Making a marketplace for farmers and customers to connect together

Why current marketplace fails

For now, in the agricultural sector in Sri Lanka has an issue about the local produce stores acting as middlemen getting an unfair share. A potential solution would be to make an online marketplace to connect both of them directly without any middlemen.

 But there were attempts to make an online marketplace that tried to replace the local produce stores, but nothing was a success. This is because only a marketplace will not solve deeper issues within this sector. 

Major issues seen in this sector are even with the marketplace is accessible, there will be instances of oversupplies and shortages. These causes both the farmers and customers to find an alternative system. 

Issue of not having the proper market insights and relying on middlemen-  Post-harvest losses amount to Rs. 18 billion a year: Study – The Island

Solution

To address those issues the app Agrimate was made, direct farmer-customer smart marketplace.

    We made the marketplace with a system to allow produce to be scheduled beforehand and that quantity will be assigned to each farmer, who must supply that weekly. This ensures that there won't be any issues for farmers when selling the harvests. They can just grow the produce at their own pace and after harvest they can sell to the scheduled customers.

But as the marketplace also has a feedback system, farmers have to maintain the quality of the produce, so they have regular customers.

For customers when searching for produce the app filters out considering the proximity, harvest quantity remaining and it will be listed from most to least feedback in the app.

When a customer selects a produce, it will stay in a pending state until the farmer confirms the availability and accepts order for delivery. After delivery both the farmer and the customer can confirm the transaction, then the customer will be able to leave feedback about the quality of the produce. 

For farmer's side the app shows the demand of each crop, for each week after the harvest time period (if tomato harvest time duration is 6 weeks, the scheduled demand for next 6-10 weeks will be shown) so farmers can decide themselves what to grow, preventing oversupply. By displaying the scheduled demand and comparing it with the actual supply, the marketplace maintains a balanced supply and demand system.


Also when a farmer gives the system about future harvests planting dates and quantity, they will get details about crop care precautions updated every week.

These precautions(made using GenAI) consider the crop type, the growth stage of the crop(is it in the germination stage, flowering stage etc.), next 7 days predicted whether (using an open-whether API) and previous year's supply(as if it is low it means the crop is currently growing in an off season) and  it gives customized precautions for each harvest, updated every week.

 This will solve issues of misinformation and farmers can get most up-to date info from this and the farmers can also get the maximized and best quality harvests.


As in the initial stage of the app, the farmers will have to deliver the produce by themselves, but the cost of delivery will be divided among the customers considering the path 
(made using the Greedy Nearest-Neighbor algorithm, which it always gives the nearest path without looking ahead so it gives the most cost-efficient path) and according to that the delivery fee is divided so there is not much of a financial burden to the customers.

 
The app also supports both Sinhala and English languages so to be more convenient farmers as most are not familiar with English

The Road Ahead 

So, for future implementations this system can be expanded to have a driver role as well so when there are many deliveries for drivers the delivery cost will be more divided between the customers to divide the burden among themselves.

Also, by collecting the marketplace data, there can be future implementations to train a model to predict correctly about the marketplace demands and have this marketplace to have a balance between supply and the demand.
Some factors for such a system could be.
  •  Initial year's supply (inversely proportional to demand as higher supply could reduce the demand) 
  •  Predicted whether (if predicted whether is unfavourable for a given crop, demand will be increased and vice versa)
  •  Marketplace activity (how many users search for a certain crop weekly) 
   and by training to the marketplace data, it should be made into a system with much accuracy.

 Source for crop planting details (such as harvest time, ideal conditions, area to quantity ratios etc) - HORDI – Crops – Department of Agriculture Sri lanka

Sources for previous year's supply quantities (Tomato, Okra, Bean) - https://doa.gov.lk/sepc-downloads_en/#1591938875068-80aa5386-6e36

Repository link for Agrimate - Chanuth412-pixel/agrimate

Comments