit.