Rainfall Impact on Commodity Pricing

"This case study showcases how Winjit’s Machine Learning algorithm was developed for a similar scenario, where with the given inputs of-month, year and rainfall, prices of different commodities were predicted. "

Introduction

Commodity losses in inundate quantities are suffered due to unforeseen changes in rainfall every year. However, if attention is paid by observing the rainfall patterns of previous years, a dedicated system can be designed to predict the price of various commodities with the given input of rainfall.

Customer

The client runs a chain of cold storages in India.The organization is known for its expertise in the domain. It caters to markets in all the states of India and takes call when to release the commodities in market and till when to put them on hold. Such an effort helps the organization to maximize its profits as well as those of the farmers. From past 2 year, the client’s profits were declining due to unpredicted rainfall. Hence, the client sought for a solution where by using the previous trends in rainfall, present and future calls could be made to get the maximum returns.

Requirement

To get aligned with the exact needs of the client, Winjit’s Machine Learning team identified the key requirements as:
Creation of a database, uploading previous rainfall patterns over years
Creating Price log of different commodities in different months of a year
Implementing regression technique for prediction
Training the system with pre-processed data