A general feedback I've been receiving is that the tech salary distribution in Portugal is biased towards lower salaries when compared with other countries. Currently, Fairly measures fairness against the learned distribution, but if the distribution itself is biased how can we disambiguate?
A possible approach is to generalize Fairly for other countries and add an extra feature to compare across countries. This would enable users to understand how salaries (normalized by cost of living) vary across Europe for example. This means we would require a similar dataset to the Tech Report 2021 by landing.jobs but collected for more countries, some potential datasets are listed below:
A general feedback I've been receiving is that the tech salary distribution in Portugal is biased towards lower salaries when compared with other countries. Currently, Fairly measures fairness against the learned distribution, but if the distribution itself is biased how can we disambiguate?
A possible approach is to generalize Fairly for other countries and add an extra feature to compare across countries. This would enable users to understand how salaries (normalized by cost of living) vary across Europe for example. This means we would require a similar dataset to the Tech Report 2021 by landing.jobs but collected for more countries, some potential datasets are listed below: