Key Takeaways: Interactive Visualization — plotly, Dashboard Thinking

This is your reference card for Chapter 17 — the chapter where your charts gained hover tooltips, zoom, animation, and geographic intelligence. Keep this nearby when you need to share explorable data with stakeholders.

The Core Idea

Static charts tell a story. Interactive charts let the audience explore. Use static (matplotlib, seaborn) when you control the narrative. Use interactive (plotly) when the audience needs to investigate, filter, and discover.

plotly.express Quick Reference

Common Chart Types

Encoding Variables (parallels to seaborn)

Tooltip Control

Choropleth Maps

Minimum Requirements

px.choropleth(df,
              locations="iso_alpha",  # ISO codes
              color="value_column",   # What to show
              hover_name="country")   # Tooltip title

Important Options

Animation

Key Parameters

px.scatter(df, x="x", y="y",
           animation_frame="year",       # Time column
           animation_group="country",    # Entity ID
           range_x=[0, 80000],           # Fix axes!
           range_y=[30, 100])            # Fix axes!

Rules

1. Always fix axis ranges — prevents disorienting rescaling.
2. Always set animation_group — enables smooth entity tracking.
3. Keep frame count reasonable — 10-30 frames, not 100.
4. Fix range_color for choropleth animations — ensures consistent color meaning.

Dash Dashboard Architecture

Structure

Layout (what the user sees)
  - html.H1("Title")
  - dcc.Dropdown(id="my-dropdown", options=[...])
  - dcc.Slider(id="my-slider", min=..., max=...)
  - dcc.Graph(id="my-chart")

Callback (what happens when user interacts)
  @app.callback(
      Output("my-chart", "figure"),    # What to update
      Input("my-dropdown", "value")    # What triggers update
  )
  def update(selected_value):
      filtered = df[df["col"] == selected_value]
      return px.scatter(filtered, ...)

Common Components

Dashboard Design Principles

1. Start with a question the user needs to answer
2. Limit to 3-5 interactive controls
3. Show the big picture first, details below
4. All charts should respond to the same controls
5. Provide sensible defaults — the dashboard should be informative before any interaction

Export Options

Note: Static image export requires the kaleido package.

Templates (Themes)

Set globally: import plotly.io as pio; pio.templates.default = "plotly_white"

When to Use Which Tool

Terms to Remember

What You Should Be Able to Do Now

• [ ] Create interactive charts with plotly.express: scatter, line, bar, histogram, box, violin
• [ ] Add informative tooltips with hover_name and hover_data
• [ ] Encode additional variables with color, size, symbol, facet_col, facet_row
• [ ] Build choropleth maps with country-level or state-level data
• [ ] Create animations with animation_frame and fixed axis ranges
• [ ] Customize layouts with update_layout(), templates, and titles
• [ ] Export to HTML with both self-contained and CDN modes
• [ ] Export static images for reports (PNG, PDF, SVG)
• [ ] Build a basic Dash app with a dropdown, graph, and callback
• [ ] Choose the right tool (matplotlib, seaborn, or plotly) based on audience and output format

If you checked every box, you are ready for Chapter 18, where you step back from code and think about design — what makes a visualization honest, accessible, and effective.