Interactive Visualization

What the source does

• FilesystemNone observedNo reads or writes in the source
• NetworkNone observedNo outbound calls in the source
• ShellNone observedNo commands run in the source
• DependenciesPinnedNothing floats
• SecretsNone foundNo keys in source

Five audit layers

Gates applied: no behavioural pass.

Audit 0.4.1 · source sha c8c1fc3a2b13 · 19 Oct 2026Report an issue or request a re-scan

• PublisherGPTomics GitHub organizationOwner of the repository this skill is read from
• Source folderGPTomics/bioSkills/data-visualization/interactive-visualizationThe exact folder we audited
• LicenceMITAs the repository states it
• Commit pinned for this verdictc8c1fc3a2b13A new audit runs when the source changes
• Repository statusArchivedLast push 65 days ago
• Sandboxed runNot runSkills are read, not executed; layer L4 is skipped

Its description: Build interactive HTML/web visualizations with plotly (Python/R), bokeh (Python), and gganimate/plotly frames for animation, with awareness of current Kaleido static-export model (post-orca-EOL), HTML file-size bloat,…

---
name: bio-data-visualization-interactive-visualization
description: Build interactive HTML/web visualizations with plotly (Python/R), bokeh (Python), and gganimate/plotly frames for animation, with awareness of current Kaleido static-export model (post-orca-EOL), HTML file-size bloat, and the limits of interactive-only output for journal submission. Use when producing zoomable/hoverable plots for notebook EDA, supplementary HTML, dashboards, or animated time-course / iteration visualizations.
tool_type: mixed
primary_tool: plotly
---

## Version Compatibility

Reference examples tested with: plotly 5.24+, plotly R 4.10+, bokeh 3.4+, kaleido 1.0+ (note: v1 dropped bundled Chrome), gganimate 1.0.9+, altair 5.4+, htmlwidgets 1.6+.

Before using code patterns, verify installed versions match. If versions differ:
- Python: `pip show <package>` then `help(module.function)`
- R: `packageVersion('<pkg>')` then `?function_name`

If code throws ImportError, AttributeError, or TypeError, introspect the installed package and adapt the example to match the actual API rather than retrying.

# Interactive Visualization

**"Build an interactive plot"** -> Render a zoomable, hoverable, panable HTML/web visualization, knowing that interactive output is a SUPPLEMENT to (not replacement for) the static figure needed for journal submission. Choose plotly for fastest onboarding and ggplot2 conversion (`ggplotly`); bokeh for streaming/server-side; altair for grammar-of-graphics; D3.js for full custom.

- Python: `plotly.graph_objects`, `plotly.express`, `bokeh`, `altair`
- R: `plotly` (via `ggplotly`), `htmlwidgets` ecosystem (leaflet, networkD3, DT)

## The Single Most Important Modern Insight -- Kaleido v1 and the Static-Export Pipeline

Interactive plots produce HTML, but journals need static PDF/PNG. The plotly static-export pipeline changed materially in 2025:

- **Orca is end-of-life** (deprecated 2021, removed pipeline 2025)
- **`fig.write_image(..., engine='orca')` removed in plotly 6.2** (post-Sept 2025)
- **Kaleido v1+ is the current standard** — pass no `engine=` argument
- **Kaleido v1 dropped bundled Chrome** — requires installed Chrome / Chromium
- **EPS export removed in Kaleido v1** (was supported via orca's bundled Chromium)

For static export of plotly figures in 2026: `pip install kaleido`; verify Chrome installed; `fig.write_image('out.pdf')`. Test by writing to a known path and inspecting file size; silent failure on missing Chrome was a 2024-2025 pain point that v1 partially addresses with clearer errors.

## Interactive vs Static — The Reproducibility Cost

Interactive HTML has hidden trade-offs:

- **File size**: a 5000-point plotly HTML is 3-5 MB (embedded JS bundle). 50000 points crashes browsers without WebGL acceleration.
- **Non-citable**: a paper figure must be static. Always export static alongside.
- **Browser version drift**: HTML from 2020 plotly may not render in 2026 browsers.
- **Cannot be alt-text described**: accessibility weaker than static.

