What datasets are used in the guide for creating charts?

The guide uses synthetic datasets from the CSV files 'plotlycountrymetrics.csv' and 'plotlycafeorders.csv'.

What are some features you can add to Plotly Express charts?

You can add hover labels, custom tooltips, map columns to color, size, symbol, and animation, and use log axes for values with large ranges.

How do you install Plotly and Pandas for this guide?

You can install Plotly and Pandas using the command 'pip install plotly pandas'.

Plotly Express: Create Interactive Charts in

Plotly Express: Interactive Scatter, Line, Bar, Histogram, Pie, Sunburst, Heatmaps, 3D Charts, Facets, and Subplots

Plotly helps you build interactive charts in Python.

Matplotlib is excellent when you want full control over static figures. Plotly is excellent when you want charts that users can hover, zoom, pan, filter, and inspect in a browser or notebook.

In this guide, you will learn Plotly Express through practical examples:

interactive scatter plots
color, size, hover labels, and tooltips
log-scaled axes
animated charts
line charts
single, grouped, and stacked bar charts
histograms
pie and donut charts
sunburst and treemap charts
heatmaps with px.imshow
3D scatter plots
scatter matrix plots
facet grids
mixed dashboards with make_subplots

The examples use original synthetic datasets, not copied notebook, course, sports, or public tutorial data.

Files Used In This Guide

Use these CSV files:

plotly_country_metrics.csv
plotly_cafe_orders.csv

Place them in the same folder as your notebook or script.

If you keep them in a data/ folder, change the paths like this:

python

country = pd.read_csv("data/plotly_country_metrics.csv")
orders = pd.read_csv("data/plotly_cafe_orders.csv")

What You Will Learn

By the end, you should be able to:

create a Plotly Express figure from a Pandas DataFrame
decide when to use Plotly Express and when to use Graph Objects
add hover labels and custom tooltips
map columns to color, size, symbol, and animation
use log axes for values with large ranges
build grouped and stacked bar charts
create distribution charts with histograms
use pie charts only when they are appropriate
visualize hierarchy with sunburst and treemap charts
convert pivot tables into heatmaps
create 3D charts without manual JavaScript
split a chart into facets
combine multiple charts with make_subplots

1. Setup

Install Plotly if needed:

bash

pip install plotly pandas

Import the libraries:

python

import pandas as pd
import plotly.express as px
import plotly.graph_objects as go
from plotly.subplots import make_subplots

Load the datasets:

python

country = pd.read_csv("plotly_country_metrics.csv")
orders = pd.read_csv("plotly_cafe_orders.csv")

orders["date"] = pd.to_datetime(orders["date"])

print(country.head())
print(orders.head())

The country table contains fictional country-level metrics across multiple years.

The orders table contains fictional cafe orders across cities, products, channels, and customer types.

2. Plotly Express vs Graph Objects

Plotly has two common APIs:

plotly.express for quick charts from tidy DataFrames
plotly.graph_objects for lower-level figure construction

Most beginner and intermediate work should start with Plotly Express.

Example with Graph Objects:

python

auralia = country[country["country"] == "Auralia"]

fig = go.Figure()

fig.add_trace(
    go.Scatter(
        x=auralia["year"],
        y=auralia["internet_users_pct"],
        mode="lines+markers",
        name="Auralia",
    )
)

fig.update_layout(
    title="Internet Adoption In Auralia",
    xaxis_title="Year",
    yaxis_title="Internet Users (%)",
)

fig.show()

The same chart with Plotly Express:

python

fig = px.line(
    auralia,
    x="year",
    y="internet_users_pct",
    markers=True,
    title="Internet Adoption In Auralia",
)

fig.show()

Use Plotly Express when your data is already in a DataFrame and you want a fast chart.

Use Graph Objects when you need precise control over traces, subplots, annotations, or custom dashboards.

3. First Interactive Scatter Plot

A scatter plot compares two numeric columns.

This chart compares life expectancy and GDP per person for the latest year:

python

latest = country[country["year"] == country["year"].max()]

fig = px.scatter(
    latest,
    x="life_expectancy",
    y="gdp_per_capita",
    title="Life Expectancy vs GDP Per Person",
)

fig.show()

Plotly charts are interactive by default.

You can hover over a point, zoom into a region, pan around the chart, and save the image from the chart toolbar.

4. Color, Size, And Hover Labels

Plotly Express becomes powerful when you map DataFrame columns to visual encodings.

Example:

x-axis: life expectancy
y-axis: GDP per person
color: continent
size: population
hover label: country

python

fig = px.scatter(
    latest,
    x="life_expectancy",
    y="gdp_per_capita",
    color="continent",
    size="population_millions",
    hover_name="country",
    title="Country Metrics With Color, Size, And Hover",
    labels={
        "life_expectancy": "Life Expectancy",
        "gdp_per_capita": "GDP Per Person",
        "population_millions": "Population (Millions)",
    },
)

fig.show()

Use hover_name for the main label users should see first.

Use labels to make axis names readable.

5. Custom Hover Data

By default, Plotly shows columns used in the chart.

Use hover_data when you want to add or hide details.

python

fig = px.scatter(
    latest,
    x="internet_users_pct",
    y="education_index",
    color="continent",
    size="population_millions",
    hover_name="country",
    hover_data={
        "population_millions": ":.1f",
        "gdp_per_capita": ":,.0f",
        "carbon_tons_per_person": ":.2f",
    },
    title="Education And Internet Access",
)

fig.show()

The format strings make tooltips easier to scan.

6. Log Axes

Use a log axis when values have a wide range.

GDP per person can vary a lot, so a log-scaled y-axis can make the pattern easier to read.

python

fig = px.scatter(
    latest,
    x="life_expectancy",
    y="gdp_per_capita",
    color="continent",
    size="population_millions",
    hover_name="country",
    log_y=True,
    title="Life Expectancy vs GDP Per Person On A Log Scale",
)

fig.show()

Do not use log scales silently. Make sure the chart title, axis, or surrounding explanation makes the scale clear.

7. Animated Scatter Plot

Animation helps when the same entities repeat across time.

python

fig = px.scatter(
    country,
    x="life_expectancy",
    y="gdp_per_capita",
    color="continent",
    size="population_millions",
    hover_name="country",
    animation_frame="year",
    animation_group="country",
    range_x=[65, 84],
    range_y=[12000, 62000],
    title="Country Development Metrics Over Time",
)

fig.show()

animation_group tells Plotly that the same country appears in multiple frames.

Fixed ranges prevent the chart from jumping as the animation moves between years.

8. Line Charts

Line charts are useful for trends.

python

fig = px.line(
    country,
    x="year",
    y="internet_users_pct",
    color="country",
    markers=True,
    title="Internet Adoption By Country",
)

fig.show()

If too many lines make the chart noisy, filter the data first:

python

focus_countries = ["Auralia", "Borevia", "Eldora", "Harboris"]
focus = country[country["country"].isin(focus_countries)]

fig = px.line(
    focus,
    x="year",
    y="life_expectancy",
    color="country",
    markers=True,
    title="Life Expectancy Trend For Selected Countries",
)

fig.show()

9. Bar Charts

A bar chart compares categories.

Create a continent-level summary:

python

continent_summary = (
    latest
    .groupby("continent", as_index=False)
    .agg(
        population_millions=("population_millions", "sum"),
        avg_gdp_per_capita=("gdp_per_capita", "mean"),
        avg_life_expectancy=("life_expectancy", "mean"),
    )
)

fig = px.bar(
    continent_summary,
    x="continent",
    y="population_millions",
    color="continent",
    text_auto=".1f",
    title="Population By Continent Group",
)

fig.show()

text_auto writes values on the bars.

Use it when the chart has enough space.

10. Grouped Bar Chart

Grouped bars compare multiple categories side by side.

python

city_category = (
    orders
    .groupby(["city", "category"], as_index=False)
    .agg(revenue=("total_bill", "sum"))
)

fig = px.bar(
    city_category,
    x="city",
    y="revenue",
    color="category",
    barmode="group",
    text_auto=".0f",
    title="Cafe Revenue By City And Category",
)

fig.show()

Grouped bars work well when each x-axis category has a small number of groups.

