DS Seaborn Visualization
Seaborn is a Python library built on top of Matplotlib. It produces polished statistical charts with short commands and works directly with Pandas tables. A chart that needs fifteen lines in Matplotlib often needs one line in Seaborn.
Install and Import
pip install seaborn
import seaborn as sns
import matplotlib.pyplot as plt
sns.set_theme()
tips = sns.load_dataset("tips")
print(tips.head())The tips dataset ships with Seaborn. It records restaurant bills, tips, days, and group sizes, which makes it a friendly practice set.
Pick the Right Chart
| Question | Chart | Seaborn Function |
|---|---|---|
| How are values spread? | Histogram | histplot |
| How do groups compare? | Box plot | boxplot |
| Do two numbers move together? | Scatter plot | scatterplot |
| How many items per category? | Bar count | countplot |
| Which columns relate to each other? | Heatmap | heatmap |
Histogram: See the Spread
sns.histplot(data=tips, x="total_bill", bins=20, kde=True)
plt.show()Bars count how many bills fall in each price range. The kde option adds a smooth curve over the bars.
Box Plot: Compare Groups
sns.boxplot(data=tips, x="day", y="total_bill")
plt.show()How to read a box plot
o <-- outlier (unusually high bill)
|
+-----+ <-- top of box: 75% of bills fall below this
| |
|-----| <-- line inside: median bill
| |
+-----+ <-- bottom of box: 25% of bills fall below this
|
| <-- whisker: normal range continues
The box covers the middle half of the values. Short boxes show steady data, and tall boxes show wide variety.
Scatter Plot: Find Relationships
sns.scatterplot(data=tips, x="total_bill", y="tip", hue="time", size="size")
plt.show()Each dot is one meal. The hue option colors dots by lunch or dinner. The size option makes dots larger for bigger groups. A cloud of dots rising from left to right shows that larger bills bring larger tips.
Count Plot: Count Categories
sns.countplot(data=tips, x="day", hue="sex")
plt.show()Heatmap: Spot Strong Links
corr = tips[["total_bill", "tip", "size"]].corr()
sns.heatmap(corr, annot=True, cmap="Blues")
plt.show()A heatmap colors each cell by the strength of the link between two columns. Numbers near 1 show a strong positive link. Numbers near 0 show no link. Numbers near negative 1 show an opposite link.
Pair Plot: Overview in One Image
sns.pairplot(tips, hue="time")
plt.show()A pair plot draws a scatter plot for every pair of numeric columns, with histograms on the diagonal. Use it early in exploration to find promising relationships.
Style Your Charts
- Call sns.set_theme(style="whitegrid") for a clean grid background.
- Set figure size with plt.figure(figsize=(8, 5)).
- Add titles with plt.title and axis labels with plt.xlabel and plt.ylabel.
- Choose color palettes with the palette option, such as "viridis" or "Set2".
Accessible colors help every reader. Pick palettes that stay readable for people with color vision differences, and label categories directly where space allows.
Seaborn versus Matplotlib
| Feature | Matplotlib | Seaborn |
|---|---|---|
| Code length | Longer | Shorter |
| Default look | Plain | Polished |
| Statistical charts | Manual work | Built in |
| Fine control | Full | Uses Matplotlib underneath |
Both libraries work together. Draw with Seaborn, and fine-tune with Matplotlib commands.
Mini Project: Explore Restaurant Tips
Combine the charts above to answer a real question: does group size change the tip percentage? Build a new column for tip percentage first, then draw a box plot by group size.
tips["tip_pct"] = tips["tip"] / tips["total_bill"] * 100
sns.boxplot(data=tips, x="size", y="tip_pct")
plt.title("Tip Percentage by Group Size")
plt.show()Look at the median line in each box. Groups of one or two diners often show wider spreads, and large groups show steadier values. A chart like this turns a vague question into a visible answer within a minute.
Common Beginner Mistakes
- Forgetting plt.show() in scripts, which leaves the chart undisplayed.
- Plotting text columns on a numeric axis.
- Using too many colors in one chart.
- Leaving out titles and axis labels.
Key Points
- Seaborn creates attractive statistical charts from Pandas tables.
- Match the chart type to the question you want to answer.
- Pair plots and heatmaps give fast overviews of new datasets.