MapAbility

Day Planner - the LLM agent that plans out your day based on your needs and preferences
Your Profile - where you can update your preferences to ensure a better, more personal experience
Explore Page - the main attraction and value proposition
Relaxing Weekend - midway through the generation of answer

Inspiration

For too long society has judged those are different from the group whether it be differently abled or different sexuality. Our app allows underrepresented people to flip the script on society and pass judgement on those buildings that are not considerate of their needs and requirements.

What it does

All a user needs to do is to search for a building they have visited before, and give it a review based on certain accessibility and friendliness metrics. Users can see the reviews for various buildings posted by other people so that they are equipped with the knowledge of which buildings are accommodating for their needs and which ones are less considerate. The user can also plan out journeys and make day planners where each building is selected so that it best fits with their desired plan as well as their specific needs and preferences.

How we built it

We used MongoDB for our backend database and docker to integrate containerization and tie everything together. Containerization allowed us to easily code in one place and then quickly deploy to AWS to host the backend infrastructure of our application. For our front end we used NextJS, ShadCN, TailwindCSS, and the Google Maps API. We integrated generative AI into our project by implementing function calling technology, which allowed us to structure user inputs and accessibility data. This structured information was then fed into OpenAI's Large Language Model (LLM) to generate highly personalized and accessible activity plans, enhancing the overall user experience.

Challenges we ran into

One of our biggest challenges was figuring out how to create and manage our own APIs using FastAPI. Something that we ended up spending a lot of time trying to debug was when an API call with a specific endpoint never ended up getting called because it would default to an API with a general endpoint in a very similar link structure. Something else that we struggled with was how to aggregate various reviews into a single rating for a building. This step was crucial as it was a part of the MVP but it was an also a prerequisite for the day planner feature. What appeared to be a trivial task of just needing to take an average became actually tricky when it came to implementing it.

Accomplishments that we're proud of

We are most proud of our explore functionality MVP and our overall UI/UX. Being an accessibility app, we are proud that our app itself caters to various accommodational features - dyslexia friendly text, text resizing optionality, contrast mode and text-to-speech and speech-to-text. Our filtering system goes beyond just the basic categories the buildings belong to. Our system allows users to filter by 8 different accessibility needs so that a wide array of needs can be addressed. Each building has a high quality of user generated reviews while also backed by AI based summaries and net aggregation giving users a seamless, dependable source of information about the buildings around them.

What we learned

Our hackathon journey was a crash course in both technical skills and teamwork. We dove into AWS, FastAPI, React, and Next.js, tackling challenges like API endpoint conflicts and data aggregation. These experiences sharpened our problem-solving abilities and taught us the value of effective communication within our team. Beyond the code, we learned crucial lessons in project management and personal growth. We mastered the art of prioritizing tasks to deliver an MVP under tight deadlines. Along the way, we discovered our capacity for resilience and adaptability, fueled by ample coffee and a flexible approach to sleep schedules. These technical and soft skills will undoubtedly prove invaluable in our future endeavors.

What's next for MapAbility

Going forward we need to encourage more and more people to keep using our app an reviewing buildings. Since a major aspect of our app is dependent on user input, this does pose a network externality barrier initially. However, with enough people using our app, we can start to drive change and ensure that new buildings to come as well as existing buildings start to accommodate the needs of underrepresented groups so that they do not feel excluded and or feel like outsiders.