What is the best Dimensionality reduction method for mixed dataset of categorical, binary and continuous variables?

Question

I have a dataset currently with around 15k entries and 100+ variables of categorical variety (gender, race, occupation), continuous variety (income, age, bank balance), and