Reading the data

Results follow people

Most records are photos people took and shared on iNaturalist; the rest come from other collections through GBIF. A result shows the weeks when people recorded a species, not when or where it lives. Where and when people go, and what they stop to photograph, shapes every result.

When people go out

People record far more from late spring to early fall, more on weekends than weekdays, and in bursts during events such as the City Nature Challenge. Far fewer records are made in winter.

So a species can look as if its season is shorter, or leans toward summer, partly because fewer people were out in winter. The grey bars under each result show how many records of all species were made each week, so you can see when recording itself was low.

Study: Di Cecco and others (2021), Observing the observers: how participants contribute data to iNaturalist and implications for biodiversity science. BioScience 71(11). (opens in a new tab)

Where people go

Records gather near roads, homes, towns and parks. Places that are hard to reach, and counties with fewer people, have fewer records, so they have fewer results. A county with few results may simply have had fewer people recording there.

Study: Geurts, Reynolds and Starzomski (2023), Turning observations into biodiversity data: broadscale spatial biases in community science. Ecosphere 14(6). In British Columbia, about 94% of iNaturalist observations were within 1 km of a road. (opens in a new tab)

What people notice

Birds, flowers, butterflies and other large or colourful species are recorded far more than small, hidden or hard-to-photograph ones, such as most insects, fungi and mosses. Many species live in Ohio without enough records for a result.

Study: Troudet and others (2017), Taxonomic bias in biodiversity data and societal preferences. Scientific Reports 7, 9132. Birds were the most over-represented group in GBIF and insects the most under-represented, following public interest more than research. (opens in a new tab)

Names can be wrong

From iNaturalist, OpenPhenology uses only Research Grade records, where the community agrees on the species. Most are right, but some are not, and mistakes are more common in groups that are hard to tell apart from a photo.

Study: iNaturalist, A second experiment to learn about the accuracy of iNaturalist observations. iNaturalist's own check of a sample of observations, not a peer-reviewed study. (opens in a new tab)

More records every year

iNaturalist keeps growing, so recent years have more records than earlier ones and weigh more in a result. Each result says which whole years it pools and how many of them have records. The first and last weeks of a season are the least certain part of it, because they rest on the fewest records.

Study: Belitz and others (2020), The accuracy of phenology estimators for use with sparsely sampled presence-only observations. Methods in Ecology and Evolution 11, 1273–1285. Estimates near the middle of a season were more accurate than those near its start and end. (opens in a new tab)

No record isn't absence

A week, county or species with no records means no one recorded it there, not that it wasn't there. The site never says a species is absent.

Locations kept vague on purpose

Records of species at risk are never shown more precisely than their county. Records whose location is only known roughly count toward the state, not a county. So county maps can show fewer records than were really made there.

Ranges are maps, not records

Where a species page shows a mapped range, it comes from a published range map, not from records. It shows where the map places the species, and its source is named beside it.

Some species have no result

A species needs enough records across enough years before the site shows its season, so one busy year can't decide it. Below that, the site says there isn't enough evidence yet instead of guessing. The numbers are on the Methods page.

What the site does about it

• It counts records by week, which smooths out the weekend effect.
• It shows recording of all species beside each result.
• It says a result shows when people recorded a species, and never claims more.
• It doesn't adjust or reweight records. It shows them as they are.