HokieRoute

Walk safer. Stay aware. AI-powered route planning with real-time safety insights around Virginia Tech and downtown.

Comment

Inspiration

As students at Virginia Tech, we receive daily campus crime alerts in our email inboxes. Over time, we began noticing a worrying trend: a visible increase in reports related to sexual harassment, assault, and unsafe late-night encounters around both campus and downtown Blacksburg.

Those alerts stopped feeling like statistics and started feeling personal. These were places we walk every day: to class, to the library, to friends’ apartments, or back home late at night.

During this hackathon, we asked ourselves a simple question:

What if navigation apps optimized for safety first, not just speed?

As strong supporters of Title IX and student safety, we built HokieRoute. A system designed to actively reduce risk exposure by helping students choose smarter, safer paths in real time.

What it does

HokieRoute is a web-based mapping and routing platform that calculates the safest route between two points on campus using six years of historical crime data.

Unlike traditional navigation apps that optimize only for ETA (Estimated Time of Arrival), HokieRoute optimizes for two variables simultaneously:

  • Safety
  • Time

This means users aren’t just shown the fastest path, they’re shown the safest reasonable path.

The app:

  • Visualizes campus risk zones on a map
  • Scores streets and areas by historical safety
  • Generates safer alternative routes
  • Provides an SOS emergency button for immediate assistance
  • Announces your live location automatically during emergencies

The goal is simple: reduce the probability of walking through high-risk areas, especially at night.

How we built it

We began with publicly available VTPD daily crime logs, which were messy, inconsistent, and spread across years.

So we:

  • Cleaned and standardized categories
  • Normalized timestamps and locations
  • Aggregated everything into a unified CSV

We designed a quantitative safety model that converts crime density into a normalized score:

$$ \mathrm{Safety} = 100 \times \left( 1 - \frac{\mathrm{risk}}{\mathrm{max_risk}} \right) $$

This produces:

100 --> safest areas

0 --> highest risk areas

Each street segment gets a safety weight that feeds into routing.

The stack we used was:

Database: Snowflake Maps & routing: OpenStreetMap API Backend: route scoring + risk weighting logic Frontend: web app interface Emergency system: ElevenLabs voice + instant SOS call

We added an SOS button that:

  • Contacts local authorities
  • Shares the user’s live location
  • Plays a spoken message aloud with coordinates

This ensures help can be dispatched even if the user cannot speak.

Challenges we ran into

One of our biggest technical hurdles was geolocation accuracy.

Initially, the app sometimes returned a user’s previous cached location instead of their live position, which could lead to unsafe or incorrect routing.

We traced this to stale server-side location data and implemented:

  • forced refresh logic
  • real-time polling
  • server reset handling

After the fix, live tracking became significantly more reliable.

We also faced challenges balancing:

  • safety vs ETA tradeoffs
  • sparse crime data in certain zones
  • and routing performance at scale

Each required careful tuning of weights and thresholds.

Accomplishments that we're proud of

We’re especially proud of:

  • A working dual-variable routing algorithm
  • A full historical safety dataset spanning 6 years
  • Real-time safety-aware navigation
  • An SOS system that can actively call for help
  • A product that directly addresses a real campus issue

But most importantly, this isn’t theoretical. It’s something students could actually use tonight.

What we learned

This project taught us that:

  • Data science can directly impact personal safety
  • Routing algorithms can be adapted for social good
  • AI audio systems can meaningfully assist in emergencies
  • Technology should prioritize people, not just efficiency

What's next for HokieRoute

We are planning to build a mobile application for the same. Next we would try to collaborate with VTPD and Blacksburg PD to integrate some more safety and security features that can help make Hokies truly safe around campus.

Built With

Share this project:

Updates