Welcome to the GESIS CSS Methods Portal for Gender Inference

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 portal will enable you to explore, evaluate and execute a range of methods for inferring gender based on names and/or images .

What do we offer?

With our Methods Portal for Gender Inference you can:

explore Explore: We provide you an overview of the state of the art methods for inferring gender from individual's names and/or images.

evaluate Evaluate: We assess the performance of the various gender inference methods on a wide range of benchmarking datasets from different sources (e.g., Wikipedia, IMDB, scholars, Twitter etc.), and containing gender labels for individuals with different socio-demographic attributes (e.g., genders and ethinicity) and different photographic conditions (e.g., poses, lighting, noise etc.). On basis of 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 Execute: We provide easy to follow instructions on how to use the different methods including directly executable demo notebooks for trying them out.

Licensing information

The GESIS CSS Methods Portal for Gender Inference is a non-commercial service intended for academic use.