getdummies

get_dummies is a function in the pandas library that converts categorical variables into dummy or indicator variables. It takes a DataFrame or Series and returns a new DataFrame where each category value in the selected columns is represented by a separate binary column. The value is 1 if the observation belongs to that category and 0 otherwise. The function can be configured with parameters such as columns to encode, prefix and prefix_sep to control new column names, and dummy_na to create an extra column detecting missing values. In addition, drop_first can be set to True to drop the first category in each encoded column, which helps avoid multicollinearity in some models. The sparse parameter can request a sparse representation to save memory on large datasets.

get_dummies can be applied to multiple columns by passing a list to columns, or to the whole DataFrame by default. It will only transform non-numeric columns, leaving numeric columns unchanged where appropriate. Example: df = pd.DataFrame({'color':['red','blue','red'], 'size':['S','M','L']}); result = pd.get_dummies(df, columns=['color','size']). The result contains columns like color_blue, color_red, size_L, size_M, size_S with binary indicators.

Limitations include high dimensionality when many categories exist, potential sparsity, and the loss of information about category order if any. For ML pipelines, alternative encoders such as scikit-learn's OneHotEncoder or target encoding may be preferred.