pyphi.plots module

Plots for pyPhi

@author: Sal Garcia <sgarciam@ic.ac.uk> <salvadorgarciamunoz@gmail.com>

Changes on Feb 25th 2026:

  • Bokeh 3.x: tooltips removed from figure(); now added via explicit HoverTool

  • Extracted shared helpers: _get_lv_labels, _get_xvar_labels, _get_yvar_labels, _new_output_file, _make_bokeh_palette, _resolve_lpls_space, _mask_by_class, _add_hover

  • Fixed loop-variable collision in predvsobs (i → j)

  • Fixed output_file inside loop in mb_r2pb

  • Fixed 2-D y_ array in score_line (.flatten())

  • Fixed nan_policy on vbar (not supported); use per-series ColumnDataSource

  • Fixed bitwise ~ on bool in lineplot

  • Added unbound-variable guard in contributions_plot

  • Replaced False sentinels with None throughout

  • Replaced list(np.arange(…)+1) with range(1, n+1)

  • Replaced string concatenation with f-strings

  • Standardised colormap keyword args (alpha=1, bytes=True)

  • Replaced math.pi with np.pi

  • loadings / weighted_loadings: PLS now produces single HTML with X (blue) and Y (red) bars on a shared axis per LV

pyphi.plots.r2pv(mvm_obj: dict, *, plotwidth: int = 600, plotheight: int = 400, addtitle: str = '', material=None, zspace: bool = False) → None[source]

Plot R² per variable per latent variable / principal component.

Produces a stacked bar chart where each bar represents a variable and each colour band represents the contribution of one LV/PC to that variable’s explained variance. For PLS models, X and Y plots are shown stacked vertically in a single HTML file.

Parameters:

  • mvm_obj (dict) – Fitted model from pyphi.pca(), pyphi.pls(), pyphi.lpls(), pyphi.jrpls(), or pyphi.tpls().

  • plotwidth (int) – Width of each plot in pixels. Default 600.

  • plotheight (int) – Height of each plot in pixels. Default 400.

  • addtitle (str) – Optional string appended to the plot title. Default "".

  • material (str) – For JRPLS / TPLS models, the name of the material block to display. If None, the R-space (blend) loadings are used. Default None.

  • zspace (bool) – For TPLS models, if True plots the Z-space (trajectory) R² instead of the X-space R². Default False.

Returns:

Saves and opens an HTML file via Bokeh’s show().

Return type:

None

pyphi.plots.loadings(mvm_obj: dict, *, plotwidth: int = 750, xgrid: bool = False, addtitle: str = '', material=None, zspace: bool = False, ScaleToPlot: bool = False) → None[source]

Plot model loadings as vertical bar charts, one chart per LV/PC.

For PCA models: each chart shows X-space loadings (P or W*) in blue.

For PLS models: each chart shows X-space loadings (W*) in blue and Y-space loadings (Q) in red on a shared categorical axis, separated by a vertical dashed divider. Without ScaleToPlot, a dual y-axis is used so both X and Y loadings are readable regardless of scale differences. With ScaleToPlot, both are normalised to [-1, 1] and share a single y-axis.

All charts are written to a single HTML file.

Parameters:

  • mvm_obj (dict) – Fitted model from pyphi.pca(), pyphi.pls(), pyphi.lpls(), pyphi.jrpls(), or pyphi.tpls().

  • plotwidth (int) – Width of each chart in pixels. Default 750.

  • xgrid (bool) – If True, show vertical grid lines. Default False.

  • addtitle (str) – Optional string appended to each chart title. Default "".

  • material (str) – For JRPLS / TPLS models, the material block to plot. Default None (uses R-space / blend loadings).

  • zspace (bool) – For TPLS models, if True plots Z-space (trajectory) loadings. Default False.

  • ScaleToPlot (bool) – If True, normalises each loading vector to [-1, 1] before plotting, so all LVs are visually comparable. Default False.

Returns:

Saves and opens an HTML file via Bokeh’s show().

Return type:

None

pyphi.plots.loadings_map(mvm_obj: dict, dims: list[int], *, plotwidth: int = 600, addtitle: str = '', material=None, zspace: bool = False, textalpha: float = 0.75) → None[source]

Scatter plot of loadings on two selected components (loadings map / biplot).

For PLS models, X-space loadings (W*, blue) and Y-space loadings (Q, red) are overlaid on a normalised [-1, 1] axis, making it easy to identify which X variables co-vary with which Y variables. For PCA models, only P loadings are shown.

