StepRight

Inspiration

As an all-girl team we have faced the same issue - walking late at night doesn't feel safe. 'This person seems suspicious. I have to stay alert.' are thoughts that often cross our minds. As first-years in college too, we quickly realized how scary it can be to walk alone at night. With late-night classes, long study sessions at the library, or heading back from campus events, we have often felt unsure about which paths were truly safe. Many streets end up being poorly lit or full of crime, making these walks very stressful. We wished there was a way to know, before stepping out, which route would be the most favorable. Using that frustration, we were inspired to build StepRight, a tool designed to help students (or anyone) walking at night feel secure.

What it does

StepRight recommends walking routes that prioritize safety, offering multiple different routes that are optimized for safety. Users can see which paths maximize visibility and safety, while also still being efficient. The goal is not just to give a route, but to provide a peace of mind. Users should feel confident walking alone, even at night. We built this with the user’s safety as the central focus, blending data-driven recommendations with an easy-to-read interface. With an emergency 911 call button, we want to make sure first-responders are quick to react if anything were to happen too. Our website considers various data such as data about crimes in Berkeley and their severity, location of street lights and police stations, user feedback on safety etc. to provide users with the safest route to the end location as well as some safe faster options.

How we built it

We started with publicly available street light location data and previous crime locations in Berkeley, generalizing the data to provide the safest routes for the user. Using this information and a designed algorithm, we calculated safety scores of street segments in Berkley and then used a minimum spanning tree based algorithm to compute the safest paths. We then built a web-based interface that displays the different routes on a map, giving users the information about each of the five paths. They can toggle between each path and follow the one they seem the most favorable. During development we experimented with multiple scoring systems to balance safety and path distance, ensuring that users wouldn’t be forced to take an unreasonably long detour just to walk under more safe segments. The algorithm we used to find the safest path once we have computed the safety scores for each street segment is a modification of Kruskal's minimum spaning tree algorithm using Disjoint Set Union, where we sort segments by their safety index in decreasing order and gradually add them in order until we find a path from the start to the end point. For the other suggested paths (the best combinations between safety and distance), we use various optimal thresholds we found during experiments.

Challenges we ran into

By far, our biggest challenge was finding the most optimal path for our users. We started with an algorithm where we focused only on safety and didn't take distance into account. That ended up with a route of over 1 mile for a destination down the street. However, that problem was quickly solved with the implementation of Google’s Routes API, we fixed a maximum walking radius. The safety scores metric was another challenge. We determined the optimal formula after multiple experiments. Another challenge was working with the datasets as they were very big. We had to do a lot of preprocessing on them, including cleaning, feature engineering, and selecting the data we needed. Additionally, another challenge we faced was designing the interface in an intuitive way so that users wouldn’t be overwhelmed by the features. Our intent is to make this an easily accessible app so users can quickly walk home and spend less time out at night. Our focus is simplicity and safety, so our website should represent that too.

Accomplishments that we're proud of

Despite our challenges, we’re proud of the working website that successfully recommends safer walking routes. With the integration of multiple datasets, designed with a scoring algorith, and a clean user-friendly interface, we’re proud that StepRight can actually create an impact. We’re particularly proud that StepRight transforms a stressful, anxiety-inducing experience into a more manageable and informed one. It’s rewarding to know that our project has the potential to improve daily life for students and the community.

What's next for Step Right

We’re looking to expand this product beyond the city of Berkeley. We want to include features like real-time updates from community reports, and obviously, improved coverage of street lighting data. Our ultimate goal is to help anyone, anywhere feel confident and safe walking alone at night. We also plan to take users input into account: users give feedback on whether or not they felt safe during their walk, each user has a trustworthiness score based on how close their opinion is to the oppinion of the other users. Then each user's input is taken into account multiplied by an index determined by their trustworthiness score. StepRight is not just a project. It’s a step right toward empowering people to take control of their safety, reducing fear, and encouraging freedom of movement after dark. We hope to inspire others to think about safety as something that can be enhanced through thoughtful, user-centered technology.

Built With

• berkeley-open-data
• css
• google-geocoding
• google-maps
• google-places
• google-routes
• html
• javascript
• python