Countries in the Asia Pacific like India, Bangladesh, Nepal, etc are agricultural countries, though the agro-industry has a total of 12-13% of the GDP share. More than 60% of rural India and Bangladesh depend on agricultural yield growth and allied agro-industry products. Crop yield prediction is an essential task for the decision-makers at national and regional levels for rapid decision-making. An accurate crop yield prediction model can help farmers to decide on what to grow and when to grow. It will act as a medium to provide the farmers with efficient information required to get high yields and thus maximize profits which in turn will reduce the suicide rates and lessen difficulties.
Hence our team came up with Project Farmiyo: Where farming is made easier :: Crop yield diagnostic tool!
The problem it solves:
- It is an important metric to understand because it helps us to understand food security.
- It has been demonstrated that an increase in crop yields significantly reduces poverty.
- Providing climate-smart agriculture to increase crop yield while facilitating the achievements of crop production in a safe environment will help to achieve the goal of the 2030 agenda of Sustainable Development of the United Nations in transforming our world formulated as end hunger, achieve food security, improve nutrition and promote sustainable development.
- Reduce farmers' suicide rates.
What it does
- Web application with integrated machine learning model developed in order to provide farmers an approximation on how much amount of crop will be produced depending upon the given input.
- This will help the farmers to know the crop yield in advance to plan and choose a crop that would give a better yield.
- The ML model aims to help farmers to cultivate proper crops for better yield production. To precisely predict the crop yield it analyzes factors like district (assuming the same weather and soil parameters in a particular district), state season, and crop type. It can be achieved using supervised and unsupervised learning algorithms.
- We used a random forest machine-learning algorithm. The chosen dataset was loaded and 75% of the dataset was used to train the model and 25% to test the model. The model achieved an accuracy of 89%.
How we built it
- Figma: After brainstorming many ideas, we moved on to wireframing in Figma, starting with lo-fidelity and working together to create clickable interactions on the high fidelity prototype.
- Machine Learning: implemented Random Forest group Algorithm using python libraries: Pandas, Matplotlib, Numpy, Sciketlearn.
- Front-end development using React.js
- Material-UI and CSS: we worked on the UI/UX, layout, CSS, and design.
Challenges we ran into
- We faced difficulty in testing the data sets.
- Data cleaning and gathering were time-consuming.
- Debugging issues and dynamic routing in React.js.
- Firebase authentication
Accomplishments that we're proud of
✅We are proud that we were able to address such an important issue and find a practical and inclusive solution to it.
✅Our teamwork and cooperative workflow helped us to build the project in its entirety.
✅We are proud to have completed the whole UI design and developed a fully functional website and an ML model.
What we learned
- The very first thing we all learned was teamwork. We thought about the various problems around us and came up with a solution that we could build in the given time.
- The second thing we could implement Machine Learning Algorithm for our project.
- We learned better tactics of collaboration and brainstorming. We also learned a lot of issues faced in the agro-industry while researching, that we were unaware of.
What's next for Family
- We want to build a mobile application since it is handier.
- We wish to integrate language translation tools so that farmers can use our applications without any linguistic barriers.