Blogs

Data Science in Agriculture

With increasing population, decreasing land sizes for agriculture and climate change ( leading to erratic rainfall patterns) increasing agricultural yield per square feet is more important than ever.

The solution lies in using smart farming and data driven precision farming. This involves employing IoT, data science and other emerging technologies to increase revenue and efficiency. 

Let’s look at some of the applications of data science and analytics in agriculture. 

  1. Crop recommendation system which takes in soil profile, weather conditions, user demand, logistics, land size etc to select the most profitable and best fit crop to cultivate for that season are available and being used. Ensemble models are a best fit for these recommendation systems as there are a large number of factors and parameters from an even larger number of sources that need to be considered before giving the final recommendation.

Websites to get agricultural data: www.fao.orghttps://data.world/datasets/agriculturehttps://data.gov.in , https://cgiar.orghttps://icar.org.in

Satellite monitoring, IoT sensors on crops and tools, GPS systems that are integrated into the vehicles and tools, weather stations, data from neighboring farms etc together form some of the main sources of data. 

2. Using data from all these sources, there are now apps, websites and other tools that help farmers make better decisions. Eg: Cropin,SatSure, Amnex etc. A $500-million big data project has been launched as a coalition between national govts, FAO and the Bill and Melinda Gates Foundation. The end result of this project would be a statistically representative database with precise geographic information about rural farms that can be used to track trends and adjust government outreach efforts accordingly.

3. Disease and Pest Detection: There are companies (Eg: Saillog, Agrosmart etc) which use drones ( to take photos of the crops) and IoT sensors to determine if, when and how much pesticides must be used and if the crop has been hit by disease. Recently there was a kaggle competition to assess how healthy a plant is based on images and the data given. Here is another dataset where ML is used to classify a given image into the 4 disease categories that Cassava crops tend to fall under or a 5th category (healthy).

4) Predict yield: This is done so farmers can plan ahead accordingly. This IEEE paper uses Big data and modeling techniques to predict crop yield. IBM Watson’s Platform for agriculture has APIs to forecast yield for corn 2–3 months in advance, predicting risk of diseases and pests for Corn, and much more.

The global agritech market is booming and is growing at 18% CAGR. Big Data and IoT are rapidly gaining a stronghold in this industry and its full potential is yet to be unleashed.

Share:

More Posts

Supervised Learning AI solutions.

All rights reserved