Citizen science projects invite members of the public to contribute to scientific research, often by performing simple but large-scale tasks such as classifying images, identifying species, or recording observations. These projects can generate enormous amounts of valuable data — far more than individual researchers could analyse alone. However, volunteer attention and effort are limited resources, which raises an important question: how can citizen science projects be designed most effectively?
In this project, you will investigate statistical and practical questions that arise in large citizen science platforms using freely available real-world data. For example: How many times should each image be classified before we can trust the result? Are experts always better than enthusiastic amateurs? How can we identify reliable contributors? How often should participants be shown “control” images with known answers to check accuracy and maintain quality?
You may also explore broader questions such as participant fatigue, disagreement between classifiers, or how to balance accuracy against time and effort. Depending on your interests, the project could involve statistical modelling, simulation, Bayesian methods, machine learning, or optimisation approaches.
This project is ideal if you enjoy working with real data, asking practical questions, and developing methods that can improve how large collaborative research projects operate. It offers considerable flexibility and can be shaped toward your interests in statistics, data science, ecology, psychology, or human decision-making.
Supervisors
Primary Supervisor: Elena Moltchanova
Key qualifications and skills
You will need to have a solid grasp of probability theory, algebra and calculus, as well as facility with a programming language such as R or Python.
Does the project come with funding
No - Student must be self-funded
Final date for receiving applications
Ongoing
How to apply
Apply by email to primary supervisor
Keywords
statistics; citizen science; data science
