Q&A 26 How do you visualize patterns across multiple numeric features using a parallel coordinates plot?

26.1 Explanation

A parallel coordinates plot lets you explore high-dimensional patterns by plotting each feature on a vertical axis. Each observation is a line crossing all axes.

It helps you:

• Spot group patterns across all variables
  
• Detect outliers and overlaps
  
• Understand feature trends in classification problems

Coloring by category (e.g., cut) helps distinguish clusters or groups.

---

26.2 Python Code

import pandas as pd
import matplotlib.pyplot as plt
from pandas.plotting import parallel_coordinates

# Load dataset
diamonds = pd.read_csv("data/diamonds_sample.csv")

# Sample for performance
subset = diamonds[["carat", "depth", "table", "price", "x", "y", "z", "cut"]].sample(300, random_state=42)

# Normalize numeric features
normalized = subset.copy()
for col in ["carat", "depth", "table", "price", "x", "y", "z"]:
    normalized[col] = (subset[col] - subset[col].min()) / (subset[col].max() - subset[col].min())

# Parallel coordinates plot
plt.figure(figsize=(12, 6))
parallel_coordinates(normalized, class_column="cut", color=["#66c2a5", "#fc8d62", "#8da0cb", "#e78ac3", "#a6d854"])
plt.title("Parallel Coordinates Plot: Diamond Features by Cut")
plt.ylabel("Normalized Value")
plt.xticks(rotation=45)
plt.tight_layout()
plt.show()

26.3 R Code

library(readr)
library(GGally)
library(dplyr)

# Load and sample dataset
diamonds <- read_csv("data/diamonds_sample.csv")
subset <- diamonds %>% select(carat, depth, table, price, x, y, z, cut) %>% sample_n(500)

# Normalize numeric columns
subset_norm <- subset %>%
  mutate(across(c(carat, depth, table, price, x, y, z), ~ (. - min(.)) / (max(.) - min(.))))

# Parallel coordinates plot
ggparcoord(data = subset_norm,
           columns = 1:7,
           groupColumn = 8,
           scale = "uniminmax",
           alphaLines = 0.5) +
  scale_color_brewer(palette = "Set2") +
  theme_minimal() +
  labs(title = "Parallel Coordinates Plot: Diamond Features by Cut",
       x = "Features", y = "Normalized Value")

---

✅ Parallel coordinates plots help you detect patterns across multiple features simultaneously, especially when colored by category.