More Complex Visualizations

Tutorial: Creating More Complex Visualizations with ScanPyImports

Here's an example showcasing the packages I frequently use in my projects. I'm sharing it to provide ideas on how to create more complex visualizations using ScanPyImports, Matplotlib and WordCloud.

Feel free to customize these visualizations to suit your needs and apply the code to your own projects.

Importing Libraries¶

In [1]:

from ScanPyImports.plotter import DataPlotter
import matplotlib.pyplot as plt
from matplotlib.colors import to_rgb
from matplotlib.patches import FancyBboxPatch
import math
import random
import warnings

# Ignore warnings for clean output
warnings.filterwarnings("ignore")

from ScanPyImports.plotter import DataPlotter import matplotlib.pyplot as plt from matplotlib.colors import to_rgb from matplotlib.patches import FancyBboxPatch import math import random import warnings # Ignore warnings for clean output warnings.filterwarnings("ignore")

Setting Up Constants and Utility Functions¶

Next, we define some constants and utility functions that will help us in creating the visualizations.

In [2]:

# Constants for styling
ROUNDING = 0.1
BOXSTYLE_BG = f'round, pad=0.01, rounding_size={ROUNDING}'

# Add custom background to a plot (Axes)
def box_bg(ax, color):
    box = FancyBboxPatch((0, 0), 1, 1, boxstyle=BOXSTYLE_BG,
                          ec='none', fc=color, clip_on=False, transform=ax.transAxes, zorder=0)
    ax.add_patch(box)

# Define a color palette
def tab20() -> dict[str, tuple]: 
    """Tableau 20 colors hex codes. 

    Returns:
        Dictionary {color name: (dark, light)}
    """
    colors = {
        'blue': ('#1f77b4', '#aec7e8'),
        'orange': ('#ff7f0e', '#ffbb78'),
        'green': ('#2ca02c', '#98df8a'),
        'red': ('#d62728', '#ff9896'),
        'purple': ('#9467bd', '#c5b0d5'),
        'brown': ('#8c564b', '#c49c94'),
        'pink': ('#e377c2', '#f7b6d2'),
        'gray': ('#7f7f7f', '#c7c7c7'),
        'olive': ('#bcbd22', '#dbdb8d'),
        'cyan': ('#17becf', '#9edae5')
    }
    return colors


# Generate the colors to be used for the words in the Cloud plot
def words_colors():
    """Colors in format `rgb(r, g, b)` as required by the WordCloud object."""
    colors = tab20()
    light_colors = [to_rgb(val[1]) for _, val in colors.items()]
    light_colors += [(1.0,) * 3] * 10  # Fill with more whites
    RGB = lambda x: math.floor(x * 255)
    colorsRGB = [f'rgb{(RGB(a), RGB(b), RGB(c))}' for a, b, c in light_colors]
    return colorsRGB

# Function to randomly choose a color for the Cloud plot
def color_func(*args, **kwargs):
    """Function to be passed as `**kargs` of the Cloud plot.
    
    It follows the requirements of the WordCloud Api."""
    colors = words_colors()
    return random.choice(colors)

# Constants for styling ROUNDING = 0.1 BOXSTYLE_BG = f'round, pad=0.01, rounding_size={ROUNDING}' # Add custom background to a plot (Axes) def box_bg(ax, color): box = FancyBboxPatch((0, 0), 1, 1, boxstyle=BOXSTYLE_BG, ec='none', fc=color, clip_on=False, transform=ax.transAxes, zorder=0) ax.add_patch(box) # Define a color palette def tab20() -> dict[str, tuple]: """Tableau 20 colors hex codes. Returns: Dictionary {color name: (dark, light)} """ colors = { 'blue': ('#1f77b4', '#aec7e8'), 'orange': ('#ff7f0e', '#ffbb78'), 'green': ('#2ca02c', '#98df8a'), 'red': ('#d62728', '#ff9896'), 'purple': ('#9467bd', '#c5b0d5'), 'brown': ('#8c564b', '#c49c94'), 'pink': ('#e377c2', '#f7b6d2'), 'gray': ('#7f7f7f', '#c7c7c7'), 'olive': ('#bcbd22', '#dbdb8d'), 'cyan': ('#17becf', '#9edae5') } return colors # Generate the colors to be used for the words in the Cloud plot def words_colors(): """Colors in format `rgb(r, g, b)` as required by the WordCloud object.""" colors = tab20() light_colors = [to_rgb(val[1]) for _, val in colors.items()] light_colors += [(1.0,) * 3] * 10 # Fill with more whites RGB = lambda x: math.floor(x * 255) colorsRGB = [f'rgb{(RGB(a), RGB(b), RGB(c))}' for a, b, c in light_colors] return colorsRGB # Function to randomly choose a color for the Cloud plot def color_func(*args, **kwargs): """Function to be passed as `**kargs` of the Cloud plot. It follows the requirements of the WordCloud Api.""" colors = words_colors() return random.choice(colors)

Fancy Spiral Plot Function¶

We define a function fancy_spiral to create a spiral plot with custom foreground and background colors.

In [3]:

def fancy_spiral(plot, ax, fg_color, bg_color, title_text= None,  fontname='Arial Rounded MT Bold'):
    # Create a spiral plot
    kwargs = dict(edgecolor=bg_color, facecolor=fg_color)
    _, ax, _, texts = plot.spiral_frequencies(top=25,
                                              label_padding=5,
                                              ax=ax,
                                              **kwargs)
    
    # Adjust the properties of the bar labels
    for text in texts:
        text.set_fontsize('x-large')
        text.set_color(fg_color)
        text.set_font(fontname)

    # Set the title properties
    if title_text:
        ax.set_title(title_text
                        , y = 0.9, x = 0.1
                        , ha = 'left', va = 'top'
                        , color = bg_color
                        , fontsize = 'x-large'
                        , fontweight = 'bold'
                        , fontname = fontname
                        , bbox = dict(facecolor=fg_color # Background color of the title box
                                    , edgecolor='none'
                                    , boxstyle='round,pad=1') 
                        )

    # Add a background round box to the plot
    box_bg(ax, bg_color)
    
    return  ax, texts

def fancy_spiral(plot, ax, fg_color, bg_color, title_text= None, fontname='Arial Rounded MT Bold'): # Create a spiral plot kwargs = dict(edgecolor=bg_color, facecolor=fg_color) _, ax, _, texts = plot.spiral_frequencies(top=25, label_padding=5, ax=ax, **kwargs) # Adjust the properties of the bar labels for text in texts: text.set_fontsize('x-large') text.set_color(fg_color) text.set_font(fontname) # Set the title properties if title_text: ax.set_title(title_text , y = 0.9, x = 0.1 , ha = 'left', va = 'top' , color = bg_color , fontsize = 'x-large' , fontweight = 'bold' , fontname = fontname , bbox = dict(facecolor=fg_color # Background color of the title box , edgecolor='none' , boxstyle='round,pad=1') ) # Add a background round box to the plot box_bg(ax, bg_color) return ax, texts

Fancy Cloud Plot Function¶

In [4]:

def fancy_cloud(plot, ax, bg_color, fontname='Impact'):

    # Add box background
    box_bg(ax, bg_color) # Has to be added first!

    # Set the font if provided
    if fontname:
        plot.settings.set_font(fontname)

    # Define `**kwargs` parameters for the cloud plot
    kwargs = dict(color_func=color_func, max_words=200)

    # Define the `imshow` parameter.
    # Here we ajust the size of the image to get some space 
    #   between the image and axis
    shrink = 0.10
    imshow = dict(extent=(0 + shrink, 1 - shrink, 0 + shrink, 1 - shrink),
                  transform=ax.transAxes)

    # Generate the cloud plot
    _, ax, wc, _ = plot.cloud_frequencies(ax=ax, imshow=imshow, **kwargs)
    return ax, wc

def fancy_cloud(plot, ax, bg_color, fontname='Impact'): # Add box background box_bg(ax, bg_color) # Has to be added first! # Set the font if provided if fontname: plot.settings.set_font(fontname) # Define `**kwargs` parameters for the cloud plot kwargs = dict(color_func=color_func, max_words=200) # Define the `imshow` parameter. # Here we ajust the size of the image to get some space # between the image and axis shrink = 0.10 imshow = dict(extent=(0 + shrink, 1 - shrink, 0 + shrink, 1 - shrink), transform=ax.transAxes) # Generate the cloud plot _, ax, wc, _ = plot.cloud_frequencies(ax=ax, imshow=imshow, **kwargs) return ax, wc

Setting Up Projects and DataPlotters¶

We initialize a DataPlotter object per project (directories).

In [5]:

# Define the projects to analyze
# These are my project directories, but you can provide your own directories and projects names!
projects = {
'@Graz' : DataPlotter(r"D:\Dropbox\Research\DigitalComp"),
'@WU' : DataPlotter(r'D:\Dropbox\Research\MigrationAT'),
'All' : DataPlotter(r'D:\Dropbox')
}

# Exclude specific packages by setting the `to_exclude` attribute
to_exclude = ['my_scripts', 'my_package', 'my_plots', 'my_module', 'my_sum']
for  plotter in projects.values():
    plotter.to_exclude = to_exclude

# Define the projects to analyze # These are my project directories, but you can provide your own directories and projects names! projects = { '@Graz' : DataPlotter(r"D:\Dropbox\Research\DigitalComp"), '@WU' : DataPlotter(r'D:\Dropbox\Research\MigrationAT'), 'All' : DataPlotter(r'D:\Dropbox') } # Exclude specific packages by setting the `to_exclude` attribute to_exclude = ['my_scripts', 'my_package', 'my_plots', 'my_module', 'my_sum'] for plotter in projects.values(): plotter.to_exclude = to_exclude

Creating the Visualization¶

Finally, we create a mosaic plot with different subplots for each project.

In [6]:

# Define subplot configuration
# Note: The projection for spiral plots must be 'polar'
per_subplot_kw = {key: {'projection': 'polar'} for key in projects.keys()}

# Set up the figure and axes
size = 4.5
fig, axs = plt.subplot_mosaic([['@WU', '@Graz'],
    ['cloud', 'All']], 
    per_subplot_kw=per_subplot_kw, 
    figsize=(size * 2, size * 2), 
    gridspec_kw={
        "bottom": 0.1,
        "top": 0.9,
        "left": 0.1,
        "right": 0.9,
        "wspace": 0.1 * 2 / 3,
        "hspace": 0.1 * 2 / 3
    }
)

# Define colors
colors = tab20()

# Create cloud plot for the project 'All' 
bg_color = colors['blue'][0]
axs['cloud'], wc= fancy_cloud(projects['All'], ax=axs['cloud'],bg_color=bg_color)

# Create spiral plots for each project
fg_color = 'white'
bg_colors = ['gray', 'orange', 'red']
for (project, plot), color  in zip(projects.items(),bg_colors):
    bg_color = colors[color][0]
    axs[project], texts = fancy_spiral(plot, axs[project]
                                , fg_color, bg_color
                                , title_text=project)
    # Manually adjust labels that go out of the plot (`All` spiral) 
    if project == 'All':
        for text in texts[-4:]:
            text.set_horizontalalignment('right') 
            x,y = text.get_position()
            text.set_position((x,y-10))
            text.set_color(bg_color)

# Define subplot configuration # Note: The projection for spiral plots must be 'polar' per_subplot_kw = {key: {'projection': 'polar'} for key in projects.keys()} # Set up the figure and axes size = 4.5 fig, axs = plt.subplot_mosaic([['@WU', '@Graz'], ['cloud', 'All']], per_subplot_kw=per_subplot_kw, figsize=(size * 2, size * 2), gridspec_kw={ "bottom": 0.1, "top": 0.9, "left": 0.1, "right": 0.9, "wspace": 0.1 * 2 / 3, "hspace": 0.1 * 2 / 3 } ) # Define colors colors = tab20() # Create cloud plot for the project 'All' bg_color = colors['blue'][0] axs['cloud'], wc= fancy_cloud(projects['All'], ax=axs['cloud'],bg_color=bg_color) # Create spiral plots for each project fg_color = 'white' bg_colors = ['gray', 'orange', 'red'] for (project, plot), color in zip(projects.items(),bg_colors): bg_color = colors[color][0] axs[project], texts = fancy_spiral(plot, axs[project] , fg_color, bg_color , title_text=project) # Manually adjust labels that go out of the plot (`All` spiral) if project == 'All': for text in texts[-4:]: text.set_horizontalalignment('right') x,y = text.get_position() text.set_position((x,y-10)) text.set_color(bg_color)