A Method for Conveying Confidence in iNaturalist Observations: A Case Study Using Non-Native Marine Species

Area of Science:

• Ecology
• Conservation Biology
• Data Science

Background:

• Citizen science platforms like iNaturalist offer cost-effective data collection for research and conservation.
• However, data quality, reliability, and accuracy concerns, stemming from unverified observations, media quality, and georeferencing, limit their utility.
• Expert verification is critical but can be time-consuming and prone to bias.

Purpose of the Study:

• To develop a protocol for assigning confidence scores to iNaturalist observations.
• To assess the effectiveness of this confidence score in streamlining the verification process.
• To evaluate the adequacy of iNaturalist's built-in quality grade as a proxy for accuracy.

Main Methods:

• Developed a confidence scoring protocol for iNaturalist observations.
• Applied the protocol to marine alien and cryptogenic species observations from South Africa.
• Compared the confidence score's accuracy with iNaturalist's quality grade.

Main Results:

• A positive relationship was found between observation accuracy and the developed confidence score.
• iNaturalist's inherent quality grade was found to be an inadequate proxy for data accuracy.
• The confidence score effectively reduces verification time and researcher bias.

Conclusions:

• Expert verification remains crucial for citizen science data.
• The proposed confidence score is an essential tool for streamlining the verification of iNaturalist data.
• Implementing this confidence score enhances the reliability and usability of citizen science data in research and conservation.