Navigation auf


Citizen Science Zurich

Research Publication: Crowdsourcing snake identification with online communities

Species identification can be challenging for biologists, healthcare practitioners and members of the general public. Snakes are no exception, and the potential medical consequences of venomous snake misidentification can be significant. In the Snake ID Challenge, a collaboration of researchers from all over Switzerland and beyond collected data on identification of 100 snake species by building a week-long online citizen science challenge which attracted more than 1000 participants from around the world. The aim was to assess the capacity of online communities to correctly identify snakes from photographs. The researchers recently published their research findings in the Royal Society Journal.

According to the World Health Organization (WHO), venomous snake bites cause around 81 000 to 138 000 deaths and around three times as many permanent disabilities every year. Mostly affected are poor and rural communities in parts of the world where snake diversity is the highest – namely in Africa, Asia and Latin America. The number of known snake species is growing rapidly, projected to reach 4,000 by 2030. However, most local people and communities in these areas can only identify a handful of species, and even professional herpetologists may have expertise in only a certain geographic area or taxonomic group. In addition, access to resources for healthcare providers are often limited.

About the project – creating tools to identify snakes

A collaboration of researchers from all over Switzerland and beyond closely working together with the WHO and Médecins Sans Frontières, is working on creating tools that anyone can use to identify snakes, using a combination of humans and artificial intelligence. Their aim is to help clinicians to better treat snakebite cases, educate people, and at the same time improve snake conservation and help scientists discover new species. The online open identification challenge was thereby used to assess the capacity of online communities to correctly identify snakes from photographs. Furthermore, the researchers wanted to understand the relative importance of main factors that affect this process (geography, photo quality, snake taxonomy, variation among participants) and also assess the potential role of citizen science and crowdsourcing in supporting snakebite epidemiology and management.

Methods – using gamification approach successfully

Our Citizen Science platform hosted two snake identification online challenges that ran in March and November 2019. In each challenge, photos of widespread snake species from around the world were shown to participants for identification. Identification could be at the species level, genus level or family level. As part of the gamification approach, participants were given extra points for correctly identifying photos, and while they were given unlimited time to suggest an identification, their response time was monitored for analysis purposes. The top scores were displayed on the platform to motivate participants to identify many photos. The three highest scored participants received a herpetological book as prize in the end.

Results – high identification accuracy of skilled participants

A total of 1027 participants submitted 117 897 unique IDs over a period of 6 days. Participants came from 49 countries primarily in North America and Europe. 24% of them identified more than 80 photos each, with identification accuracy that varied considerably across individual participants, species and photographs. The most skilled participants were able to identify 81% of photos correctly at the species level, 89% at the genus level and 95% at the family level!

Conclusion – great enthusiasm and engagement of participants

The results demonstrate the potential of a large and active community of professionals and skilled amateurs to quickly and accurately identify snakes from photos, with only a minor to moderate influence of the geographical origin of the photo or its quality. Photo quality had a relatively small effect on identification, suggesting that even relatively poor photos taken with low-resolution cell phone cameras or in the rushed circumstances of a snakebite may be identifiable and provide useful information to field workers.

Overall, participants showed great enthusiasm and engagement, and the study provides evidence that innovative citizen science / crowdsourcing approaches can play a significant role in training and building capacity. Furthermore, involving skilled avocational snake enthusiasts in decision making could build the capacity of healthcare workers to identify snakes more quickly, specifically and accurately, and ultimately improve snakebite treatment data and outcomes.

Coming up very soon – North Carolina Snake ID Challenge

The North Carolina Snake ID Challenge is the first of many regionally-focused follow-ups to the here presented published Royal Society Open Science paper. It is hosted in cooperation with the North Carolina State Museum of Natural Sciences in Raleigh, North Carolina, which holds an annual Reptile & Amphibian Day event (this year virtual) that typically draws >50 000 people from around the southeastern USA. There are 38 species of snakes in North Carolina, six of which are venomous, but they are frequently mis-identified. The challenge will open with a video introduction on March 8th, and Dr. Andrew Durso will give a seminar on March 12th to present the data gathered so far, before the challenge closes on March 21st. The three participants who identify the most snakes correctly will get copies of the book the Amphibians and Reptiles of the Carolinas and Virginia.

Learn more and contribute to the challenge here!

Read the full published paper here:

Ursina Roffler