aperta.data_processing

Tabular helpers that flow through aperta’s (Geo)DataFrame pipeline.

Small composable column-level transforms — dedupe an index after a join, add a straight-line origin→destination distance, take a weighted mean per group. None of these reason about geometry as the primary concern (for that see geo_processing); they operate on tables that may or may not carry a geometry column.

aperta.data_processing.remove_duplicate_indices(df)[source]

Remove duplicate indices from DataFrame or Series. Often useful after spatial join.

Parameters:

df (DataFrame)

Return type:

DataFrame

aperta.data_processing.add_straight_line_dist(df, orig_prefix='orig', dest_prefix='dest', out_col='dist_line')[source]

Add euclidean origin→destination distance, in CRS units (typically metres).

Parameters:
Return type:

DataFrame

aperta.data_processing.weighted_group_mean(values, weights, group_id)[source]

Weighted mean of values per group, indexed by group ID.

NaN-aware: rows with NaN in either values or weights are dropped before aggregation. Groups with no surviving rows (or with all weights ≤ 0) yield NaN.

All three inputs must share the same index (per-row alignment); the result is indexed by the unique group IDs. Typical use case: reduce per-cell values to per-zone (group_id = cells[‘zone_id’], weights = cells[‘population’] or similar).

Parameters:

  • values (Series) – per-row numeric Series.

  • weights (Series) – per-row non-negative weight Series (same index as values).

  • group_id (Series) – per-row group-membership Series (same index as values). Values become the result’s index.

Returns:

pd.Series indexed by the unique group IDs, with the weighted mean of values per group.

Return type:

Series