Leveraging citizen science as a source of ecological data

Presented by 
Thomas Vattakaven 
Strand Life Sciences
Thomas Vattakaven, R. Prabhakar


The Observation module of the India Biodiversity Portal utilizes crowd-sourcing to gather ecological data. Species images are collected with species ID, location and time of observation, allowing a spatial and temporal distribution for the species data. With increasing participation and data inflow, we were interested in analyzing interesting patterns emerging with user submitted data and the validity of the identified species on the portal by laymen users.


In an ongoing study, over 7000 identified and unidentified observations on the portal were classified by taxonomic species groups and segregated. The number of interactions per observation was generated for all observations to determine the number of identification agreed upon. Taxonomic experts in the species groups of plants, insects, amphibians, reptiles and birds were consulted to validate the identification suggested by primary users of the portal.


Less than 5% of the observations were found to be un-identified with birds being the most identified with less than 1% remaining to be identified and fungi being the least identified with over 56% remaining unidentified. Most Identifications and validations suggested by portal users were found to be reliable as corroborated by independent experts in the respective species-groups


This study divulges the popularity of various species-groups among the “crowd” that contribute data and their comfort levels at identifying each of these accurately. Verifying the validity of user suggested identifications shows that over 80% of the identifications suggested are reliable and that scientists who utilize the data generated by citizen science portals have to be prepared to scrutinize the data carefully and discard data that is suspect or unreliable.