finds.utils.plots

Convenience functions for data plotting

• chart types: date axis, time axis, confidence bands, bar, hist, scatter

• regression diagnostics

Copyright 2023, Terence Lim

MIT License

finds.utils.plots.ColorMap(n, colormap='brg')

Return sampled colors from spectrum of a color map :param n: number of discrete colors :param colormap: matplotlib colormaps, e.g. tab20, viridis, Blues, cool

Returns:

n x 4 2D-array of RGBA values

finds.utils.plots.plot_bands(mean: Series, stderr: Series, width: float = 1.96, x: List[int] = [], ylabel: str = '', xlabel: str = '', c: str = 'b', loc: str = 'best', legend: List[str] = [], ax: Any = None, fontsize: int = 10, title: str = '', hline: List = [], vline: List = [])

Line plot a series with confidence bands

Parameters:

• mean – Mean values to plot

• stderr – Stderr values to plot confidence bands

• width – Multipler on stderr for confidence bands

• c – Color to fill bands

• x – X-axis values

• ylabel – Axis labels

• xlabel – Axis labels

• legend – List of legend labels

• loc – Location to display legend

• ax – Axis object

• fontsize – Base font size

• title – Main title string

• hline – List of y-axis values to plot horizontal lines

• vline – List of x-axis values to plot vertical lines

finds.utils.plots.plot_bar(y: DataFrame, ax: Any = None, labels: List[str] = [], xlabel: str = '', ylabel: str = '', fontsize: int = 12, title: str = '', legend: List[str] = [], loc: str = 'best', alpha: float = 0.8, labelsize: int = 8, rotation: float = 0.0)

Bar plot with annotated points

Parameters:

• y – DataFrame of y-values, observations in rows, variables in columns

• ax – Axis object

• labels – List of labels to annotate

• labelsize – Font size of annotation labels text

• rotation – Rotate annotation labels text

• xlabel – Axis labels

• ylabel – Axis labels

• fontsize – Base font size

• alpha – Transparency of bars

• title – Main title text

• legend – List of legend names

• loc – Location for legend

finds.utils.plots.plot_date(y1: DataFrame, y2: DataFrame | None = None, ax: Any = None, xmin: int = 0, xmax: int = 99999999, cn: int = 0, fontsize: int = 12, rescale: bool = False, yscale: bool = False, ls: str = '-', marker: str | None = '', ms: int | None = None, hlines: List[float] = [], vlines: List[int] = [], nbins: int = 0, vspans: List[Tuple[int, int]] = [], xlabel: str = '', points: DataFrame | Series | None = None, rotation: float = 0, title: str = '', ylabel1: str = '', ylabel2: str = '', legend1: List[str] = [], legend2: List[str] = [], loc1: str = 'upper left', loc2: str = 'upper right')

Line plot with int date on x-axis, and primary and secondary y frames

Parameters:

• y1 – Plot on primary y-axis

• y2 – Plot on secondary y-axis

• ax – Matplotlib axes object from plt.subplots() or plt.gca()

• cn – Starting CN color to cycle through

• marker – Marker style, None to cycle (default ‘’ means no marker)

• xmin – Minimum of x-axis date range

• xmax – Maximum of x-axis date range (default is auto)

• nbins – Number of bins for xticks placement (default 0 is auto)

• rotation – Rotation of x-axis ticks

• hlines – Y-axis points where to place horizontal lines

• vlines – X-axis points where to place vertical lines

• vspans – Vertical regions to highlight

• xlabel – X-axis label

• ylabel1 – Y-axis labels

• ylabel2 – Y-axis labels

• fontsize – Base font size

• points – Points and labels to annotate

• title – Main title

• legend1 – Lists of legend labels

• legend2 – Lists of legend labels

• loc1 – Locations to place legends

• loc2 – Locations to place legends

• ls – Linestyle

• ms – Marker size

finds.utils.plots.plot_fitted(fitted: Series, resid: Series, n: int = 3, ax: Any = None, title: str = 'Residuals vs Fitted', fontsize: int = 12, strftime: str = '%Y-%m-%d') → Series

Plot residuals and identify outliers

Parameters:

• ax – Axis object

• fitted – Fitted Series

• residual – Residual Series

• n – Number of outlier points in each end to identify

• strftime – string to format time display

• title – Main title text

• fontsize – Base font size

finds.utils.plots.plot_groupbar(df: DataFrame, ax: Any = None, labels: DataFrame | None = None)

Plot a grouped bar chart

Parameters:

