Interactive viz guide

Skill brycewang-stanford/Auto-Empirical-Research-Skills/skills/43-wentorai-research-plugins/skills/analysis/dataviz/interactive-viz-guide

Interactive data visualization with Plotly, ECharts, and D3From its SKILL.md

Install

npx -y skills add brycewang-stanford/Auto-Empirical-Research-Skills --skill interactive-viz-guide

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.

SKILL.md

8.2 KB, ~2.2k tokens by cl100k_base, as published. Nobody here has run it

Interactive Visualization Guide

Create interactive, publication-ready visualizations using Plotly, ECharts, Altair, and Bokeh for academic papers, presentations, and supplementary materials.

When to Use Interactive Visualizations

Scenario	Static	Interactive
Journal PDF figure	Preferred	Not supported
Supplementary materials	Optional	Excellent
Conference poster (digital)	Common	Increasingly popular
Presentation slides	Standard	Engaging
Online appendix / project website	Limited	Ideal
Exploratory data analysis	Quick	Detailed exploration

Plotly (Python)

Plotly produces interactive HTML charts with hover tooltips, zoom, pan, and export capabilities.

Scatter Plot with Hover Details

import plotly.express as px
import pandas as pd

# Example: visualize paper citations vs. year
df = pd.DataFrame({
    "title": ["Paper A", "Paper B", "Paper C", "Paper D", "Paper E"],
    "year": [2019, 2020, 2021, 2022, 2023],
    "citations": [150, 320, 89, 450, 210],
    "field": ["NLP", "CV", "NLP", "RL", "CV"],
    "venue": ["ACL", "CVPR", "EMNLP", "NeurIPS", "ICCV"]
})

fig = px.scatter(
    df, x="year", y="citations",
    color="field", size="citations",
    hover_data=["title", "venue"],
    title="Citation Counts by Year and Field",
    labels={"citations": "Citation Count", "year": "Publication Year"}
)
fig.update_layout(
    template="plotly_white",
    font=dict(size=14),
    width=800, height=500
)
fig.write_html("citations_interactive.html")
fig.show()

Grouped Bar Chart

import plotly.graph_objects as go

methods = ["Baseline", "Method A", "Method B", "Ours"]
accuracy = [82.1, 85.3, 87.0, 89.4]
f1_score = [79.8, 83.1, 85.2, 87.9]

fig = go.Figure(data=[
    go.Bar(name="Accuracy", x=methods, y=accuracy,
           text=[f"{v}%" for v in accuracy], textposition="auto"),
    go.Bar(name="F1 Score", x=methods, y=f1_score,
           text=[f"{v}%" for v in f1_score], textposition="auto")
])
fig.update_layout(
    barmode="group",
    title="Model Performance Comparison",
    yaxis_title="Score (%)",
    yaxis_range=[70, 95],
    template="plotly_white"
)
fig.write_html("comparison.html")

Heatmap (Confusion Matrix)

import plotly.figure_factory as ff
import numpy as np

z = [[85, 5, 3, 7],
     [4, 90, 2, 4],
     [6, 3, 88, 3],
     [5, 2, 7, 86]]
labels = ["Class A", "Class B", "Class C", "Class D"]

fig = ff.create_annotated_heatmap(
    z, x=labels, y=labels,
    colorscale="Blues",
    showscale=True
)
fig.update_layout(
    title="Confusion Matrix",
    xaxis_title="Predicted",
    yaxis_title="Actual"
)
fig.write_html("confusion_matrix.html")

Altair (Python - Declarative)

Altair uses Vega-Lite grammar for concise, declarative visualization.

import altair as alt
import pandas as pd

# Interactive scatter with selection
df = pd.DataFrame({
    "x": range(100),
    "y": [v**2 + 10 for v in range(100)],
    "category": ["A" if i % 3 == 0 else "B" if i % 3 == 1 else "C" for i in range(100)]
})

selection = alt.selection_point(fields=["category"], bind="legend")

chart = alt.Chart(df).mark_circle(size=60).encode(
    x="x:Q",
    y="y:Q",
    color="category:N",
    opacity=alt.condition(selection, alt.value(1), alt.value(0.2)),
    tooltip=["x", "y", "category"]
).add_params(
    selection
).properties(
    width=600, height=400,
    title="Interactive Scatter with Legend Selection"
).interactive()  # Enable zoom/pan

chart.save("altair_scatter.html")

ECharts (JavaScript)

Apache ECharts is a powerful JavaScript charting library ideal for web dashboards and complex visualizations.

