It was the first day of college when I was having dinner in the mess when I noticed the amount of food that was being thrown out. Being curious I researched a bit about it and came to know about the massive problem of food waste. After some time when i had to go buy some vegetables, i asked the sabsi-wale bhaiya if at all any food was wasted, and I was shocked to hear when he said he had to separately pay for a rickshaw to throw roughly 10-15kg's daily. i though there obviously must be someone solving this that the vendor didnt know about but there wasnt anything out there. So I did some talked to more vendors to get more clarity and after some brainstorming landed on this beautiful idea.
What it does
Reduco uses cutting edge technology to predict the quantity of vegetables a vendor would need for the following day. After that if there are still some items left that are currently fresh but might go to waste in a few hours we upload it on our app where users can buy them at discounted prices.
How we built it
We use computer vision to detect which vegetables and how much of it is left. This data is added to our already existing data-base and merged with other factors such as weather, no. of customers, shelf life among others to predict the amount of produce he needs to buy.
Challenges we ran into
I only knew very basic level programming and learned to code basic OpenCV programs besides this we had no idea on the business aspect of things and what goes into going from idea to product Also in order to make efficient predictions we need a large data set which is a problem we are still facing. there are more than 50 vegetables and fruits and to train the computer vision model we need a varaity and a large amount of pictures. Also we are working with vendors who send us daily reports of various data points.
Accomplishments that we're proud of
made vegetable recogniser and basic data analysing model, made 3-d model of product in mind
What we learned
Our learning curve grew exponentially, read 2 books( 100$ startup, zero to one), learned python(mit ocw) and opencv( udacity), also learned using Blender to make 3-d prootypes besides this we got to know what type o questions to ask to conduct proper market research( talked to 150 vendors)
What's next for PINE
within the next 60 days we plan to launch our MVP for free and after 45 days we plan to launch to do a proper launch with the app optimizedvand facial recognition features built in