Parameters:

  • mvm_obj (dict) – Fitted model from pyphi.pca(), pyphi.pls(), pyphi.lpls(), pyphi.jrpls(), or pyphi.tpls().

  • dims (list[int]) – Two component indices to plot, e.g. [1, 2] for the first two LVs/PCs (1-indexed).

  • plotwidth (int) – Width of the plot in pixels. Default 600.

  • addtitle (str) – Optional string appended to the plot title. Default "".

  • material (str) – For JRPLS / TPLS, the material block to use. Default None.

  • zspace (bool) – For TPLS, if True uses Z-space loadings. Default False.

  • textalpha (float) – Transparency of variable name labels (0–1). Default 0.75.

Returns:

Saves and opens an HTML file via Bokeh’s show().

Return type:

None

pyphi.plots.weighted_loadings(mvm_obj: dict, *, plotwidth: int = 750, xgrid: bool = False, addtitle: str = '', material=None, zspace: bool = False, ScaleToPlot: bool = False) → None[source]

Plot loadings weighted by their R² contribution, one chart per LV/PC.

Each loading coefficient is multiplied by the corresponding per-variable R² value, so variables that explain more variance appear larger. This helps identify truly important variables vs. those that load highly but explain little variance.

For PCA models: shows P × R²X (blue).

For PLS models: shows W* × R²X (blue, left axis) and Q × R²Y (red, right axis) on a shared categorical axis with a dual y-axis layout (or single normalised axis with ScaleToPlot).

All charts are written to a single HTML file.

Parameters:

  • mvm_obj (dict) – Fitted model from pyphi.pca(), pyphi.pls(), pyphi.lpls(), pyphi.jrpls(), or pyphi.tpls().

  • plotwidth (int) – Width of each chart in pixels. Default 750.

  • xgrid (bool) – If True, show vertical grid lines. Default False.

  • addtitle (str) – Optional string appended to each chart title. Default "".

  • material (str) – For JRPLS / TPLS, the material block to plot. Default None.

  • zspace (bool) – For TPLS, if True plots Z-space weighted loadings. Default False.

  • ScaleToPlot (bool) – If True, normalises loadings to [-1, 1] before weighting, placing X and Y on a single shared axis. Default False.

Returns:

Saves and opens an HTML file via Bokeh’s show().

Return type:

None

pyphi.plots.vip(mvm_obj: dict, *, plotwidth: int = 600, material=None, zspace: bool = False, addtitle: str = '') → None[source]

Plot Variable Importance in Projection (VIP) scores.

VIP summarises the contribution of each X variable across all LVs, weighted by the R²Y explained by each LV. Variables are sorted from highest to lowest VIP. Only applicable to PLS models (silently returns for PCA models).

Parameters:

  • mvm_obj (dict) – Fitted PLS model from pyphi.pls(), pyphi.lpls(), pyphi.jrpls(), or pyphi.tpls().

  • plotwidth (int) – Width of the plot in pixels. Default 600.

  • material (str) – For JRPLS / TPLS, the material block to use. Default None.

  • zspace (bool) – For TPLS, if True computes VIP in Z-space. Default False.

  • addtitle (str) – Optional string appended to the plot title. Default "".

Returns:

Saves and opens an HTML file via Bokeh’s show().

Return type:

None

pyphi.plots.score_scatter(mvm_obj: dict, xydim: list[int], *, CLASSID: DataFrame | None = None, colorby: str | None = None, Xnew=None, add_ci: bool = False, add_labels: bool = False, add_legend: bool = True, legend_cols: int = 1, addtitle: str = '', plotwidth: int = 600, plotheight: int = 600, rscores: bool = False, material=None, marker_size: int = 7, nbins=None, include_model: bool = False) → None[source]

Score scatter plot for two selected latent variables or principal components.

Plots training scores (or scores for new observations) on two chosen components. Supports class-based colouring, confidence ellipses, observation labels, and overlaying new observations on top of training data.

Parameters:

  • mvm_obj (dict) – Fitted model from pyphi.pca(), pyphi.pls(), pyphi.lpls(), pyphi.jrpls(), or pyphi.tpls().

  • xydim (list[int]) – Two component indices to plot, e.g. [1, 2] (1-indexed).

  • CLASSID (pd.DataFrame) – Optional class-membership DataFrame. First column should be observation IDs; colorby specifies the column to colour by. Default None.

  • colorby (str) – Column name in CLASSID to use for colouring. Required if CLASSID is provided. Default None.

