Python Visualisations Repo

What is this?

This is a repository of helpful snippets to generate illustrations in python.

Repo Index

1. Annotated Heat Maps
2. Polar Plots Redux

Annotated Heatmaps

Create annotated heatmaps that have scaled, and unscaled values on each line.

Sample code here

if scale:
  df_copy=df.copy()
  ss = MinMaxScaler()
  scaled_features = ss.fit_transform(df) # scaling is done to allow comparisons across features
  df[df.columns] = scaled_features
  
  # Supplementary text in the case where scaling is used
  sup_text = '\nValues are normalized based on the minimum/maximum values across the week for each Facility.'

  # labels in each cell
  labels = (np.asarray(["Scaled: {0:.2f} \nOriginal: {1:.0f}".format(original, scaled)
                for original, scaled in zip((df.to_numpy()).flatten(),
                                           (df_copy.to_numpy()).flatten())]
   )).reshape(df.shape).T
  
else:
  sup_text = ''
        
font_title = {'family': 'sans-serif',
    'weight': 'bold',
    'size': 16,
    }
# We use ax parameter to tell seaborn which subplot to use for this plot

fig = plt.figure(figsize=(15,9))
# generate our heatmap
ax = sns.heatmap(df.T, cmap="PuBu", annot=labels if scale else True, fmt='' if scale else '.0f',cbar=False)
ax.tick_params(axis='both', which='both', length=0)
ax.set_xlabel('Day of Week')
ax.set_ylabel('Facility Name')

# Main title
ax.text(x=0, y=1.1, s=f'Visualizing Facilities in {monthDict[month]}', fontsize=16, weight='bold', ha='left', va='bottom', transform=ax.transAxes)

# Supplementary text just below the main title
ax.text(x=0, y=1.05, s='Distribution of Utilization throughout the week (Mondays - Fridays).' + sup_text, fontsize=8, alpha=0.75, ha='left', va='bottom', transform=ax.transAxes)

plt.show()

Polar Plots Redux

Plotting events across a clock-like dimensional axis (Polar)