subplots

subplots(array=None, ncols=1, nrows=1, ref=1, order='C', aspect=1, figsize=None, width=None, height=None, journal=None, axwidth=None, axheight=None, hspace=None, wspace=None, space=None, hratios=None, wratios=None, width_ratios=None, height_ratios=None, flush=None, wflush=None, hflush=None, left=None, bottom=None, right=None, top=None, tight=None, pad=None, axpad=None, panelpad=None, span=None, spanx=None, spany=None, align=None, alignx=None, aligny=None, share=None, sharex=None, sharey=None, basemap=False, proj=None, projection=None, proj_kw=None, projection_kw=None, autoformat=True, includepanels=False)[source]

Analogous to matplotlib.pyplot.subplots, creates a figure with a single axes or arbitrary grids of axes, any of which can be map projections.

Parameters
  • array (2D array-like of int, optional) – Array specifying complex grid of subplots. Think of this array as a “picture” of your figure. For example, the array [[1, 1], [2, 3]] creates one long subplot in the top row, two smaller subplots in the bottom row. Integers must range from 1 to the number of plots.

    0 indicates an empty space. For example, [[1, 1, 1], [2, 0, 3]] creates one long subplot in the top row with two subplots in the bottom row separated by a space.

  • ncols, nrows (int, optional) – Number of columns, rows. Ignored if array is not None. Use these arguments for simpler subplot grids.

  • order ({‘C’, ‘F’}, optional) – Whether subplots are numbered in column-major ('C') or row-major ('F') order. Analogous to numpy.array ordering. This controls the order axes appear in the axs list, and the order of subplot a-b-c labeling (see format).

  • figsize (length-2 tuple, optional) – Tuple specifying the figure (width, height).

  • width, height (float or str, optional) – The figure width and height. Units are interpreted by units.

  • journal (str, optional) – String name corresponding to an academic journal standard that is used to control the figure width (and height, if specified). See below table.

    Key

    Size description

    Organization

    'pnas1'

    1-column

    Proceedings of the National Academy of Sciences

    'pnas2'

    2-column

    'pnas3'

    landscape page

    'ams1'

    1-column

    American Meteorological Society

    'ams2'

    small 2-column

    'ams3'

    medium 2-column

    'ams4'

    full 2-column

    'agu1'

    1-column

    American Geophysical Union

    'agu2'

    2-column

    'agu3'

    full height 1-column

    'agu4'

    full height 2-column

  • ref (int, optional) – The reference axes number. The axwidth, axheight, and aspect keyword args are applied to this axes, and aspect ratio is conserved for this axes in tight layout adjustment.

  • axwidth, axheight (float or str, optional) – Sets the average width, height of your axes. Units are interpreted by units. Default is rc[‘subplots.axwidth’] = '18em'.

    These arguments are convenient where you don’t care about the figure dimensions and just want your axes to have enough “room”.

  • aspect (float or length-2 list of floats, optional) – The (average) axes aspect ratio, in numeric form (width divided by height) or as (width, height) tuple. If you do not provide the hratios or wratios keyword args, all axes will have identical aspect ratios.

  • hratios, wratios – Aliases for height_ratios, width_ratios.

  • width_ratios, height_ratios (float or list thereof, optional) – Passed to GridSpec, denotes the width and height ratios for the subplot grid. Length of width_ratios must match the number of rows, and length of height_ratios must match the number of columns.

  • wspace, hspace, space (float or str or list thereof, optional) – Passed to GridSpec, denotes the spacing between grid columns, rows, and both, respectively. If float or string, expanded into lists of length ncols-1 (for wspace) or length nrows-1 (for hspace).

    Units are interpreted by units for each element of the list. By default, these are determined by the “tight layout” algorithm.

  • left, right, top, bottom (float or str, optional) – Passed to GridSpec, denotes the width of padding between the subplots and the figure edge. Units are interpreted by units. By default, these are determined by the “tight layout” algorithm.

  • sharex, sharey, share ({3, 2, 1, 0}, optional) – The “axis sharing level” for the x axis, y axis, or both axes. Default is 3. This can considerably redundancy in your figure. The options are as follows:

    1. No axis sharing. Also sets the default spanx and spany values to False.

    2. Only draw axis label on the leftmost column (y) or bottommost row (x) of subplots. Axis tick labels still appear on every subplot.

    3. As in 1, but forces the axis limits to be identical. Axis tick labels still appear on every subplot.

    4. As in 2, but only show the axis tick labels on the leftmost column (y) or bottommost row (x) of subplots.

  • spanx, spany, span (bool or {0, 1}, optional) – Default is False if sharex, sharey, or share are 0, True otherwise. Toggles “spanning” axis labels for the x axis, y axis, or both axes. When True, a single, centered axis label is used for all axes with bottom and left edges in the same row or column. This can considerably redundancy in your figure.

    “Spanning” labels integrate with “shared” axes. For example, for a 3-row, 3-column figure, with sharey > 1 and spany=1, your figure will have 1 ylabel instead of 9.

  • alignx, aligny, align (bool or {0, 1}, optional) – Default is False. Whether to align axis labels for the x axis, y axis, or both axes. Only has an effect when spanx, spany, or span are False.

  • proj, projection (str or dict-like, optional) – The map projection name. The argument is interpreted as follows.

    • If string, this projection is used for all subplots. For valid names, see the Proj documentation.

    • If list of string, these are the projections to use for each subplot in their array order.

    • If dict-like, keys are integers or tuple integers that indicate the projection to use for each subplot. If a key is not provided, that subplot will be a XYAxes. For example, in a 4-subplot figure, proj={2:'merc', (3,4):'stere'} draws a Cartesian axes for the first subplot, a Mercator projection for the second subplot, and a Stereographic projection for the second and third subplots.

  • proj_kw, projection_kw (dict-like, optional) – Keyword arguments passed to Basemap or cartopy Projection classes on instantiation. If dictionary of properties, applies globally. If dictionary of dictionaries of properties, applies to specific subplots, as with proj.

    For example, with ncols=2 and proj_kw={1:{'lon_0':0}, 2:{'lon_0':180}}, the projection in the left subplot is centered on the prime meridian, and the projection in the right subplot is centered on the international dateline.

  • basemap (bool or dict-like, optional) – Whether to use Basemap or Projection for map projections. Default is False. If boolean, applies to all subplots. If dictionary, values apply to specific subplots, as with proj.

Other Parameters
  • tight (bool, optional) – Toggles automatic tight layout adjustments. Default is rc.tight = True.

    If you manually specify a spacing, it will be used to override the tight layout spacing – for example, with left=0.1, the left margin is set to 0.1 inches wide, while the remaining margin widths are calculated automatically.

  • pad, axpad, panelpad (float or str, optional) – Padding for automatic tight layout adjustments. See Figure for details.

  • includepanels (bool, optional) – Whether to include panels when calculating the position of certain spanning labels. See Figure for details.

  • autoformat (bool, optional) – Whether to automatically format axes when special datasets are passed to plotting commands. See Figure for details.

Returns