HydroHomie

Framing the Problem

Our problem solves issues with the busy world of today and how we approach self-care - 80% of people do not drink enough water, so our avenue of automation was creating making drinking water easier, with gamification elements. Our app doesn't just add pointless reminders that you will ignore - but a friendly, collaborative force on a journey to forming a universally healthy habit!

Idea Explanation

Our idea is an extremely simple app that is easy to adopt into your daily life, and takes zero overhead setup. Our team can sympathise with the pain of using self-care apps that take hours to record your life in, add tasks into, customise it to your liking etc. and it all leads to one spending all their time on the prep, rather than getting results. Essentially, it just becomes another avenue for procrastination.

Therefore, all our app has is a simple login, and you can start recording your sips right away after verifying your details. When you click the 'sip!' button, your sip is recorded, the glass at the centre of the screen fills up to show you the impact, and you get closer and closer to reaching your custom hydration goal. No frills.

Your friends can join you in enjoying HydroHomie too! Our app is driven by the concept of the accountability buddy being a powerful motivator - your friend and you can compete on a leaderboard, which shows their sips and yours. In addition, if you feel like your friend hasn't sipped in a while, you can 'nudge' them to send them a notification to sip some water. This shared journey can help you form a great habit!

In addition, we also have built-in AI integration with ChatGPT sending you encouraging prompts every once in a while, which take into account your location, local weather, time since last sipped and other information that is personalised to you to give you context-aware hydration advice. No friends? No problem - ChatGPT can be a great motivator too!

Implementation

Our frontend is built with React, and it makes requests to our backend APIs. The backend APIs are containerized using Docker and deployed on AWS EC2 instances. For the database, we use NoSQL MongoDB.

Additionally, our web app calls an IP API to determine the user's location and a Weather API to get the current weather information. This data is combined with a prompt and sent as a request to OpenAI. The result returned by OpenAI is processed by the backend and then sent back to the frontend through our API for use in the app.

Challenges

One of the main challenges we faced was conceptualizing the project from start to finish, translating a broad idea into a workable product. We had to carefully think through how to integrate the frontend interface in React with AI-driven features in Python.

Another significant hurdle was determining how to source real-time weather data and use it effectively to train the AI to deliver timely and relevant prompts. Additionally, building the right prototype posed its own set of difficulties, from designing the app’s visual layout to deciding what features it should integrate, how frequently reminders should be sent, and how to time them for maximum impact. We also spent considerable time deciding on how to gamify the user experience in a meaningful way, using the 'nudge'. Through brainstorming, continuous testing, and collaborative refinement, we were ultimately able to arrive at a solution that balanced functionality, user experience, minimal congitive load and engagement.

Accomplishments

We all learned or improved to work in a team under pressure, as this was the first hackathon for some of us - we learned how to get an idea off the ground quickly in such a varied team too. All of us had completely different competencies - ML Engineer, iOS developer, Finance student, front-end developer and backend developer - putting us all together to work on a shared goal was a challenge, but became one of our biggest strengths. Our skillsets were definitely stretched as we didn't have any dedicated design team members, so all of us had to pitch in a bit to every visual component, which definitely taught us a lot!

A major focus was balancing functionality with user experience; we wanted to build a minimal cognitive burden-focused product, without creating notification overload, which would ultimately defeat its purpose. Moving from ideation to execution within such a tight timeframe put our creativity, technical abilities, and teamwork to the test, which we were able to overcome through leveraging all of our different strengths.

Next Steps

To enhance the offering of our product, the following future features are planned:

1. Expand Scope of Reminders- Integrate reminders for other activities such as walking, self-care, diet

2. Customize Dashboard and Analytics- Provide deeper insights through trend analytics and tracking, and implement adaptive recommendations based on user behaviour

3. Integrate with Health Apps and Wearables- Sync data with third-party apps like Apple Health, Google Fit, to create a holistic monitoring system, leverage wearable data to provide real-time feedback and alerts.

Team Roles and Contributions

Nicole Ho

Nicole Ho was responsible for creating the front-end prototype in React.

Cristina Sabau

Cristina Sabau created wireframes for the product's design in Figma.

Deepali Desai

Deepali Desai worked on the ChatGPT integration, Figma wireframes, and pitch deck.

Shaoyun Jian

Shaoyun Jian was our backend expert who created our architecture diagram and backend structure.

Karolina Kowalska

Karolina Kowalska managed the project lifecycle, created Figma wireframes, pitch deck and Python integration with external APIs.