Helps users in North America, Australia, and New Zealand find nearby parks with low light pollution, cloud coverage, humidity, and moon illumination. These parks are ranked according to our proprietary algorithm - approved by an astronomer from McMaster University - which assigns a score indicating naked-eye star visibility. A score of 80% or greater represents good visibility.
This saves time and energy by not having to read light pollution maps. Currently, the search query "where to stargaze" returns blogs and niche websites with either minimal or complex information.
The dataset consists of:
Satellite Data - a satellite image provided by the Earth Observations Group (EOG) at the National Centers for Environmental Information. This Day/Night Band image captures the average radiance values of North America, Australia, and New Zealand. The pixel values we extract correspond to a number on the Bortle scale, which gives an approximate measure of the night sky's brightness at a particular location.
GIS/Park Data - Parks within 140km of the user in lower light-pollution zones are ranked based on not only the light pollution, but cloud coverage (the % of the sky covered by clouds), humidity %, and moon phase.
Weather Forecast Data - Weather forecast data is obtained from OpenWeather. Due to the data limitations and the wide area required for forecasting, doing forecast requests for each park individually was infeasible, as well as not financially viable. As a result, a k-means clustering algorithm was used to cluster nearby parks together, since all parks in an area could share a forecast. After some testing, there was not a significant difference between using nearest neighbor and the more elaborate methods, so nearest neighbor was used. As a result, the centroid of each cluster of parks is used as the forecast request point, and all parks in a cluster share the same forecast.
More detailed information can be found on the FAQs
