About the data
New England Ski Data is an open collection of everything measurable about lift-served skiing in Vermont, New Hampshire, Maine, Massachusetts, Connecticut and Rhode Island. A crawler gathers it from public sources on a schedule and keeps a record of where every number came from.
Sources
- Wikipedia
- The list of ski areas per state, and each area's infobox: elevations, vertical drop, skiable acres, trail and lift counts, snowmaking coverage, ownership and opening year.
- OpenStreetMap
- Individual lifts (type, capacity, length, year built) and trails (rating, length, lighting, glades) with their geometry, queried through the Overpass API around each area's coordinates.
- Open-Meteo (ERA5 reanalysis)
- Daily snowfall, precipitation, temperature and wind at each area's coordinates since 1990. Reanalysis is a physical model constrained by observations, so it is consistent everywhere but does not capture summit-level snow stakes. Expect modeled totals to sit below what a resort reports.
- Resort snow reports
- Trails and lifts open, base depth, recent and season snowfall and surface conditions, parsed from each resort's own conditions page. Today's page is fetched daily; the Internet Archive's captures of the same page reconstruct past seasons.
- Resort ticket pages
- Day-ticket and pass prices, from the resort's current page and archived copies.
How derived figures are computed
- A season runs from August to the following July and is named for its starting year, so “2024-25” covers winter 2024 to 2025.
- Opening and closing dates are the first and last snow report in a season that showed terrain open. They can only be as precise as the archive's capture frequency.
- Average terrain open is the mean of “percent of trails open” over all captured reports in the season, so seasons with few captures are noisier.
- Long-run snowfall averages use seasons from 1991-92 onward.
Two kinds of snowfall numbers
Reported totals are what a resort publishes from its own snow stake, collected here via OnTheSnow's season history and archived snow reports. Modeled totals come from the ERA5 reanalysis at the area's coordinates, which represents a grid cell tens of kilometres wide, usually far below the summit. Across the seasons where both exist, reported totals run 1.3x above modeled at the median and 3x to 4x above at the big northern mountains, so the site never mixes the two: reported numbers lead wherever they exist and modeled numbers are labeled as such and used for trends, monthly patterns and pre-2012 history.
Caveats
Snow-report parsing is heuristic and each page layout is different, so a small share of reports will be misread. Every stored value links back to its source page so it can be checked. Small community hills often have no Wikipedia infobox or OpenStreetMap coverage and will show gaps until someone maps them.