SafeRoute Navigator
SafeRoute Navigator is a pollution-aware navigation app that helps users find walking and cycling routes with lower estimated air pollution exposure by comparing AQI, distance, and travel time.
magine walking to college, cycling to work, or travelling across a busy city. You open a navigation app, enter your destination, and choose a route based on distance or travel time.
But there is another question worth asking:
What if we could compare routes based not only on how far we travel, but also on the air pollution estimated along the way?
That question inspired SafeRoute Navigator, our hackathon project focused on making route planning more air-quality-aware.
Instead of treating pollution as information separate from navigation, SafeRoute explores how air-quality estimates can become part of the route decision itself.
Our goal is simple: help people make more informed travel choices, one journey at a time.
The Problem: Navigation Tells Us Where to Go, but Not the Whole Story
Air pollution is a major public-health concern. The World Health Organization reports that 99% of the world's population breathes air exceeding its air-quality guideline limits. WHO: Air pollution
For people moving through busy cities, the environment along a journey matters. Traffic corridors, changing weather conditions, and local surroundings can make air quality an important consideration when choosing how to travel.
Yet route planning traditionally emphasizes familiar metrics such as distance, estimated travel time, and convenience.
Air-quality information, meanwhile, is usually presented through separate dashboards, monitoring stations, or forecasting tools. That leaves users with a practical gap: they may be able to check pollution levels, but it is much harder to incorporate that information into a specific journey.
SafeRoute Navigator explores a different approach: bring route planning and air-quality information together in one experience.
Our Solution: Navigation With Air Quality in Mind
SafeRoute Navigator is a map-based route-planning prototype designed to incorporate estimated pollution exposure into travel decisions.
Users can enter an origin and destination by searching for places, select locations on the map, or use their current location as the starting point. They can choose a travel mode, request routes, and view the resulting route geometry alongside available pollution information.
Depending on the API response, route information includes:
- Route distance and estimated travel time to understand the practical cost of each option.
- AQI and PM2.5 estimates to make air-quality information visible during route comparison.
- Data freshness and coverage to provide context about the returned estimates.
- Confidence information to communicate uncertainty instead of presenting model outputs as ground truth.
The idea is not to promise a perfectly clean route. It is to make air quality another factor people can consider when deciding how to travel.
How SafeRoute Works: From Coordinates to Route Insights
Behind the interface is a pipeline that connects a mobile application, a route API, a road network, and pollution estimates.
1. Choose a journey
The Flutter application provides the map and route-planning interface. Users can set their origin and destination through place search, map interaction, or current-location selection.
The application converts these selections into geographic coordinates: latitude and longitude.
2. Send a structured request
When the user requests routes, the frontend sends a POST request to the backend route endpoint.
The request contains the origin, destination, travel mode, and route preferences. This gives the backend the information it needs to attempt to produce relevant candidate routes.
3. Bring in road-level estimates
Our data pipeline works with New Delhi road-network data and modeled air-quality estimates.
In the current dataset, modeled estimates have been generated for 14,142 of 32,079 road segments. These estimates are stored in MongoDB alongside segment information and fields such as estimated PM2.5, estimated AQI, exposure score, confidence, and contributing factors.
This creates a foundation for connecting route geometry with estimated pollution conditions.
The distinction matters: a modeled estimate associated with a road segment is not the same as a pollution measurement taken directly on that street.
4. Return route geometry and exposure data
The route API returns candidate routes as GeoJSON line geometries, together with route information such as distance, travel time, and available exposure metadata.
The Flutter application converts the returned longitude-latitude coordinates into map points and draws the actual route geometry. It can then display the pollution information returned for each candidate route.
The result is a single workflow that connects a user's journey request to a visual route and its associated air-quality estimates.
The Technology Behind the Prototype
SafeRoute combines familiar application technology with a data-oriented approach to route planning.
Flutter and Dart power the mobile interface, including the map, origin and destination selection, travel-mode controls, route visualization, and results.
Flutter Map and OpenStreetMap provide the map display and geographic context.
Geocoding and location services help users select places and obtain coordinates for route requests.
An HTTP-based backend API receives journey requests and returns route candidates and exposure metadata.
MongoDB stores modeled pollution estimates associated with mapped road segments.
Open-Meteo air-quality data contributes modeled pollutant information to the estimation pipeline. Its documentation describes forecast products with different spatial resolutions, including a global atmospheric-composition forecast. This is important when interpreting the results: model-based estimates must not be mistaken for independent street-level sensor readings. Open-Meteo Air Quality API
Together, these components form the foundation of an air-quality-aware navigation experience.
What Makes SafeRoute Different?
We believe the value of SafeRoute lies in three design choices.
It puts air quality in the travel decision
Rather than requiring users to interpret a separate pollution map and then plan their journey elsewhere, SafeRoute brings route information and pollution estimates into one interface.
It makes uncertainty part of the experience
Environmental models are not perfect representations of real-world conditions. Data can be incomplete, inputs can be coarse, and observations can become outdated.
SafeRoute includes fields for freshness, coverage, and confidence so these limitations can become visible to users. Our next step is to make those signals even clearer and ensure they are interpreted consistently.
It connects the user experience to an actual data pipeline
The project is more than a static map concept. It includes a Flutter frontend, a route-request workflow, GeoJSON route rendering, a backend API contract, road-network data, and persisted model-based pollution estimates.
That connection between data processing and a user-facing decision tool is the foundation we want to develop further.
Building Responsibly: What We Know and What We Still Need to Validate
A meaningful environmental application must be honest about its limitations.
Our current prototype is a starting point, not a certified pollution-monitoring system. The modelled estimates do not establish the exact pollution concentration a traveller will encounter on every street. Existing data coverage is incomplete, and reliable ground-truth observations are needed to evaluate how accurately the estimates reflect local conditions.
We also need to continue validating route geometry, the consistency of exposure scores, confidence calculations, and backend reliability across different journeys.
For that reason, SafeRoute should present pollution information as estimates for decision support, not guarantees of personal exposure or proof that one route is safe.
We see this transparency as part of the product, not a disclaimer added at the end.
What's Next?
Our roadmap focuses on turning the prototype into a more reliable and useful experience.
Better validation: Compare modelled estimates with suitable, time-aligned observations and clearly document where the model performs well and where uncertainty remains.
More dependable routing: Validate returned routes against the road network, test a wider range of journeys, and strengthen backend error handling.
More useful route comparisons: Present distance, travel time, estimated pollution, data freshness, and confidence together so users can understand the trade-offs between available options.
Improved coverage: Expand and improve the road-segment estimate dataset while communicating where estimates are missing or insufficiently supported.
A more accessible experience: Continue refining the mobile interface so entering a journey, comparing routes, and understanding pollution information feels straightforward.
Each improvement brings us closer to our broader objective: making environmental context easier to use in everyday mobility decisions.
Beyond the Hackathon: A Small Change in How We Think About Travel
SafeRoute Navigator began with a simple question: why should route planning consider only how quickly we can reach a destination?
Distance matters. Time matters. Convenience matters.
But so does the environment through which we travel.
By exploring how modeled air-quality estimates can be connected to route geometry and presented transparently in a mobile interface, SafeRoute takes a step toward a more context-aware approach to navigation.
We are not claiming that software can eliminate air pollution or guarantee a healthier journey. We are building toward something more practical: giving people additional information to make their own travel decisions.
SafeRoute Navigator — not just a way to find your route, but a way to understand more about the journey.
Enjoyed reading this content? Let the author know!
Your likes, comments, shares, and saves help creators reach more builders.
Loading recommendations
Loading article