11. Stacked Bar Chart

Stacked bars show composition.

python

city_channel = (
    orders
    .groupby(["city", "channel"], as_index=False)
    .agg(revenue=("total_bill", "sum"))
)

fig = px.bar(
    city_channel,
    x="city",
    y="revenue",
    color="channel",
    title="Revenue Composition By Sales Channel",
)

fig.show()

The default px.bar behavior stacks bars when multiple rows share the same x-axis category.

Use stacked bars for part-to-whole comparison.

Use grouped bars for exact side-by-side comparison.

12. Log-Scaled Bar Chart

If one category is much larger than the others, a log axis can help.

python

product_revenue = (
    orders
    .groupby("product", as_index=False)
    .agg(revenue=("total_bill", "sum"))
    .sort_values("revenue", ascending=False)
)

fig = px.bar(
    product_revenue,
    x="product",
    y="revenue",
    log_y=True,
    text_auto=".0f",
    title="Product Revenue On A Log Scale",
)

fig.update_layout(xaxis_tickangle=-35)
fig.show()

Rotate long labels when they crowd the x-axis.

13. Histograms

Histograms show distributions.

python

fig = px.histogram(
    orders,
    x="total_bill",
    nbins=14,
    color="channel",
    marginal="box",
    title="Distribution Of Order Values",
)

fig.show()

marginal="box" adds a compact box plot above the histogram.

Other useful marginal options include:

python

"rug"
"violin"
"box"

14. Histogram With Facets

Facets split one chart into small multiples.

python

fig = px.histogram(
    orders,
    x="service_rating",
    color="customer_type",
    facet_col="channel",
    nbins=10,
    title="Service Ratings By Channel",
)

fig.show()

Facets are often better than putting too many colors into one chart.

15. Pie And Donut Charts

Pie charts are useful only for simple part-to-whole comparisons.

Use them when:

there are few categories
the values add up to a meaningful whole
exact comparison is not the main goal

python

category_share = (
    orders
    .groupby("category", as_index=False)
    .agg(revenue=("total_bill", "sum"))
)

fig = px.pie(
    category_share,
    values="revenue",
    names="category",
    title="Revenue Share By Category",
)

fig.show()

A donut chart is a pie chart with a hole:

python

fig = px.pie(
    category_share,
    values="revenue",
    names="category",
    hole=0.45,
    title="Revenue Share By Category",
)

fig.show()

If you need accurate category comparison, prefer a bar chart.

16. Sunburst Charts

Sunburst charts show hierarchy.

This example shows revenue from city to category to product:

python

order_hierarchy = (
    orders
    .groupby(["city", "category", "product"], as_index=False)
    .agg(
        revenue=("total_bill", "sum"),
        avg_rating=("service_rating", "mean"),
    )
)

fig = px.sunburst(
    order_hierarchy,
    path=["city", "category", "product"],
    values="revenue",
    color="avg_rating",
    color_continuous_scale="Blues",
    title="Cafe Revenue Hierarchy",
)

fig.show()

Sunburst charts are useful when hierarchy matters more than exact bar-to-bar comparison.

17. Treemap Charts

A treemap uses rectangles instead of radial slices.

python

fig = px.treemap(
    order_hierarchy,
    path=["city", "category", "product"],
    values="revenue",
    color="avg_rating",
    color_continuous_scale="Greens",
    title="Cafe Revenue Treemap",
)

fig.show()

Treemaps often use screen space more efficiently than sunburst charts.

18. Heatmaps With `px.imshow`

Heatmaps are useful for matrix-shaped summaries.

Create a pivot table:

python

rating_matrix = orders.pivot_table(
    index="city",
    columns="category",
    values="service_rating",
    aggfunc="mean",
)

fig = px.imshow(
    rating_matrix,
    text_auto=".2f",
    color_continuous_scale="Blues",
    title="Average Service Rating By City And Category",
)

fig.show()

The input to px.imshow can be a NumPy array, a list of lists, or a Pandas DataFrame.

