Generating nice smooth parallel plots!

How to install

Just run

pip install parallelplot

Little Demo on the Wine Quality Dataset

First lets import some packages we need to get some sample data

# Import libraries to handle data 
import numpy as np
import pandas as pd

# The only thing that is really needs to be imported 
# is the plot function from the parallelplot module 
# and the pyplot module from matplotlib to display the plot
import parallelplot.plot as pp
import matplotlib.pyplot as plt


# There is also a module that contains a nice colormap. In addition you can use the matplotlib colormap module
from parallelplot.cmaps import purple_blue
import matplotlib.cm as cm

# Function to download and load the wine quality dataset
def load_wine_quality_dataset():
    # URLs for the Wine Quality datasets 
    red_wine_url = "https://archive.ics.uci.edu/ml/machine-learning-databases/wine-quality/winequality-red.csv"
    white_wine_url = "https://archive.ics.uci.edu/ml/machine-learning-databases/wine-quality/winequality-white.csv"
    
    # Download and read the datasets
    red_wine = pd.read_csv(red_wine_url, sep=';')
    white_wine = pd.read_csv(white_wine_url, sep=';')
    
    # Add a wine type column
    red_wine['wine_type'] = 'red'
    white_wine['wine_type'] = 'white'
    
    # Combine the datasets
    wine_df = pd.concat([red_wine, white_wine], axis=0, ignore_index=True)
    
    return wine_df


wine_df = load_wine_quality_dataset()

print("Wine Quality Dataset:")
wine_df

Wine Quality Dataset:

.dataframe tbody tr th {
    vertical-align: top;
}

.dataframe thead th {
    text-align: right;
}

# Manipulate the dataset to simulate small and large numbers
wine_df["fixed acidity"] = wine_df["fixed acidity"] * 1e6
wine_df["volatile acidity"] = wine_df["volatile acidity"] / 1e6

Create the plots from the imported data!

# Example 1: Basic parallel plot with default style
fig1, axes1 = pp.plot(
    df=wine_df,
    target_column='quality',
    title="Wine Quality Dataset - All Features",
    figsize=(16, 8),
    tick_label_size=10,
    alpha=0.3,
    cmap=cm.hot,
    order='max',
    lw=0.5,
    
)
plt.show()

# Example 2: Parallel plot with dark background
fig2, axes2 = pp.plot(
    df=wine_df,
    target_column='quality',
    title="Wine Quality Dataset - Dark Background",
    figsize=(16, 8),
    style="dark_background",
    lw=0.2,
    # axes_to_reverse = [0, 1, 2, 5]
)
plt.show()

# Example 3: Different cmap 
fig3, axes3 = pp.plot(
    df=wine_df,
    target_column='quality',
    title="Wine Quality Dataset - Colored by Wine Type",
    figsize=(16, 8),
    cmap=purple_blue,
    style="dark_background",
    lw=0.1,
    order='min',
    alpha = 0.2,
    axes_to_reverse = [1,2]
)
plt.show()

# Example 4: Select top features with highest correlation to quality
# Calculate correlations with quality
corr_with_quality = wine_df.drop(columns=['wine_type']).corr()['quality'].abs().sort_values(ascending=False)
top_features = corr_with_quality.index[:8]  # Top 8 features

# Create subset with only the top features
wine_top_features = wine_df[top_features]

fig4, axes4 = pp.plot(
    df=wine_top_features,
    target_column='quality',
    title="Wine Quality - Top Correlated Features",
    figsize=(14, 7),
    cmap=cm.viridis,
    style="dark_background",
    lw=0.2,
    axes_to_reverse = [1,2]


)
plt.show()

# Example 3: Different cmap 
fig3, axes3 = pp.plot(
    df=wine_df,
    target_column='quality',
    title="Wine Quality Dataset - Colored by Wine Type",
    figsize=(16, 8),
    cmap=cm.plasma,
    style="dark_background",
    lw=0.1,
    axes_to_reverse = [1,2]

)
plt.show()

# Example 3: Different cmap and hide all axes
fig3, axes3 = pp.plot(
    df=wine_df,
    target_column='quality',
    title="Wine Quality Dataset - Colored by Wine Type",
    figsize=(16, 8),
    cmap=cm.cool.reversed(),
    style="dark_background",
    lw=0.1,
    # order='random',
    hide_axes=True,
    axes_to_reverse = [0]


)
plt.show()