JD Farm Fresh Redefining horticulture | food delivery

Supply chain OLD vs NEW
Admin deployment to manage fleet of vehicles + quality control

We didn't build your typical boiler-plate hack. We actually went and solved (YES SOLVED) an actual - real life problem that affects millions of lives in India and abroad. Our only hope is that this problem is recognised more, and something is down to solve it. Our solution is & problem are primarily targeted at the Indian market.

The problem | the inspiration

Currently in the farming industry, both the farmer and the consumer are in a loose-loose situation - this is due to the sheer number of middlemen who exist between the food being grown, and you buying it, which leads to a hike in food prices, yet the farmer only gets a small percentage of the final value thus, both are in a loose loose situation. THIS IS AN INCREDIBLY COMPLICATED PROBLEM, and one needs to understand the inner workings of this sector in order to solve this problem effectively, which is why one of the team members spent one month, going to different farms in India, purly to understand the supply and value chain surrounding this market - ideas which we have incorporated deeply into our solution, over the last 24h

Solution

our solution is simple - cut the middle men from the equation and sell directly to the consumer. Farm Fresh is a platform, where consumers can buy their groceries directly from the farmer . MAKE NO MISTAKE, WE ARE NOT YOUR JOE AVERAGE E-COMMERCE SITE!! This is because of our deep integration with our custom predictive algorithms, using machine learning and neural networks.

See the thing is, food, is not like any other product , where supply and demand can simply be predicted, because food takes into account factors suck as:

temperature , rainfall , humidity ,type of fruit/vegetable , historical data , light intensity, among others

and all this affects the quality, yield, and price of the food. This is why we have developed algorithms that take these factors into account, to predict the supply, and price of crops, weeks in advance, such that it matches closely with supply and demand of the day it is delivered, rather than ordered.

Additional (and crucially ) Farm Fresh has also deployed independent solutions for 4 separate platforms, for points along the supply chain, and quality monitoring: Farmer platform, Customer order site, Transport services app, and admin monitoring. (refer to images).

small hardware data logging boxes will be placed in vans/ trucks to ensure quality, especially in refrigerated vans, to keep independent contractors in check .

what we are offering, is a complete range of solutions to deploy on all platforms. However, the ** backbone of our solution are our algorithms, which define the predictions **.

Benefits:

Allows farmers to relatively constant prices
Consumers get significantly better quality for the same price
Food quality maintained and tracked
ML and AI used to increase accuracy of the algorithm

How I built it

We created realistic data about different factors that could affect the growth of crops and trained a machine learning model on the data to predict the growth of plants. Therefore, by inputting the number of seeds planted and the weather conditions the model is able to predict how many vegetables are expected to be produced. We also wrote an algorithm to help calculate the price to charge consumers based on the supply of different vegetables.

In addition, an intricate website was established to guide both farmers and consumers to input farm data and easily place food orders respectively. This was primarily based on flask in python, with the use of HTML, JS and CSS.

Challenges I ran into

Other than the lack of sleep, and the coooold weather yesterday, we initially had a lot of trouble with out business model, which we then ironed out, in the next couple of hours. We also had issues with our first ML algorithm, which was outputting data in the wrong format.

process to order:

customer places order for a set amount of veggies to be received the following week 1.1 algorithm predicts price based on a predicted supply to show the customer
veggies shipped from warehouse to customer

Accomplishments that I'm proud of

As a team, we each had vastly different skillsets, so mostly we are just very happy to have learnt , and expanded our skillsets, because that is ultimately the point of a hackathon .