vivainsights.network_p2p

Perform person-to-person network analysis and visualization.

vivainsights.network_p2p.network_p2p(data, hrvar='Organization', return_type='plot', centrality=None, community=None, weight=None, comm_args=None, layout='mds', path='', style='igraph', bg_fill='#FFFFFF', font_col='grey20', legend_pos='best', palette='rainbow', node_alpha=0.7, edge_alpha=1, edge_col='#777777', node_sizes=[1, 20], node_scale=1, seed=1, legend_ncols=0, figsize=None)[source]

Return a network plot given a data frame containing a person-to-person query.

Parameters:

  • data (pandas.DataFrame) – Data frame containing a person-to-person query.

  • hrvar (str) – Label for the HR attribute. Defaults to "Organization".

  • return_type (str) –

    Type of output to return. Valid values:

    • "plot" (default): matplotlib Figure.

    • "plot-pdf": save network plot as PDF.

    • "sankey": sankey plot of communities × HR attribute.

    • "table": vertex summary table.

    • "data": vertex-level DataFrame.

    • "network": igraph object.

  • centrality (str, optional) – Centrality measure used to scale node sizes. Valid values: "betweenness", "closeness", "degree", "eigenvector", "pagerank". When None (default), nodes are uniform size.

  • community (str, optional) – Community detection algorithm. Valid values: "multilevel", "leiden", "edge_betweenness", "fastgreedy", "infomap", "label_propagation", "leading_eigenvector", "optimal_modularity", "spinglass", "walk_trap". Defaults to None.

  • weight (str, optional) – Column to use as edge weights. None creates an unweighted graph.

  • comm_args (dict, optional) – Keyword arguments passed to igraph’s clustering algorithm.

  • layout (str) – Node placement algorithm. Defaults to "mds".

  • path (str) – File path for PDF output. Defaults to an auto-generated name.

  • bg_fill (str) – Background fill colour. Defaults to "#FFFFFF".

  • font_col (str) – Font colour. Defaults to "grey20".

  • legend_pos (str) – Legend position (e.g., "best", "upper left").

  • palette (str) – Colour palette name. Defaults to "rainbow".

  • node_alpha (float) – Node transparency (0–1). Defaults to 0.7.

  • edge_alpha (float) – Edge transparency (0–1). Defaults to 1.

  • edge_col (str) – Edge colour. Defaults to "#777777".

  • node_sizes (list of int) – Two-element list [min, max] for rescaling node sizes when centrality is set. Defaults to [1, 20].

  • node_scale (float) – Multiplier applied to node sizes. Defaults to 1.

  • seed (int) – Random seed for community detection reproducibility.

  • legend_ncols (int) – 0 for horizontal legend, 1 for vertical.

  • figsize (tuple, optional) – Figure size as (width, height) in inches. Defaults to (8, 6).

Returns:

Output depends on return_type:

  • "plot": matplotlib Figure.

  • "plot-pdf": saves PDF and returns None.

  • "sankey": sankey plot Figure.

  • "table": vertex summary DataFrame.

  • "data": vertex-level DataFrame.

  • "network": igraph object.

Return type:

matplotlib.figure.Figure, pandas.DataFrame, or igraph.Graph

Examples

>>> import vivainsights as vi
>>> sample_data = vi.p2p_data_sim()
>>> vi.network_p2p(data=sample_data, return_type="plot")
>>>
>>> # Community detection with custom resolution
>>> vi.network_p2p(
...     data=sample_data,
...     community="leiden",
...     comm_args={"resolution": 0.01},
...     return_type="table",
... )
>>>
>>> # Centrality-based node sizing
>>> vi.network_p2p(data=sample_data, centrality="betweenness", return_type="table")