  • Xnew (pd.DataFrame or np.ndarray) – New observations to project and plot. If None, training scores are used. Not supported for LPLS / JRPLS models. Default None.

  • add_ci (bool) – If True, overlays 95% (gold) and 99% (red) confidence ellipses. Default False.

  • add_labels (bool) – If True, annotates each point with its observation ID. Default False.

  • add_legend (bool) – If True and CLASSID is provided, adds a click-to-hide legend. Default True.

  • legend_cols (int) – Number of legend columns when using CLASSID. Default 1.

  • addtitle (str) – Optional string appended to the plot title. Default "".

  • plotwidth (int) – Width of the plot in pixels. Default 600.

  • plotheight (int) – Height of the plot in pixels. Default 600.

  • rscores (bool) – For LPLS / JRPLS / TPLS models, if True plots R-space scores instead of X-space scores. Default False.

  • material (str) – For JRPLS / TPLS, the material block to plot when rscores=True. Default None (plots all materials, coloured by material name).

  • marker_size (int) – Marker size in pixels. Default 7.

  • nbins (int) – If provided, bins a continuous colorby variable into nbins groups before colouring. Default None.

  • include_model (bool) – If True and Xnew is provided, overlays training set scores (grey, labelled “Model”) alongside the new observations. Default False.

Returns:

Saves and opens an HTML file via Bokeh’s show().

Return type:

None

pyphi.plots.score_line(mvmobj: dict, dim, *, CLASSID: DataFrame | None = None, colorby: str | None = None, Xnew=None, add_ci: bool = False, add_labels: bool = False, add_legend: bool = True, plotline: bool = True, plotwidth: int = 600, plotheight: int = 600) → None[source]

Score line plot: one component plotted against observation sequence.

Useful for monitoring or batch data where the observation order is meaningful. When CLASSID is provided, markers are coloured by class with a single grey connecting line underneath.

Parameters:

  • mvmobj (dict) – Fitted model from pyphi.pca() or pyphi.pls().

  • dim (int or list[int]) – Component index to plot (1-indexed). Pass a single integer or a one-element list, e.g. 1 or [1].

  • CLASSID (pd.DataFrame) – Optional class-membership DataFrame. First column is observation IDs; colorby specifies the column used for colouring. Default None.

  • colorby (str) – Column name in CLASSID to colour by. Required if CLASSID is provided. Default None.

  • Xnew (pd.DataFrame or np.ndarray) – New observations to project and plot. If None, training scores are used. Default None.

  • add_ci (bool) – If True, overlays 95% (gold) and 99% (red) univariate confidence limits as dashed horizontal lines. Default False.

  • add_labels (bool) – If True, annotates each point with its observation ID. Default False.

  • add_legend (bool) – If True and CLASSID is provided, adds a click-to-hide legend. Default True.

  • plotline (bool) – If True, draws a line connecting the observations in sequence. Default True.

  • plotwidth (int) – Width of the plot in pixels. Default 600.

  • plotheight (int) – Height of the plot in pixels. Default 600.

Returns:

Saves and opens an HTML file via Bokeh’s show().

Return type:

None

pyphi.plots.diagnostics(mvmobj: dict, *, Xnew=None, Ynew=None, score_plot_xydim=None, plotwidth: int = 600, ht2_logscale: bool = False, spe_logscale: bool = False) → None[source]

Hotelling’s T² and SPE diagnostic dashboard.

Produces a column of linked plots:

  1. Hotelling’s T² vs. observation sequence, with 95% (gold) and 99% (red) control limits.

  2. SPE X-space vs. observation sequence, with control limits.

  3. Outlier map: T² vs. SPE X, with 99% limits as crosshairs.

  4. SPE Y-space (PLS models only, when Ynew is supplied or using training data).

  5. Score scatter (optional, if score_plot_xydim is provided).

Parameters:

  • mvmobj (dict) – Fitted model from pyphi.pca() or pyphi.pls().

  • Xnew (pd.DataFrame or np.ndarray) – New X observations to diagnose. If None, training set diagnostics are plotted. Default None.

  • Ynew (pd.DataFrame or np.ndarray) – New Y observations (optional). Used only for PLS models to compute Y-space SPE. Default None.

  • score_plot_xydim (list[int]) – If provided (e.g. [1, 2]), appends a score scatter plot for the specified components, with confidence ellipses overlaid. Default None.

  • plotwidth (int) – Width of each plot in pixels. Default 600.

  • ht2_logscale (bool) – If True, plots Hotelling’s T² on a log₁₀ scale. Default False.