Use interactive for notebooks (exploration), supplementary HTML (online journal supplement), dashboards (Streamlit/Dash/Shiny). For the journal figure, always also produce static.

## plotly (Python) — Standard Interactive

**Goal:** Build an interactive HTML plot with zoom, pan, and hover-tooltip behavior; export both interactive HTML for supplements and static PDF for the journal figure.

**Approach:** Use `plotly.express` for declarative high-level plots OR `graph_objects` for fine control; enable WebGL via `render_mode='webgl'` or `Scattergl` for >5000 points; export HTML with `write_html()` and static with `write_image()` after installing Kaleido v1+ and Chrome.

```python
import plotly.express as px
import plotly.graph_objects as go

# Express: high-level, declarative
fig = px.scatter(df, x='PC1', y='PC2', color='cluster',
                  hover_data=['gene_count', 'sample_id'],
                  color_discrete_sequence=['#0072B2', '#D55E00', '#009E73'],
                  title='PCA')
fig.update_layout(template='plotly_white', width=600, height=500)

# WebGL acceleration for >5000 points
fig = px.scatter(df, x='PC1', y='PC2', color='cluster', render_mode='webgl')

# Save
fig.write_html('pca.html')
fig.write_image('pca.pdf')                   # requires kaleido + Chrome

# Graph_objects: low-level
fig = go.Figure(go.Scattergl(                 # Scattergl == WebGL scatter
    x=df['PC1'], y=df['PC2'],
    mode='markers',
    marker=dict(color=df['cluster_code'], colorscale='Tab10', size=4),
    text=df['sample_id'], hoverinfo='text'))
```

## plotly (R) — ggplotly Conversion

```r
library(plotly)
library(ggplot2)

p <- ggplot(df, aes(x = PC1, y = PC2, color = cluster, text = sample_id)) +
    geom_point() + theme_classic()

# Convert ggplot to interactive plotly
p_int <- ggplotly(p, tooltip = c('text', 'x', 'y', 'colour'))

# Save
htmlwidgets::saveWidget(p_int, 'pca.html', selfcontained = TRUE)
```

ggplotly is the lowest-friction R interactive path — write ggplot, get plotly.

## bokeh (Python) — Server-Side / Streaming

```python
from bokeh.plotting import figure, output_file, save
from bokeh.models import ColumnDataSource, HoverTool

output_file('pca_bokeh.html')

source = ColumnDataSource(df)
p = figure(title='PCA', x_axis_label='PC1', y_axis_label='PC2',
           tools='pan,wheel_zoom,box_zoom,reset,hover,save')
p.scatter('PC1', 'PC2', source=source, size=8, alpha=0.7,
          color={'field': 'cluster', 'transform': cluster_cmap})
p.add_tools(HoverTool(tooltips=[('Sample', '@sample_id'), ('Cluster', '@cluster')]))
save(p)
```

bokeh is stronger than plotly for streaming dashboards and server-side aggregation. Static export via `bokeh.io.export_png` requires selenium + Chrome.

## Animation — gganimate (R) and plotly frames (Python)

```r
library(gganimate)
p <- ggplot(df, aes(x, y, color = condition)) +
    geom_point(size = 3) +
    theme_classic() +
    transition_time(time) +                  # animate over time

Excerpt, first 6,000 characters of SKILL.md, verbatim from the audited commit. It is text from someone else’s repository, quoted so you can judge it before your agent reads it.

pip install plotly bokeh kaleido altair

git clone https://github.com/GPTomics/bioSkills
git -C bioSkills checkout c8c1fc3a2b13

Works in

1. 2026-10-19
   ✓SAFE B 89 / 100

   first audit · source c8c1fc3a2b13

Questions about Interactive Visualization

• github.com/GPTomics/bioSkills/tree/main/data-visualization/inter — the audited source at commit c8c1fc3a2b13 and every finding aboveObserved · 19 Oct 2026

Spotted something wrong? Report a correction · Methodology