Skip to content
jahagirdar edited this page Sep 17, 2015 · 5 revisions

Agricultural sector consists of various part's. For local crop pattern an excellent predictive engine can be built based on the following data set's

  • Sale of seeds, fertilizers and pesticide
  • Soil fertility reports
  • Rainfall pattern

For a global look we will have to sample data from datasets made available by other organisations e.g.

We do not have data on the first two and the third item cannot be processed in isolation.

Excellent dataset is available for the place where transaction takes place between farmers and traders i.e. APMC market. at http://agmarknet.nic.in/ we find data related to each commodity that is grown across India and sold at the APMC Market.

e.g. for Fruits --> Papaya The following data is recorded for each APMC Market across India, on a daily basis.

Market Center Arrivals Unit of Arrivals Origin Variety Minimum Price Maximum Price Modal Price Unit of Price
Andhra Pradesh --- --- --- --- --- --- --- ---
Koduru NR Tonnes NR Papaya 600 650 625 Rs/Quintal
Siddavatam NR Tonnes NR Papaya 750 850 800 Rs/Quintal
Goa --- --- --- --- --- --- --- ---
Mapusa 1.2 Tonnes NR Papaya 700 800 750 Rs/Quintal

The website provides a basic descriptive analysis by reporting the weekly and monthly stats across state and district. It also has tieup's with a few agricultural universities to provide market intelligence, but the reports for these market intelligence are either missing or buried deep in some thesis by some students.

So what are the questions that can be asked about the agricultural sector?

  • What are the crop's grown in a particular area?
  • Are farmers moving away from certain crops?
  • Can we identify early indicators of crop failure and potential market price rise?
  • Where and when does a particular crop fetch the best price?
  • What species of plants are grown for a given crop(hybrid etc.) how has it changed over the years?
  • How long does it take for a new hybrid crop to swamp the market?
  • Can we have a delayed prediction model between rainfall and crop?

Clone this wiki locally