<!DOCTYPE html>
<html>
<head>
  <script src="https://cdn.jsdelivr.net/npm/echarts@5/dist/echarts.min.js"></script>
</head>
<body>
  <div id="chart" style="width: 800px; height: 500px;"></div>
  <script>
    const chart = echarts.init(document.getElementById('chart'));

    const option = {
      title: { text: 'Research Output by Year', left: 'center' },
      tooltip: {
        trigger: 'axis',
        axisPointer: { type: 'shadow' }
      },
      legend: { data: ['Papers', 'Citations'], top: 30 },
      xAxis: {
        type: 'category',
        data: ['2019', '2020', '2021', '2022', '2023']
      },
      yAxis: [
        { type: 'value', name: 'Papers' },
        { type: 'value', name: 'Citations' }
      ],
      series: [
        {
          name: 'Papers',
          type: 'bar',
          data: [12, 15, 18, 22, 28],
          itemStyle: { color: '#3B82F6' }
        },
        {
          name: 'Citations',
          type: 'line',
          yAxisIndex: 1,
          data: [45, 120, 280, 450, 680],
          itemStyle: { color: '#EF4444' },
          smooth: true
        }
      ],
      dataZoom: [{ type: 'slider', start: 0, end: 100 }]
    };

    chart.setOption(option);
    window.addEventListener('resize', () => chart.resize());
  </script>
</body>
</html>

Network Visualization

Plotly Network Graph

import plotly.graph_objects as go
import networkx as nx

# Create a citation network
G = nx.karate_club_graph()
pos = nx.spring_layout(G, seed=42)

# Edge traces
edge_x, edge_y = [], []
for edge in G.edges():
    x0, y0 = pos[edge[0]]
    x1, y1 = pos[edge[1]]
    edge_x.extend([x0, x1, None])
    edge_y.extend([y0, y1, None])

edge_trace = go.Scatter(x=edge_x, y=edge_y, mode="lines",
                        line=dict(width=0.5, color="#888"), hoverinfo="none")

# Node traces
node_x = [pos[n][0] for n in G.nodes()]
node_y = [pos[n][1] for n in G.nodes()]
node_degree = [G.degree(n) for n in G.nodes()]

node_trace = go.Scatter(
    x=node_x, y=node_y, mode="markers",
    marker=dict(size=[d*3 for d in node_degree], color=node_degree,
                colorscale="Viridis", showscale=True,
                colorbar=dict(title="Connections")),
    text=[f"Node {n}: {G.degree(n)} connections" for n in G.nodes()],
    hoverinfo="text"
)

fig = go.Figure(data=[edge_trace, node_trace],
                layout=go.Layout(title="Citation Network",
                                 showlegend=False,
                                 xaxis=dict(showgrid=False, zeroline=False, showticklabels=False),
                                 yaxis=dict(showgrid=False, zeroline=False, showticklabels=False)))
fig.write_html("network.html")

Exporting for Publication

Static Export from Plotly

# Export as high-res static image for journals
fig.write_image("figure.pdf", width=1200, height=800, scale=2)  # Vector PDF
fig.write_image("figure.png", width=1200, height=800, scale=3)  # 300 DPI PNG
fig.write_image("figure.svg", width=1200, height=800)            # Vector SVG

# Requires: pip install kaleido

Embedding in Jupyter Notebooks

# Plotly renders natively in Jupyter
fig.show()

# For Altair in Jupyter
chart  # Just display the chart object

# For ECharts in Jupyter, use pyecharts
from pyecharts.charts import Bar
from pyecharts import options as opts

bar = (Bar()
    .add_xaxis(["2019", "2020", "2021", "2022", "2023"])
    .add_yaxis("Papers", [12, 15, 18, 22, 28])
    .set_global_opts(title_opts=opts.TitleOpts(title="Research Output")))
bar.render_notebook()

Best Practices

Start with a static version: Ensure your visualization works as a static figure first; interactivity is an enhancement, not a replacement.
Meaningful tooltips: Show relevant context on hover (paper title, exact values, metadata), not just coordinates.
Responsive design: Use percentage-based sizing or window.addEventListener('resize') for ECharts.
Accessibility: Provide text alternatives, use colorblind-friendly palettes, and ensure keyboard navigation.
Performance: For datasets over 10,000 points, use WebGL renderers (Plotly's scattergl, Deck.gl) or server-side aggregation.
Reproducibility: Save the data alongside the visualization so others can recreate it.

What ships with it

Read from the repository

Just SKILL.md. No reference files, no scripts.