Welcome to the Gender Inference Methods Insight
Who is it for?
If you are a (social) scientist interested in utilizing gender inference methods for your research then you are at the right place. Do not worry if you think you do not have an expertise in computational techniques for gender inference. Our methods insight will enable you to explore, evaluate and execute a range of methods for inferring gender based on first and last names and/or images of people .
Dislaimer: Gender inference' in the current literature means the assignment of either a "male" or female" label, which we adopt here for consistency. There are of course several ethical issues with inferring binary 'gender'. For a further discussion please read our guidelines for responsible usage.
What do we offer?
Explore
Get an overview of the state of the art methods for inferring gender from individual's names and/or images.
Evaluate
Explore how the gender inference methods perform in our assessment on a wide range of benchmarking datasets from different sources (e.g., Wikipedia, IMDB, scholars, Twitter, etc.). These datasets contain gender labels and socio-demographic attributes for individuals (e.g., genders and skin type) and images with varying photographic conditions (e.g., poses, lighting, noise, etc.). Based on the performance benchmarking measures, we also provide suggestions or hints on which method may be most suitable or best performing for different use-case scenarios.
Execute
Get easy-to-follow instructions on how to use the different methods, including directly executable demo notebooks for trying them out.
Examine images methods
Examine names methods
Examine images+names
methods
Licensing information
The GESIS CSS Methods Insight for Gender Inference is a non-commercial service intended for academic use. Please check out the individual methods and datasets for licensing information.