AI species detection data visualisation

AI species detection is valuable to both conservations and ecologists, who have different goals when reporting on acoustic monitoring data. User interviews and bat monitoring field trips (in the pitch black) were done to understand the field work use case alongside the data analysis and reporting requirements. How the data is captured affects the statistical validity of the survey.

Not being a bat expert, ecologist or statistician, upskilling in this area was intense but rewarding. A ‘Batbot’ was created by building a context library of 100+ best practice guidelines, academic documents and example reports into Claude, to assist with strategic planning, requirements definition and the mathematical models that needed to underpin the reporting.

LLM-driven data visualisation prototyping using real data from acoustic monitoring devices, processed by automated species detection model, to display distribution and activity over time, geographic distribution and relationships between species activity and environmental factors such as sunrise, sunset, temperature and rainfall.