Multimodal & Crowdsourced Data

Goals: To understand the strengths and biases of various biodiversity image data sources, particularly citizen science and crowdsourced data, for a variety of tasks.

Overview: This project investigates the utility and limitations of citizen science and crowdsourced image data for biodiversity applications such as population size estimation and Species Distribution Modeling (SDMs). The team is analyzing how data quality and source characteristics impact the effectiveness of these methods. The research aims to inform future model development and data collection strategies by identifying both opportunities and biases present in existing biodiversity image datasets.