top of page

#WorldInnovationDay Hack 2021 Winner Showcase: Team Berkeley Data Science #1

This is the final project in a six-part series which will showcase the projects built by the Top 6 teams of the #WorldInnovationDay Hackathon. This was the biggest hackathon by Hackmakers in collaboration with the UN Environmental Programme, dedicated to developing innovative and creative solutions for the Sustainable Development Goals established by the UN.

​Team Berkeley Data Science


Kevin Hartman, Shaji K Kunjumohan, Kevin Hanna, Vaishnavi Rajagopal, Christina Chen, Padma Sridhar, Qian (Cathy) Deng

​Team Berkeley Data Science created an application called Carby to address the UN Sustainability Goal 8(4). Carby, an easy-to-use tool which can augment decision making process by working seamlessly with consumer behaviour and life integration. The team secured the 1st position among all the participants in the hackathon.


Challenge Addressed

Decent Work and Economic Growth – Sustainable Development Goal #8

The UN Sustainable Development Goal of Decent Work and Economic Growth promotes inclusive and sustainable economic growth, full and productive employment and decent work for all. The goal aims at ensuring the economic sector of every country provides the necessary need for its citizens to have a good life irrespective of their background, race, culture or other extraneous circumstances.


Problem Statement

The current state is that 65% of the consumers want to buy from purpose-driven brands that advocate sustainability, but out of which only 26% end up doing so.

Meanwhile, the average carbon footprint per person in the US is 16 times more than what it should be.



Carby has been developed based primarily on customer segmentation. The customers of the application are environmentally conscious.

70% of the consumers in the US and Canada think that it is important that a brand is sustainable and eco-friendly.

Climate conscious: Gen Z grew up with climate change at their forefront, and now they’re well conscious to make good buying decisions on their own, educate others on this scenario and make a change.



Carby, a concept based application that aims to augment the decision making process with an easy to use tool that works seamlessly with consumer behaviour and life integration.

The Carby mobile app is used to take a photo of a product for identification. With the photo, details of the product are made available to the consumer that allows them to make a more green buying decision by doing their part in going for products with lesser carbon footprint.


Featured Highlights

  • During the visual demonstration of the application, it is show how the app is successfully able to identify the carbon footprints from each of the products available

  • Apart from identifying the carbon footprint, Carby also suggests alternate options to the consumers of products which has lower carbon footprint

  • The Product Detection Model used has been trained and validated with a 100% accuracy

  • The current data being used is limited to French Food Data


What does the Application do?

  • Using the photographs of the items, Carby can successfully identify the carbon footprint of the products that a consumer is planning to opt for

  • For detection of objects, the app has its own system design specifically for its mobile client end. The system is fed with algorithms to identify the object

  • A photo of the object is captured and sent to the system’s cloud service in the next level where advanced machine learning algorithm identifies the product and suggests a more carbon friendly option, if the score is too high

  • The product detection model was trained on augmented images and validated on images provided by open food facts where it achieved 100% accuracy


Tools Used

  • Docker & Ubuntu – Micro Services

  • Python & Flask – API Development

  • Flutter – Mobile Development

  • Tensorflow Lite – Client side Object Detection

  • Bootstrap, HTML, Javascript, CSS – Web App Development

  • Yolo & Pytorch – Model Training

  • Roboflow – Data Augmentation

  • Google & Bing Search – Finding Additional Image Example

Future Vision

  • The team is planning to scale its data in terms of the number and types of products

  • For training such large amount of data, labeling more products or developing a proxy solution of carbon footprint score could be an option

  • Introduce a ‘Search’ feature to the application, and also add the feature of showing the ‘Nutritional Value’ of the product to the consumer in order to tackle the UN Sustainable Development Goal 3(d)

  • Partnering with public and private sectors to drive user adoption

  • Finalizing Revenue Stream – Use a certain proportion of sales from suggested better carbon footprint products

What did you think of this project? Do share your thoughts and ideas with us on Hackmakers LinkedIn Page.

Follow us for more interesting and innovative projects!

Don’t forget to check out our Previous Winner Showcase Blogs for the Smart Cities Hackathon 2021:

  1. Team Sign Hand – 2nd Position in World Innovation Day Hackathon 2021

  2. Team Swasthya – 3rd Position in World Innovation Day Hackathon 2021

  3. Team Edu-Fx – 4th Position in World Innovation Day Hackathon 2021

14 views0 comments


bottom of page