Inspiration
One of the biggest hurdles in finding the right pet for everyone is considering factors that can determine what kind of pets would be best suited for each person.
We realized that while various adoption platforms has introduced a wide array of methods to help improve compatibility, they often tend to focus on basic filters, and don't take into account contexts such as personality, living situations or disabilities.
What it does
The platform uses a smart AI-powered system to help you find the best match possible - based on your lifestyle, preferences and needs.
Users begin by answering a series of questions touching on various topics such as schedule, housing situations, routine, and personal needs. Using this information, our system picks out adoptable animals from local shelters using a custom-trained compatibility model.
Users can further narrow down their preferences - such as choosing certain kind of pets or breeds. Our system picks out what appears to be the most compatible and suggest based on the information you have given.
The goal is to reduce mismatches, lower return rates and create long-lasting, happy adoptions.
How we built it
The frontend solutions for AI-Powered Pet Matchmaker:
- React
- Redux
- Tailwind
- PropelAuth
The backend solutions for AI-Powered Pet Matchmaker:
- Node.js
- Express.js
- MongoDB
- MongoDB Atlas
- AWS
The ai solutions for AI-Powered Pet Matchmaker:
- SKLearn
- Pandas
- Vertex AI
- AutoML
The projects for AI-Powered Pet Matchmaker is hosted on:
- Render for both frontend and backend
Challenges we ran into
- One major hurdle for us was coordination, as when we started many of us were out of country - this leads to timezone mismatches, which made meetings and cooperation rather difficult.
- We also had to think about balancing technical implementation with user experience as well as accessibility, and ensuring the matches fit with various individuals.
- Effective communications between team members working on different branches was difficult, but taught us a lot.
- We have only 30 pets in our database, therefore, sometime the mtachmaker cannot find any pets matched the users answers
Accomplishments that we're proud of
After two months of planning, developing, and testing, we are extremely proud of our application. Here are our proudest accomplishments:
- We successfully launched the site, which features a functional pet recommendation engine. We made something that centers empathy and accessibility, ensuring that people from various lifestyles can find the right pet.
- Our team comprises of individuals with diverse backgrounds, and we effectively leveraged each other’s expertise to create a comprehensive application. We committed ourselves fully from start to finish.
- We have an amazing document. Developers can easily run our website on their computers, making website improvements smoother.
What we learned
- We learned a lot about the React framework, the Express framework, and AI models, as well as how to connect them into one application.
- We also learned how to create useful documentation so that developers can understand how to run and use it effectively.
- Good projects don't start and end at code - it's about empathy, problem solving, effective team coordination and understanding various needs.
- How to break down large tasks and build something shippable from day 0, with realistic time frames.
- The success of an app or software depends not only on the quality of the code but also on planning, testing, and suggestions. Without the significant help of mentors, we would not have been able to develop our application to be this successful.
- It is fine not to know everything right away. Helping and learning is how we improve: as coders and as people.
What's next for AI-Powered Pet Matchmaker
There is significant potential to offer our application to other nonprofits for research purposes.
- We can include more questions to help improve matchmaking accuracy.
- Many undeveloped features, such as options for voluntary donations and fostering, would provide substantial opportunities for the website.
- Additionally, the website could be made accessible to other shelter organizations.
- We can add volunteer roles to the official account databse
- Since this project is a base, we can integrate other projects from OHack on top of it, such as pet sponsorship and social media pet story sharing
At the end of the day, our goal is to ensure that the website effectively helps Sequoia Humane Society and other shelters facilitate successful pet adoptions. We aim to support customers in finding lifelong companions, enriching their lives and reducing the likelihood of pets being returned.
How we addressed each of Judging Criteria
Scope of Solution
We have finished the P0 and P1 project requirements, which includes:
- User questionnaire capturing lifestyle factors
- Basic matching algorithm
- Pet profile database
- Mobile-responsive UI
- Storing user preferences
- Filter and sort system
- Email notifications
Documentation
Our team has created API documents, the solution's architecture, and how to run locally on GitHub. They will allow developers to use and test our software with ease. It is attached in this Devpost.
Polish
Both frontend and backend are deployed by Render. The main features of the site are fully functional.
Some potential features we would like to add are:
- Post creation system
- Improved email notifications/newsletter system.
- Donation, fostering, and volunteer features.
Security
We have implemented a secure login system for both admins and users using PropelAuth, which utilizes OAuth2 technology. In addition to authentication, PropelAuth also assists with user management and can be integrated into both the front-end and back-end through various useful classes and properties.
Only the Sequoia Human Society has the authority to create admin accounts, and only admins are permitted to add, remove or edit pets from the site. Admins will have their own usernames and passwords, which will be kept confidential from users.
Log in or sign up for Devpost to join the conversation.