  • spe_logscale (bool) – If True, plots SPE X on a log₁₀ scale. Default False.

Returns:

Saves and opens an HTML file via Bokeh’s show().

Return type:

None

pyphi.plots.predvsobs(mvmobj: dict, X, Y, *, CLASSID: DataFrame | None = None, colorby: str | None = None, x_space: bool = False) → None[source]

Observed vs. predicted scatter plots for all Y (and optionally X) variables.

One square scatter plot is produced per variable, with a 45° reference line (cyan, dashed). All plots are stacked vertically in a single HTML file.

Parameters:

  • mvmobj (dict) – Fitted model from pyphi.pca() or pyphi.pls().

  • X (pd.DataFrame, np.ndarray, or dict) – Predictor data. For multi-block models pass a dict of DataFrames keyed by block name.

  • Y (pd.DataFrame or np.ndarray) – Observed response values. Must have the same number of rows as X.

  • CLASSID (pd.DataFrame) – Optional class-membership DataFrame for colouring points. First column is observation IDs. Default None.

  • colorby (str) – Column in CLASSID to use for colouring. Required if CLASSID is provided. Default None.

  • x_space (bool) – If True, also plots observed vs. predicted for each X variable (X-space reconstruction). For PCA models this is always True. Default False.

Returns:

Saves and opens an HTML file via Bokeh’s show().

Return type:

None

pyphi.plots.contributions_plot(mvmobj: dict, X, cont_type: str, *, Y=False, from_obs=None, to_obs=None, lv_space=False, plotwidth: int = 800, plotheight: int = 600, xgrid: bool = False) → None[source]

Bar chart of variable contributions to a monitoring statistic.

Shows how much each variable contributes to the T², SPE, or score of a specific observation (or the difference between two observations). For SPE with PLS models, separate X and Y contribution bars are shown.

Parameters:

  • mvmobj (dict) – Fitted model from pyphi.pca() or pyphi.pls().

  • X (pd.DataFrame, np.ndarray, or dict) – Data containing the observation(s) to diagnose. For multi-block models, pass a dict of DataFrames.

  • cont_type (str) – Type of contribution to compute and plot. 'scores': contribution to each latent variable score. 'spe': contribution to SPE (X and, for PLS, Y). 't2': contribution to Hotelling’s T².

  • Y (pd.DataFrame or np.ndarray) – Y data. Required for cont_type='spe' with a PLS model. Default False (not provided).

  • from_obs (int, str, or list) – Reference observation for difference-based contributions. Can be an integer index, an observation ID string, or a list thereof. If None, contributions are computed absolutely for to_obs. Default None.

  • to_obs (int, str, or list) – The observation(s) to diagnose. Can be an integer index, observation ID string, or list thereof. This argument is required. Default None.

  • lv_space (bool) – If True, computes contributions in latent variable space. Default False.

  • plotwidth (int) – Width of the plot in pixels. Default 800.

  • plotheight (int) – Height of the plot in pixels. Default 600.

  • xgrid (bool) – If True, shows vertical grid lines. Default False.

Returns:

Saves and opens an HTML file via Bokeh’s show().

Return type:

None

pyphi.plots.mb_weights(mvmobj: dict, *, plotwidth: int = 600, plotheight: int = 400) → None[source]

Bar chart of super-weights (block weights) for a Multi-Block PLS model.

Shows the weight of each X block on each latent variable, indicating which blocks drive the model. One bar chart is produced per LV, stacked vertically in a single HTML file.

Parameters:

  • mvmobj (dict) – Fitted MBPLS model from pyphi.mbpls(). Must contain the 'Wt' (super-weights) and 'Xblocknames' keys.

  • plotwidth (int) – Width of each plot in pixels. Default 600.

  • plotheight (int) – Height of each plot in pixels. Default 400.

Returns:

Saves and opens an HTML file via Bokeh’s show().

Return type:

None

pyphi.plots.mb_r2pb(mvmobj: dict, *, plotwidth: int = 600, plotheight: int = 400) → None[source]

Stacked bar chart of R² per block per LV for a Multi-Block PLS model.

Each bar represents an X block; colour bands show the contribution of each LV to that block’s explained variance.

Parameters:

  • mvmobj (dict) – Fitted MBPLS model from pyphi.mbpls(). Must contain the 'r2pbX' and 'Xblocknames' keys.

  • plotwidth (int) – Width of the plot in pixels. Default 600.

  • plotheight (int) – Height of the plot in pixels. Default 400.