• df – DataFrame to plot

• ax – Axis object

• labels – optional DataFrame of annotations

finds.utils.plots.plot_hist(*args, kde: bool = True, hist: bool = False, bins: ~typing.List[float] = [], pdf: ~typing.Callable | ~typing.List | ~typing.Dict = <bound method rv_continuous.pdf of <scipy.stats._continuous_distns.norm_gen object>>, ax: ~typing.Any = None, title: str = '', xlabel: str = '', ylabel: str = 'density', fontsize: int = 12)

Histogram bar plot with a benchmark probability density

Parameters:

• ax – Axis object

• bins – List of bin values

• hist – Plots histogram

• kde – Plots kernel density curve

• pdf – Benchmark probability density

• ylabel – Axis labels

• xlabel – Axis labels

• fontsize – Base font size

• title – Main title text

finds.utils.plots.plot_leverage(resid: Series, hat: array, dist: array, ddof: int, title: str = 'Residuals vs Leverage', ax: Any = None, strftime='%Y-%m-%d') → DataFrame

Plot leverage and identify influential points

Parameters:

• resid – Residual Series

• hat – Hat values

• dist – Distance values

• ddof – Degrees of freedom of model

• ax – Axis object

• title – Main title text

• strftime – string to format time display

finds.utils.plots.plot_qq(resid: Series, title: str = 'Normal Q-Q', ax: Any = None, z: float = 2.807, strftime: str = '%Y-%m-%d') → DataFrame

QQ Plot

Parameters:

• resid – Residual Series

• strftime – string to format time display

• title – Main title text

• z – Z-value to identify outliers

finds.utils.plots.plot_scale(fitted: Series, resid: Series, ax: Any = None, title: str = 'Scale-Location', n: int = 3, alpha: float = 0.5, strftime: str = '%Y-%m-%d')

Plot scale of residuals with outliers

Parameters:

• fitted – Fitted Series

• resid – Residual Series

• ax – Axis object

• title – Main title text

• alpha – Transparency of points

• n – number of outliers in each end to identify

• strftime – string to format time display

finds.utils.plots.plot_scatter(x: Series, y: Series, labels: List = [], ax: Any = None, xlabel: str = '', ylabel: str = '', c: Any = None, cmap: Any = None, alpha: float = 0.75, edgecolor: Any = None, s: float = 10, marker: str = 'o', title: str = '', abline: bool | None = True, fontsize: int = 12)

Scatter plot, optionally with abline slope and point labels

Parameters:

• x – Series to plot on horizontal axis

• y – Series to plot on horizontal axis

• labels – List of annotations for points

• ax – Matplotlib axes object, from plt.subplots() or plt.gca()

• xlabel – Horizontal axis label

• ylabel – Vertical axis label

• title – Title of plot

• abline – To plot abline or 45-degree line. If none, do not plot slope

• labels – List of 3-tuples (text, x, y) to annotate

• alpha – transparency of scatter points

• edgecolor – edge color of scatter points

• marker – marker type of scatter points

• s – Marker area size

• cmap – Color map to use for scatter points

• abline – True for fitted slope, False for 45-degree line, None is no plot

• fontsize – Base font size

• title – Main title string

finds.utils.plots.plot_time(y1: DataFrame, y2: DataFrame | None = None, ax: Any = None, xmin: Timestamp = Timestamp('1900-01-01 09:30:00'), xmax: Timestamp = Timestamp('1900-01-01 16:00:00'), title: str = '', marker: str = ' ', fontsize: int = 8)

Plot lines with Timestamp time on x-axis; primary and secondary y-axis

Parameters:

• y1 – DataFrame to plot on left axis

• y2 – DataFrame (or None) to plot on right axis

• ax – Axes object to plot in

• xmin – Left-most x-axis time, None to include all

• xmax – Right-most x-axis time, None to include all

• marker – Style of marker to plot

• fontsize – Font size of tick labels

finds.utils.plots.set_xticks(ax: Axes, nbins: int = 0, nskip: int = 0, **kwargs)

Set locations for xticks

Parameters:

• ax – Matplotlib axes object from plt.subplots() or plt.gca()

• nbins – Number of bins for major ticks, or

• nskip – Number of ticks to skip (ignore if nbins > 0)

• **kwargs – Arguments for tick_params(), e.g. labelsize, colors, rotation

finds.utils.plots.subplots(nfigs: int = 1, ncols: int = 1, nrows: int = 1, squeeze: bool = False, **kwargs) → List[Axes]

Wrapper over plt.subplots to create subplots across multiple figures

Parameters:

• nfigs – Number of figures

• ncols – Number of columns per figure

• nrows – Number of rows per figure

• squeeze – ignored, will always return 1D flattened list of axes unless only 1

• **kwargs – other keyword arguments passed on to plt.subplots

Returns:

Tuple of (List of figures, List of axes)