Methodology & Data

How we transform raw citizen science data into actionable birding information.

Last Great Places for Birds & Birders summarizes weekly birding patterns using publicly available eBird checklist data, as well as data on human modification, anchored to a curated set of important bird areas and birding destinations.

Places

Places are drawn primarily from Audubon Important Bird Areas (IBAs), supplemented by a number of recommended birding sites outside of IBAs that are identified in the "National Geographic Guide to Birding Hotspots of the United States". Within several large landscape-level IBAs (e.g. Chenier Plain) specific protected areas are noted by National Geographic; this approach parallels Audubon's treatment of the ACE Basin in South Carolina, where three specific protected areas were selected as globally important IBAs rather than the greater ACE Basin landscape. Birders will want to explore eBird Hotpsots for localized birding sites within places. For more information on IBAs see Audubon Important Bird Areas.

eBird Data

eBird histogram data (i.e. bar charts) are directly provided for most IBAs; for other places, eBird data are used from the most populated hotspot. Data were compiled in early 2026. eBird Explore

Sightings Normalization

Rather than reporting raw sighting counts, the methodology focuses on relative likelihood—how consistently species or bird groups are reported at a place during a given time window. To reduce the influence of incidental or one-off sightings, eBird data are filtered to emphasize recurring patterns. Species are included only when they are reported across multiple weeks, supported by a minimum number of checklists, and observed alongside other species at a place. Likelihood of species at places only show likelihoods of 20% of higher, for beginner and intermediate birders. Advanced birders seeking special species sightings will typically use other resources.

Integrity

The Global Human Modification dataset was used to assess Integrity at places. It assesses the degree of roads, development, agriculture, industry and other human impacts on the landscape. Only places with relatively lower levels of modification (less than 40%) are included in the active list of sites. This "Last Great Places" screen excludes some urban and other sites with higher bird sightings. The "Optimal" ranking option identifies places with relatively high sightings that are less modified -- i.e. Probability * Integrity / 100

Temporal Design

Seasonality is handled using a 48 week-based model aligned with eBird’s standard reporting bins. This data is presented in a month-friendly interface designed for intuitive trip planning. For all places a "best season" for sightings is determined, and used along with expected average sightings over a full year, to evaluate qualifying sites.

Final Site Selection

A final set of 1000 "Last Great Places for Birds and Birders" was derrived from the larger universe of IBAs. Sites were excluded if highly modified or low sightings both annually and in their best season. At least 5 sites were selected from every state, and at least 10 sites from every Bird Conervation Region. Sites were scored on three factors: expected average annual sightings, expected sightings in the best season (in order to qualify sites with high seasonal variability) and the degree of human modification. Z-scores were used to normalize and rank the top sites.

Note: Rankings reflect observed patterns in the data and are intended to help birders compare places efficiently, not to predict exact outcomes. Results are influenced by reporting density; less visited places and seasons may be underrepresented.

The Technical Stack

  • Data Source: eBird histogram data for IBAs and notable hotspots.
  • Database: Compiled in Postgres with custom SQL queries for data normalization.
  • GIS: QGIS served as the primary platform for spatial analysis and mapping coordination.
  • Web Infrastructure: Data exported to Supabase for high-performance public access.