A DataFrame is convenient because it keeps row and column labels.

19. Another Heatmap: Fulfillment Time

python

time_matrix = orders.pivot_table(
    index="city",
    columns="channel",
    values="delivery_minutes",
    aggfunc="mean",
)

fig = px.imshow(
    time_matrix,
    text_auto=".1f",
    color_continuous_scale="Oranges",
    title="Average Fulfillment Minutes By City And Channel",
)

fig.show()

Use heatmaps when the same metric is measured across two category axes.

20. 3D Scatter Plot

3D charts can be helpful for exploration, but they are not always the best presentation choice.

Use them when rotation helps users inspect the data.

python

fig = px.scatter_3d(
    latest,
    x="internet_users_pct",
    y="education_index",
    z="gdp_per_capita",
    color="continent",
    size="population_millions",
    hover_name="country",
    title="3D Country Metrics",
)

fig.show()

Do not use 3D just because it looks impressive.

For reports, a 2D scatter plot is often easier to read.

21. Scatter Matrix

A scatter matrix compares multiple numeric columns pairwise.

python

fig = px.scatter_matrix(
    latest,
    dimensions=[
        "life_expectancy",
        "gdp_per_capita",
        "internet_users_pct",
        "education_index",
    ],
    color="continent",
    hover_name="country",
    title="Scatter Matrix For Country Metrics",
)

fig.update_traces(diagonal_visible=False)
fig.show()

Use a scatter matrix for quick relationship scanning.

It can become crowded if you include too many columns.

22. Facet Columns

Facets create one small chart per category.

python

fig = px.scatter(
    country,
    x="life_expectancy",
    y="gdp_per_capita",
    color="continent",
    hover_name="country",
    facet_col="year",
    facet_col_wrap=3,
    title="Life Expectancy vs GDP Across Years",
)

fig.show()

facet_col_wrap=3 places three facet panels per row.

23. Facet Rows And Columns

You can facet by two category variables at once.

python

fig = px.scatter(
    orders,
    x="total_bill",
    y="tip",
    color="customer_type",
    facet_col="channel",
    facet_row="city",
    hover_name="product",
    title="Tips vs Order Value By City And Channel",
)

fig.show()

This can produce many panels.

Use it only when the grid remains readable.

24. Updating Layout

Every Plotly Express chart returns a Figure object.

You can update that object before showing it.

python

fig = px.scatter(
    latest,
    x="internet_users_pct",
    y="life_expectancy",
    color="continent",
    hover_name="country",
    title="Internet Access And Life Expectancy",
)

fig.update_layout(
    template="plotly_white",
    legend_title_text="Continent Group",
    title_x=0.02,
)

fig.update_xaxes(title="Internet Users (%)")
fig.update_yaxes(title="Life Expectancy")

fig.show()

Common layout changes:

python

fig.update_layout(template="plotly_white")
fig.update_layout(height=550, width=900)
fig.update_layout(legend_title_text="Group")
fig.update_xaxes(tickangle=-30)

25. Saving Interactive Charts

Save an interactive chart as HTML:

python

fig.write_html("country_scatter.html")

This creates a file that can be opened in a browser.

Save a static image only if you installed the static export dependency:

bash

pip install kaleido

Then:

python

fig.write_image("country_scatter.png")

Use HTML when interactivity matters.

Use PNG or SVG for static reports, slide decks, and documents.

26. Subplots With `make_subplots`

Plotly Express is convenient, but mixed dashboards often use Graph Objects.

