Data sources
LikelyBird combines bird observation data with environmental and temporal context. The public site will document how each source is decoded, audited and recoded before it contributes to species-level predictions.
Bird records are collected through human behaviour as much as ecology. A checklist can reflect where people choose to go, how long they watched, what time they visited, whether conditions were favourable and whether the list was complete.
LikelyBird keeps those issues visible by turning source material into separate layers rather than blending everything into one opaque score.
Source
Checklist, weather, terrain, water, vegetation, fire and seasonal context.
Decode
Translate fields, dates, locations, effort measures and category codes into consistent project terms.
Audit
Check coverage, missingness, impossible values, join quality and spatial or temporal gaps.
Recode
Create modelling-ready variables while preserving the meaning and limitations of the source.
Publish safely
Prioritise derived summaries, explanations and predictions. Observation detail is only shown or linked where approvals and safeguards allow.
Current source families
Observation evidence, checklist completeness, effort, date and location context. Any public observation detail is subject to appropriate approvals and safeguards.
Read about eBird checklist evidenceConditions that can affect whether birds are detected and reported, such as wind, rain, temperature and recent rainfall.
Elevation, landform and water-related covariates that help describe habitat and access context across Australia.
Environmental covariates used to describe habitat structure, disturbance history and broad ecological setting.