About Hikeability
AI-powered trail condition monitoring for Washington State hikes.
What is Hikeability?
Hikeability combines real hiker trip reports from the Washington Trails Association (WTA)
with live weather data to classify trail conditions across Washington State.
Instead of reading through dozens of reports, you get a single, up-to-date label
telling you whether a trail is safe to hike right now.
The map updates automatically as new reports come in and weather changes —
so conditions reflect the latest information, not a report from months ago.
Condition Labels
Every trail is classified into one of three conditions:
Hikeable
Trail is in good condition. Safe and enjoyable for most hikers.
Modest
Trail is passable but has notable challenges — patchy snow, mud, or downed trees. Experienced hikers with proper gear can manage.
Unhikeable
Trail is closed, unsafe, or impassable — deep snow, washed-out sections, or extreme weather.
How It Works
1
Scrape trip reports
Recent hiker reports are collected daily from the Washington Trails Association website.
2
Fetch weather data
Current trailhead conditions — temperature, snow depth, wind, AQI, and precipitation — are pulled from Open-Meteo.
3
Classify with LLM
An NVIDIA Llama 3.1 model reads the reports and weather together and assigns a condition label with a plain-English explanation.
4
Display on the map
Results are published to the map. Click any trail to see the full conditions summary, weather, and trail details.
Data Sources
Washington Trails Association
Trip reports and hike metadata — distance, elevation gain, region, and trail descriptions.
Open-Meteo
Free, open-source weather API providing hourly forecasts, snow depth, AQI, and wind data at trailhead coordinates.
NVIDIA Llama 3.1 Nemotron
Large language model used for few-shot classification of trail conditions from unstructured report text.
Google Cloud Storage
Stores raw reports, weather data, and model predictions. Serves as the data layer between pipeline runs and the web app.