Welcome to GirlGreens!
Girl Greens uses AI (via an interactive chatbot) to address food insecurity by utilizing natural language understanding (NLU) to connect users -- food insecure individuals, food pantry workers, community gardeners, and prospective volunteers -- with the important resources they need.
Inspiration
Finding solutions to food insecurity has been an ongoing endeavor, and in the midst of the pandemic, finding ways to feed America's families has quickly risen to the forefront.
Through our project, we aimed to create an empathetic platform to help our community work through this time of great uncertainty.
The amount of resources being donated to food banks has decreased, but the community need has not. Due to COVID-19, food banks have seen demand increase by 60% on average, and we see the impact of this in the places we live in -- Ohio's Butler and Cuyahoga counties.
In Cuyahoga County alone the food insecurity rate has increased 5% from 2018, meaning over 60,000 people are unsure as to where their next meal will come from. Butler County had a food insecurity rate of 12.4% in 2018-- nearly 50,000 people...and these numbers are only rising as the pandemic rages on.
Source: Feeding America (2020), Data USA (2020)
How can we use tech to connect people to valuable resources and build stronger, more nourished communities during this difficult time in our nation?
With this question in mind, GirlGreens came to life.
What It Does
Girl Greens uses AI to address food insecurity by utilizing natural language understanding (NLU) to recognize input and produce personalized responses that connect our users with the important community resources they need. Access to these community gardens means access to a steady food source that both the food banks and local families can easily get to.
On our website, we use IBM Watson Assistant to make the user experience more efficient and personal, especially in this isolating, virtual era. Users can navigate GirlGreens operations and be matched to food banks and/or community gardens that are closest to them. We also feature a digital recipe book to provide families and food banks with quick, easy meal ideas that can be made with ingredients grown at the community gardens.
GirlGreens Helper Functionalities:
Trained to provide locations, hours, addresses, and additional resources for food banks/gardens according to user input-- helper is enabled to respond to these situations with information that best fits the user's needs
- Directs users to our virtual cookbook which is free to download-- the cookbook features numerous recipes, based off ingredients grown at community gardens and donated to local food pantries
- Recruit volunteers to start their own GirlGreens garden - growing produce for community members in need from their own homes-- produce goes to food pantries that are struggling to receive donations due to COVID-19
- Foster relationships between community gardens and food pantries as we match them depending on their location-- following the match, they are able to share resources (primarily produce)
How We Built It
All GirlGreens' information is hosted on a website. On the home page, users can start a dialogue with the GirlGreen Helper, a chatbot created with IBM Watson Assistant. Through the use of machine learning, the GirlGreens Helper has been trained with 12 Intents, 6 Entities, and 40 Dialog nodes to respond uniquely to each user input.
Our Challenges
Our original website template on Wix was not supporting our chatbot code like we expected it to. We had already began filling in our service information for the template and the chatbot was nearly finished. We were terrified: no experience with web development, running low on time, and battling Zoom fatigue. However, our resilience shined through our adversity. We began researching other sites we could use to embed our code and decided to host the website using Google's tools. We were able to demonstrate the GirlGreens Helper as we planned to, and discovered a newfound confidence in our computer science abilities.
Accomplishments We Are Proud Of
Finished a project in our first hackathon!
Developed a functioning chatbot despite limited experience in coding
Designed a website that shares our service
Created a scalable, multifaceted solution to tackle food insecurity in our home counties and throughout the US
What We Learned
How to leverage IBM Watson Assistant's Slot function which allows the GirlGreens Helper to receive user input and confirm correctness*
Debugging code and troubleshooting errors
UI/UX Principles, Calligraphy, Hackathon Ideation, Project Management + Planning through the awesome workshops
How to network online and connect with a great community of femme non-binary and female technologists!
What's Next?
The GirlGreens App
A mobile experience made to help food pantries monitor the amount of food in their facilities. Through the app, they could scan products and input them in a database
This database would be reviewed by those working at community gardens, who could then help restock the supply
Watering Sensors (powered by Raspberry Pi)
Installed in the grounds of community gardens to turn water pipes on and off
Boost sustainability in the local community
Expansion
- Expand our database and current service map to increase accessibility to users nationwide
Built With
- css
- google-docs
- google-sites
- html
- ibm-watson
- ibm-watson-assistant
- indesign
- javascript
Log in or sign up for Devpost to join the conversation.