Inspiration

When we went to Chicago for our college visits, we were constantly scared to take public transportation or the bus and racked up hundreds of dollars in parking fees, gas, and rentals. We shouldn't have to make financial and emotional sacrifices to explore the places we love. In our own towns, we hear by word of mouth where to avoid but in a strange city, we are most at risk. The number of red dotted markers across our own city, DC, as well as Chicago, made us concerned to do good within our communities to help visitors feel safe. So, our team decided to build an app that combines all the best parts of travel, communication, and the mobile platform into one.

What it does

When you're visiting a new city, our app allows you to plan your experience through a whole new lens. You can customize where you're going, how long you're spending, modes of transportation, receive helpful travel notifications as you go, all while being guided through city that's unfamiliar. We built an iOS app using Swift that'll get you where you want to go by the best means imagineable.

How we built it

We used Open Data sets from Kaggle to construct clusters of crime within DC. For the year 2017 alone, there were over 200,000 violent crimes. We used unsupervised learning to train & fit our model to identify trends of risky areas within the city. Then, we applied greedy algorithm on bus stops across the city to identify safest and most convenient routes through the city: for this, we used WMATA's rest API to represent stops along all bus routes in the WMATA system. We used a heuristic for our algorithms combining distance from violence centers we gathered through our model, as well as time-cost and distance-cost.

Challenges I ran into

One of our team members left the first day of brainstorming so not only was this one of our group member's first hackathons, we needed to do the work of 4 people on limited bases. We worked through the whole night. In addition, we had very little Google Maps experience before building this iOS app, which ended up being heavily reliant on the power of Google Places, Firebase, Firestore, SnaptoRoads, Cloud Functions, and Geolocations. In addition, turning all of DC's bus stops into a graph proved to be tricky, as well as dealing with the enormous amounts of data we had from Kaggle.

Accomplishments that I'm proud of

We're incredibly proud of how clean, beautiful, and intuitive our interface was in the end: we incorporated Technica colors, a fluid search interface, login authentication, path highlighting, and design. For a 2-person team, we accomplished everything we set out to do and learned so much about UI, machine learning, thinking algorithmically, and leveraging API's for the better of us all. For one of us, it was our first time ever using REST API's, and we're so proud of the opportunity to build an app that matters to us while exploring how to use Google Cloud.

Built With

Share this project:
×

Updates