iNaturalist Data Quality Guidelines

Abstract

Data collected by volunteers (citizen science) significantly contributes to the biodiversity data available through the GBIF global portal. However, many professional researchers are skeptical about the possibility of using citizen science data for scientific analysis. The aim of the study was to develop guidelines for the quality assessment of iNaturalist amateur observation data. We obtained iNaturalsit data through the GBIF portal. Completeness and data quality were assessed. It is shown that up to a third of the observations contained in correct or suspicious values, however, most of the observations were of high quality and can be used for scientific analysis. Based on the obtained results, the following criteria for high quality iNaturalist observations were developed: (1) the observation must be identified by a qualified expert, (2) the observation date must be indicated correctly, (3) geographical coordinates must correspond to the real place of observation. Methods for detecting incorrect values are proposed. It is shown that GBIF users can check the identification of observations, evaluate the scientific qualifications of experts who identified the observation in iNaturalist, and verify the date of observation. Estimating the geo data correctness is more difficult. Verification of these data requires the most careful analysis with the involvement of additional information sources.