Crowdsourcing Citizen Science Data Quality with a Human-Computer Learning Network.
Published in The Conference on Neural Information Processing Systems (NIPS) Workshop, 2010
Citizen science presents opportunities for crowdsourcing to produce new sources of data that were previously unavailable and even unimaginable. While engaging distributed observer networks is a well-established method for collecting spatiotemporally diverse observational data, ensuring data quality remains a key concern. Related problems of observer reliability, scalability of data verification processes, geospatial biases of observations, and motivating participation are central challenges for citizen science. This paper describes a multidisciplinary strategy for addressing these concerns in eBird through the development of a human-computer learning network.