KishanMitra

Learning Based Crop Prediction and Yield Calculation Project and API

Comment

Inspiration

Lack of knowledge and proper resources has always been a problem over all these years to our farmers. Not knowing what is best for the field, being totally unaware of what could happen in the future leads to thousands of acres of land going to waste every year. We are still in a development stage of Precision Agriculture, I am trying to implement it from the base itself. As sowing is one of the first steps after soil preparation, it is necessary to create a model for crop-related guidance, henceforth I present to you - Kishan Mitra

What it does

Kishan Mitra is a 3-Step Web Application, which allows you to perform two tasks -

  1. Predict what crop would be best for the piece of land you own, based on factors like NPK Values and Environmental Factors.
  2. Predict the Approximate Yield for a crop, based on Location and Type of Crop. Combined with an easy-to-use, versatile form-based website, the UI allows the user to enter the required values, and processes it at the back-end to produce the desired results.

How we built it

ARCHITECTURE

Front-end * Powered by HTML, CSS, Bootstrap and JS, we have a robust front-end design which allows anyone to use our models with ease. Combined with Django, we have Dynamic Drop-downs which only populate the data that actually exists in the data-set. **Back-end The Backend is coded in Python using Django. Django is a web framework that allows creation of complex website in a easy manner. The entire machine-learning part is done at the backend, where the data fetched from the user is put into the Algorithms and the result is processed and returned to the user. Our first service, the Crop Prediction is based on KNeighbors Classifier algorithm, and the second service i.e. the Yield Calculator is based on Decision Tree Regression, where both models work with 80-20 Train Test Split and produce around accuracy > 96% in the results. Expanding Usability - API We created our very own API for anyone to use our models in their applications. It allows developers to input NPK Values along with other Environmental Parameters and get the results as JSON in the body, which can be further used in various other projects.

Datasets used in the project - Crop Prediction (2200+ Entries) https://www.kaggle.com/datasets/atharvaingle/crop-recommendation-dataset Yield Calculator (600K+ Entries) https://www.kaggle.com/datasets/pyatakov/india-agriculture-crop-production

Datasets used in the project - Crop Prediction (2200+ Entries) : https://www.kaggle.com/datasets/atharvaingle/crop-recommendation-dataset

Yield Calculator (600K+ Entries) : https://www.kaggle.com/datasets/pyatakov/india-agriculture-crop-production

Technologies used in this project - Machine Learning Pandas, Numpy, Scikit-learn and Matplotlib Web Framework Django (Python) Design from Envato (Bootstrap5)

Accomplishments that we're proud of

The mission of our project is to make the farmer 100% confident about what decision he takes about his field, thus improving his net yield and income which is accomplished in such a short time.

What we learned

We learned the Team Work, Co-operation along with different Tech Stack experience and resolved bugs.

What's next for KishanMitra

Integrating APIs for a more Versatile Experience

This project has space for API Integrations in various steps of the form. Finding the location of the user using IP Geolocation, weather data of the region, and integrating with autofill, is something that can be done in the future to increase usability. Moreover, an API to detect crop season can also help save the time of the user when he visits the website.

Adding More Services Increasing the number of services the website provides, and bringing in other models as well to verify the integrity of the results can greatly highlight the web app, as it would be an ALL-AT-ONE PLACE model, being a farmer's best partner in any decision he wishes to take.

Revenue Generation through Paid Services The project can be also expanded by providing Paid Services to the farmers. This is an addition that is sought to add more value to the project. As cloud costs come into action to keep the versatility intact, this is something that needs to be implemented in due time to keep the project running.

Built With

Share this project:

Updates