3.4.4 Interactive Visualization

Section overview

Many beginners, when they first see Plotly, tend to think:

• The charts look cooler
• You can move the mouse around

But the really important question is:

When do you actually need interactivity, and when is a static chart enough?

So the most important thing in this section is not “making flashy charts,” but learning to decide:

• Which scenarios are worth using interactive charts for
• Which scenarios are clearer with static charts

Optional content

This section is optional. If you are short on time, you can skip it for now and come back later when you need it. But we recommend at least skimming it once to understand what interactive charts can do.

Learning Objectives

• Understand the difference between static charts and interactive charts
• Learn to use Plotly Express for quick charting
• Understand the concept of interactive dashboards

---

First, build a mental map

Plotly is best understood through the question: “When do I want users to explore the chart themselves?”

So what this section really tries to answer is:

• What extra capability do interactive charts provide beyond static charts?
• When are they a plus, and when do they feel unnecessary?

---

Why do we need interactive charts?

Comparison	Static charts (Matplotlib/Seaborn)	Interactive charts (Plotly)
Mouse hover	Not supported	Shows detailed data
Zoom and drag	Not supported	Freely zoom and pan
Data filtering	Not supported	Click legends to hide/show
Export	Save as image	Image + HTML page
Use cases	Papers, reports	Data exploration, web display

flowchart LR
    A["Does my chart need<br/>interactivity?"] --> B{"Use case"}
    B -->|"Paper / report / PPT"| C["Matplotlib / Seaborn<br/>Static charts are enough"]
    B -->|"Data exploration / demo"| D["Plotly<br/>Interactive is more convenient"]
    B -->|"Website / dashboard"| E["Plotly + Dash<br/>Interactive app"]

    style C fill:#e3f2fd,stroke:#1565c0,color:#333
    style D fill:#e8f5e9,stroke:#2e7d32,color:#333
    style E fill:#fff3e0,stroke:#e65100,color:#333

A better analogy for beginners

You can think of Plotly as:

• A chart you can hold in your hands and inspect, zoom into, and filter yourself

A static chart is more like:

• A printed poster

An interactive chart is more like:

• A digital map with zoom and filter buttons

So it is best for:

• Readers who want to keep exploring

Not for forcing interactivity into every single scenario.

---

Installation and imports

# Install
# python -m pip install --upgrade plotly

# Plotly Express: quick charting (recommended)
import plotly.express as px

# Plotly Graph Objects: lower-level control
import plotly.graph_objects as go

import pandas as pd
import numpy as np

---

Quick start with Plotly Express

Plotly Express is Plotly’s high-level interface. You can create beautiful interactive charts with just one line of code.

The safest learning order for Plotly beginners

A good order is usually:

1. Start with px.scatter(), px.line(), and px.bar() to get familiar with basic interactions
2. Then look at heatmaps, 3D charts, and animated charts
3. Finally, study dashboards and web display

This is much less confusing than jumping straight into Dash at the beginning.

Scatter plot

# Use a built-in dataset
df = px.data.iris()

fig = px.scatter(df, x="sepal_width", y="sepal_length",
                 color="species",        # Color by species
                 size="petal_length",     # Control point size by petal length
                 hover_data=["petal_width"],  # Show petal width on hover
                 title="Iris Dataset - Interactive Scatter Plot")
fig.show()

After running this, you will see a chart that supports hovering, zooming, and dragging!

Line chart

df = px.data.gapminder()
# Filter data for China, the United States, and Japan
countries = df[df["country"].isin(["China", "United States", "Japan"])]

fig = px.line(countries, x="year", y="gdpPercap",
              color="country",
              title="Changes in GDP per Capita for China, the US, and Japan",
              labels={"gdpPercap": "GDP per Capita (USD)", "year": "Year", "country": "Country"})
fig.show()

Bar chart

df = px.data.tips()

fig = px.bar(df, x="day", y="total_bill", color="sex",
             barmode="group",          # "group" for grouped, "stack" for stacked
             title="Daily Spending (Grouped by Sex)",
             labels={"total_bill": "Bill Amount", "day": "Day", "sex": "Sex"})
fig.show()

Histogram

df = px.data.tips()

fig = px.histogram(df, x="total_bill", color="time",
                   nbins=20,
                   marginal="box",    # Marginal plot: "box", "violin", "rug"
                   title="Distribution of Bill Amounts (with Marginal Box Plot)")
fig.show()

Box plot

df = px.data.tips()

fig = px.box(df, x="day", y="total_bill", color="smoker",
             notched=True,             # Notched box plot
             title="Distribution of Daily Spending (by Smoking Status)")
fig.show()

Pie / donut chart

# Pie chart
fig = px.pie(df, names="day", values="total_bill",
             title="Proportion of Spending by Day",
             hole=0.3)  # hole > 0 turns it into a donut chart
fig.show()

---

Advanced charts

Heatmap

# Compute correlation coefficients
df = px.data.iris()
numeric_cols = df.select_dtypes(include="number")
corr = numeric_cols.corr()

fig = px.imshow(corr, text_auto=".2f",
                color_continuous_scale="RdBu_r",
                title="Correlation Heatmap of Iris Features")
fig.show()

3D scatter plot

df = px.data.iris()

fig = px.scatter_3d(df, x="sepal_length", y="sepal_width", z="petal_length",
                    color="species",
                    title="Iris 3D Scatter Plot (rotatable!)")
fig.show()

You can drag the mouse to rotate this 3D chart and view the data from different angles.

