Pandas flashcards, generated for you.

A Series is a 1D labeled array with index and values (like a column). A DataFrame is a 2D labeled table with multiple Series as columns, indexed by rows and columns.

Use df['column_name'] or df.column_name. Returns a Series object (1D). Use df[['column_name']] with double brackets to return a DataFrame (2D).

.loc[] uses label-based indexing (row/column names or boolean masks). .iloc[] uses integer position-based indexing (0, 1, 2...). Both support row and column selection.

Boolean indexing filters rows using a condition that returns True/False values. Example: df[df['age'] > 30] returns rows where age exceeds 30. Useful for data filtering and subsetting.

.groupby() splits data into groups by one or more columns, applies an aggregation function, and recombines results. Workflow: groupby → aggregate (sum, mean, count, etc.) → result as new DataFrame.

.merge() joins on specified columns (SQL-like, flexible); .join() joins on indices by default. .merge() is more explicit for column-based joins; .join() is simpler for index-based joins.

A MultiIndex has multiple levels of row or column labels, enabling hierarchical indexing. Use it for organizing complex data (e.g., time-series by region and product) and enabling advanced slicing and grouping operations.

.apply() applies a function element-wise or row/column-wise to Series/DataFrame. .map() is Series-only, applies a function element-wise or uses a dict for value mapping. Use .apply() for row operations; .map() for simple value replacements.

.pivot_table() reshapes data by specifying index, columns, and values, with automatic aggregation (default: mean). Useful for summarizing data by multiple dimensions, creating cross-tabulations, and exploratory analysis.

.fillna() replaces NaNs with a value, forward-fill (.ffill()), or backward-fill (.bfill()). .dropna() removes rows/columns with NaNs. Fillna preserves rows; dropna loses data. Choose based on analysis intent and data importance.