Create summaries:

python

monthly_revenue = (
    orders
    .assign(month=orders["date"].dt.to_period("M").astype(str))
    .groupby("month", as_index=False)
    .agg(revenue=("total_bill", "sum"))
)

rating_by_city = (
    orders
    .groupby("city", as_index=False)
    .agg(avg_rating=("service_rating", "mean"))
)

Build a 2x2 dashboard:

python

fig = make_subplots(
    rows=2,
    cols=2,
    subplot_titles=[
        "Monthly Revenue",
        "Average Rating By City",
        "Order Value Distribution",
        "Internet Adoption Trend",
    ],
)

fig.add_trace(
    go.Scatter(
        x=monthly_revenue["month"],
        y=monthly_revenue["revenue"],
        mode="lines+markers",
        name="Revenue",
    ),
    row=1,
    col=1,
)

fig.add_trace(
    go.Bar(
        x=rating_by_city["city"],
        y=rating_by_city["avg_rating"],
        name="Rating",
    ),
    row=1,
    col=2,
)

fig.add_trace(
    go.Histogram(
        x=orders["total_bill"],
        name="Order Values",
    ),
    row=2,
    col=1,
)

for country_name in ["Auralia", "Eldora", "Harboris"]:
    temp = country[country["country"] == country_name]
    fig.add_trace(
        go.Scatter(
            x=temp["year"],
            y=temp["internet_users_pct"],
            mode="lines+markers",
            name=country_name,
        ),
        row=2,
        col=2,
    )

fig.update_layout(
    title="Cafe And Country Metrics Dashboard",
    height=750,
    showlegend=True,
)

fig.show()

This pattern is useful when one dashboard needs different chart types.

27. Practice Problems

Use the two CSV files to solve these problems.

Problem 1: Best GDP Scatter Plot

Create a scatter plot with:

x-axis: life_expectancy
y-axis: gdp_per_capita
color: continent
size: population_millions
hover label: country
only the latest year

Add a clear title and readable axis labels.

Problem 2: Animated Internet Adoption

Create an animated scatter plot with:

x-axis: internet_users_pct
y-axis: education_index
animation by year
animation group by country
color by continent

Use fixed x and y ranges.

Problem 3: Cafe Revenue Bars

Create a grouped bar chart showing revenue by:

x-axis: city
y-axis: total total_bill
color: category

Use text_auto.

Problem 4: Channel Composition

Create a stacked bar chart showing revenue by:

x-axis: city
color: channel

Explain which city has the strongest delivery contribution.

Problem 5: Order Value Histogram

Create a histogram of total_bill.

Add:

color="customer_type"
nbins=12
marginal="box"

Problem 6: Category Donut Chart

Create a donut chart showing revenue share by product category.

Then create a bar chart with the same data and compare which chart is easier to read.

Problem 7: Sunburst Hierarchy

Create a sunburst chart with:

city
category
product

Use revenue as the value and average rating as the color.

Problem 8: Heatmap From A Pivot Table

Create a pivot table where:

rows are city
columns are channel
values are average delivery_minutes

Display it with px.imshow.

Problem 9: 3D Country Metrics

Create a 3D scatter plot with:

x-axis: internet_users_pct
y-axis: education_index
z-axis: gdp_per_capita
color: continent
size: population_millions

Write one sentence explaining whether the 3D version adds insight.

Problem 10: Faceted Tips Chart

Create a faceted scatter plot using orders:

x-axis: total_bill
y-axis: tip
color: customer_type
facet column: channel

Then add facet_row="city" and decide whether the result is too dense.

28. Plotly Express Checklist

Use this checklist when building Plotly charts:

Is the data in a tidy DataFrame?
Does each axis have a clear label?
Does color encode a real variable?
Is the hover information helpful but not noisy?
Would a log scale make the pattern clearer?
Are there too many categories for one chart?
Would facets make the comparison easier?
Is a pie chart really better than a bar chart?
Does a 3D chart add insight, or only visual complexity?
Should the output be saved as HTML for interactivity?

29. Key Takeaways

Plotly Express lets you build interactive charts quickly from Pandas DataFrames.

Start with px.scatter, px.line, px.bar, and px.histogram.

Then add color, size, hover labels, facets, animation, and layout updates.

Use hierarchy charts, heatmaps, 3D charts, and subplots when the data actually needs them.

The goal is not to use every Plotly feature.

The goal is to make data easier to explore and explain.

Plotly Express: Create Interactive Charts in Python

AI Insights

Frequently Asked Questions

Related Work

FastAPI Browser-Workflow Automation

AI Visibility & SEO Auditing Platform

RAG-Powered Learning Platform with AI Assistant