Returns:

Saves and opens an HTML file via Bokeh’s show().

Return type:

None

pyphi.plots.mb_vip(mvmobj: dict, *, plotwidth: int = 600, plotheight: int = 400) → None[source]

Bar chart of block-level VIP scores for a Multi-Block PLS model.

Computes a VIP score for each X block (using block super-weights and R²Y per LV) and displays them sorted from highest to lowest.

Parameters:

  • mvmobj (dict) – Fitted MBPLS model from pyphi.mbpls(). Must contain the 'Wt', 'r2y', and 'Xblocknames' keys.

  • plotwidth (int) – Width of the plot in pixels. Default 600.

  • plotheight (int) – Height of the plot in pixels. Default 400.

Returns:

Saves and opens an HTML file via Bokeh’s show().

Return type:

None

pyphi.plots.barplot(yheights, xtick_labels: list[str], *, plotwidth: int = 600, plotheight: int = 600, addtitle: str = '', xlabel: str = '', ylabel: str = '', tabtitle: str = 'Bar Plot') → None[source]

Generic vertical bar plot.

A simple utility for plotting any set of values as a bar chart with custom tick labels.

Parameters:

  • yheights (list or np.ndarray) – Bar heights (one value per bar).

  • xtick_labels (list[str]) – Category labels for the x-axis. Length must match yheights.

  • plotwidth (int) – Width of the plot in pixels. Default 600.

  • plotheight (int) – Height of the plot in pixels. Default 600.

  • addtitle (str) – Plot title. Default "".

  • xlabel (str) – X-axis label. Default "".

  • ylabel (str) – Y-axis label. Default "".

  • tabtitle (str) – Browser tab / HTML file title. Default "Bar Plot".

Returns:

Saves and opens an HTML file via Bokeh’s show().

Return type:

None

pyphi.plots.lineplot(X, *, ids_2_include=None, x_axis=None, plot_title: str = 'Main Title', tab_title: str = 'Tab Title', xaxis_label: str = 'X-axis', yaxis_label: str = '', plotheight: int = 400, plotwidth: int = 600, legend_cols: int = 1, linecolor: str = 'blue', linewidth: int = 2, add_marker: bool = False, individual_plots: bool = False, add_legend: bool = True, markercolor: str = 'darkblue', markersize: int = 10, fill_alpha: float = 0.2, line_alpha: float = 0.4, ncx_x_col=None, ncx_y_col=None, ncx_id_col=None, CLASSID: DataFrame | None = None, colorby: str | None = None, yaxis_log: bool = False) → None[source]

Multi-series line plot from a DataFrame or a list of DataFrames.

Supports two input modes:

Common x-axis (X is a pd.DataFrame): Each selected column is plotted as a separate line against a shared x-axis (either the row index or a custom x_axis array).

Non-common x-axis (X is a list of DataFrames): Each DataFrame represents one series with its own x-values. Use ncx_x_col, ncx_y_col, and ncx_id_col to specify the column names for x-values, y-values, and series IDs respectively.

Parameters:

  • X (pd.DataFrame or list[pd.DataFrame]) – Data to plot.

  • ids_2_include (list[str] or str) – Column names (common x-axis mode) or series ID values (non-common x-axis mode) to include. If None, all columns / series are included. Default None.

  • x_axis (list or np.ndarray) – Custom x-axis values for common x-axis mode. If None, observation sequence numbers are used. Default None.

  • plot_title (str) – Main plot title. Default "Main Title".

  • tab_title (str) – Browser tab / HTML file title. Default "Tab Title".

  • xaxis_label (str) – X-axis label. Default "X-axis".

  • yaxis_label (str) – Y-axis label. Default "".

  • plotheight (int) – Height of the plot in pixels. Default 400.

  • plotwidth (int) – Width of the plot in pixels. Default 600.

  • legend_cols (int) – Number of side-by-side legend columns. Default 1.

  • linecolor (str) – Line colour when a single individual_plots line is drawn. Default "blue".

  • linewidth (int) – Line width in pixels. Default 2.

  • add_marker (bool) – If True, overlays scatter markers on each line. Default False.

  • individual_plots (bool) – If True, produces one separate plot per series stacked vertically. If False, all series are overlaid on a single plot. Default False.

  • add_legend (bool) – If True, adds a click-to-hide legend. Default True.

  • markercolor (str) – Marker fill colour when add_marker=True. Default "darkblue".

  • markersize (int) – Marker size in pixels. Default 10.