Animated chart

df = px.data.gapminder()

fig = px.scatter(df, x="gdpPercap", y="lifeExp",
                 size="pop", color="continent",
                 hover_name="country",
                 animation_frame="year",      # Play animation by year
                 animation_group="country",
                 log_x=True,
                 range_x=[100, 100000],
                 range_y=[25, 90],
                 title="Development Trajectories by Country (1952-2007)")
fig.show()

Click the play button to see how each country changes over time. This is the famous Hans Rosling bubble chart.

---

Chart customization

Update layout

fig = px.scatter(px.data.iris(), x="sepal_width", y="sepal_length",
                 color="species")

fig.update_layout(
    title=dict(text="Custom Title", font=dict(size=20)),
    xaxis_title="Sepal Width (cm)",
    yaxis_title="Sepal Length (cm)",
    template="plotly_white",      # Theme template
    width=800,
    height=500,
    legend_title="Species"
)

fig.show()

Common templates

Template name	Style
"plotly"	Default blue background
"plotly_white"	White background (recommended)
"plotly_dark"	Dark background
"ggplot2"	R ggplot style
"simple_white"	Minimal white background

A simple decision table worth remembering first

Your goal	Safer first choice
Make charts for a paper or report	Prefer static charts
Explore data details yourself	Plotly is a good fit
Build a web demo for others	Plotly is a good fit
Just want it to “look cooler”	Pause and think whether it is really necessary

This table is especially useful for beginners because it brings “interactive charts are cool” back to a question of use case.

Saving charts

# Save as interactive HTML
fig.write_html("my_chart.html")

# Save as a static image (requires kaleido)
# python -m pip install --upgrade kaleido
fig.write_image("my_chart.png", scale=2)
fig.write_image("my_chart.svg")

---

Plotly vs Matplotlib vs Seaborn

Feature	Matplotlib	Seaborn	Plotly
Interactivity	No	No	Strong
Visual appeal	Moderate	Good	Good
Learning curve	Medium	Low	Low
Customization	Very strong	Medium	Strong
Output formats	Image/PDF	Image/PDF	HTML/Image
Best for	Papers / fine control	Fast statistical charts	Data exploration / presentation

---

Introduction to interactive dashboards

In real-world work, you may need to combine multiple interactive charts into a single dashboard.

Common tools:

Tool	Features
Plotly Dash	Built with Python code, very powerful
Streamlit	Easiest to use; you can build an app with just a few lines of code
Gradio	Designed for AI model demos
Panel	Good integration with the Jupyter ecosystem

A simple Dash example (just for awareness):

# python -m pip install --upgrade dash
from dash import Dash, html, dcc
import plotly.express as px

app = Dash(__name__)

df = px.data.gapminder().query("year == 2007")
fig = px.scatter(df, x="gdpPercap", y="lifeExp", size="pop",
                 color="continent", hover_name="country", log_x=True)

app.layout = html.Div([
    html.H1("World Development Indicators Dashboard"),
    dcc.Graph(figure=fig)
])

# Run: python app.py, then open http://127.0.0.1:8050
# app.run(debug=True)

Follow-up courses

Detailed dashboard development is outside the scope of this stage. There will be hands-on exercises in the project stage later.

---

Evidence to Keep

Keep this page’s proof of learning as a small evidence card:

Question

what comparison, distribution, trend, or relationship the chart answers

Chart Choice

line, bar, scatter, histogram, box, heatmap, or interactive dashboard

Artifact

saved chart image/html plus the data slice used

Failure Check

misleading scale, overloaded chart, wrong aggregation, or missing labels

Expected Output

chart artifact with one sentence explaining the insight

Summary

Plotly Express function	Chart type
px.scatter()	Scatter plot
px.line()	Line chart
px.bar()	Bar chart
px.histogram()	Histogram
px.box()	Box plot
px.pie()	Pie chart
px.imshow()	Heatmap
px.scatter_3d()	3D scatter plot

The core advantage of Plotly is interactivity—hover to inspect data, zoom to look at details, and use animation to show changes over time. It is very useful for data exploration and presenting results.

What should you take away from this section?

• The core value of Plotly is not being flashy, but letting readers continue exploring the data
• Interactive charts are best for data exploration, web display, and presentations
• When learning for the first time, mastering px.scatter / px.line / px.bar is enough

---

Hands-on exercises

Exercise 1: Interactive exploration

# Load the gapminder dataset
# 1. Use px.scatter to plot the relationship between gdpPercap and lifeExp for each country in 2007
#    - Use pop for point size, continent for color
#    - Show country on hover
# 2. Use px.line to plot China's lifeExp changes from 1952 to 2007

Exercise 2: Data distribution

# Load the tips dataset
# 1. Use px.histogram to plot the distribution of bill amounts, and add a marginal plot with marginal="violin"
# 2. Use px.box to compare bill amounts by day

Exercise 3: Animated chart

# Use the gapminder dataset to create an animated scatter plot
# Show changes in gdpPercap vs lifeExp across countries from 1952 to 2007
# Use animation_frame="year"

Reference implementation and walkthrough

• For the Gapminder-style scatter plot, map x, y, size, color, and hover text deliberately. The hover fields should help answer the question, not just show every column.
• For a country or product trend line, sort by time and check that the x-axis is a date or ordered year field. Unsorted time data can draw misleading zigzags.
• Animated charts need a stable question. If the animation distracts from comparison, add a static summary or a final-frame note.