Counts, proportions and cardinality

Table of each categorical column: unique count, top category, top %, rare categories (<1%). Concepts: Cardinality; rare categories; Boolean features Project A: Rare levels (< 5%): credit_history 2 levels, purpose 4 levels (retraining = 9 rows), savings '>= 1000 DM', co-applicant, stores, unemployed/non-resident job, foreign worker = no (37). Show bad rate + n for each. Project B: Few categoricals: weather_main/weather_description per city (high cardinality text labels), plus derived hour/weekday/month/holiday later.