  • fill_alpha (float) – Marker fill transparency (0–1). Default 0.2.

  • line_alpha (float) – Line transparency (0–1). Default 0.4.

  • ncx_x_col (str) – Column name for x-values in non-common x-axis mode. Default None.

  • ncx_y_col (str) – Column name for y-values in non-common x-axis mode. Default None.

  • ncx_id_col (str) – Column name used to identify each series in non-common x-axis mode. Default None.

  • CLASSID (pd.DataFrame) – Optional class-membership DataFrame. First column is series IDs; colorby specifies the column to colour by. Only used in common x-axis mode. Default None.

  • colorby (str) – Column in CLASSID to use for colouring. Required if CLASSID is provided. Default None.

  • yaxis_log (bool) – If True, uses a logarithmic y-axis scale. Default False.

Returns:

Saves and opens an HTML file via Bokeh’s show().

Return type:

None

pyphi.plots.plot_spectra(X, *, xaxis=None, plot_title: str = 'Main Title', tab_title: str = 'Tab Title', xaxis_label: str = 'X-axis', yaxis_label: str = 'Y-axis', linecolor: str = 'blue', linewidth: int = 2) → None[source]

Plot spectral data: one line per sample (row).

Draws all spectra as a multi-line plot where each row of X is a separate spectrum. The x-axis values are taken from the DataFrame column headers (if a DataFrame) or from xaxis (if an ndarray).

Parameters:

  • X (pd.DataFrame or np.ndarray) – Spectral matrix (n_samples × n_wavelengths). If a DataFrame, the first column must contain sample IDs and the remaining column headers are used as x-axis tick values.

  • xaxis (list or np.ndarray) – Custom x-axis values (e.g. wavelengths). Only used when X is an ndarray. If None, sequential integers are used. Default None.

  • plot_title (str) – Main plot title. Default "Main Title".

  • tab_title (str) – Browser tab / HTML file title. Default "Tab Title".

  • xaxis_label (str) – X-axis label (e.g. "Wavelength (nm)"). Default "X-axis".

  • yaxis_label (str) – Y-axis label (e.g. "Absorbance"). Default "Y-axis".

  • linecolor (str) – Colour applied to all spectral lines. Default "blue".

  • linewidth (int) – Line width in pixels. Default 2.

Returns:

Saves and opens an HTML file via Bokeh’s show().

Return type:

None

pyphi.plots.scatter_with_labels(x, y, *, xlabel: str = 'X var', ylabel: str = 'Y var', labels=None, tabtitle: str = 'Scatter Plot', plottitle: str = 'Scatter', legend_cols: int = 1, CLASSID: DataFrame | None = None, colorby: str | None = None, plotwidth: int = 600, plotheight: int = 600, markercolor: str = 'darkblue', markersize: int = 10, fill_alpha: float = 0.2, line_alpha: float = 0.4) → None[source]

Generic scatter plot with hover labels and optional class colouring.

A flexible utility for plotting any two arrays against each other, with per-point labels shown on hover and optional class-based colouring. Origin lines (x=0, y=0) are added automatically.

Parameters:

  • x (list or np.ndarray) – X coordinates of each point.

  • y (list or np.ndarray) – Y coordinates of each point. Must have the same length as x.

  • xlabel (str) – X-axis label. Default "X var".

  • ylabel (str) – Y-axis label. Default "Y var".

  • labels (list[str]) – Per-point labels shown in the hover tooltip. If None, points are labelled "Obs 0", "Obs 1", etc. Default None.

  • tabtitle (str) – Browser tab / HTML file title. Default "Scatter Plot".

  • plottitle (str) – Main plot title. Default "Scatter".

  • legend_cols (int) – Number of side-by-side legend columns when CLASSID is used. Default 1.

  • CLASSID (pd.DataFrame) – Optional class-membership DataFrame for colouring points. First column is point IDs. Default None.

  • colorby (str) – Column in CLASSID to colour by. Required if CLASSID is provided. Default None.

  • plotwidth (int) – Width of the plot in pixels. Default 600.

  • plotheight (int) – Height of the plot in pixels. Default 600.

  • markercolor (str) – Marker colour when no CLASSID is provided. Default "darkblue".

  • markersize (int) – Marker size in pixels. Default 10.

  • fill_alpha (float) – Marker fill transparency (0–1). Default 0.2.

  • line_alpha (float) – Marker edge transparency (0–1). Default 0.4.

Returns:

Saves and opens an HTML file via Bokeh’s show().

Return type:

None