Data Visualization for Public Policy

The most rigorous analysis doesn’t help if decision-makers can’t read what you’re showing them. Here’s how I approach visualization for government and policy work.

Audience first

Who’s looking at this chart changes everything about how you build it. Elected officials need a clear headline and a single takeaway — minimal jargon, direct connection to constituent impact. They have thirty seconds for your chart, not thirty minutes.

Policy analysts want to go deeper: underlying data, methodology notes, statistical significance, benchmarks to compare against. Building for this audience means giving them ways to explore, not just a summary.

The general public needs context and plain language. Abstract numbers need real-world anchors. If they’re reading on a phone, it needs to work on a phone.

2. Choose the Right Chart Type

Time Series: Line Charts

Perfect for tracking changes over time:

Budget trends
Population growth
Service utilization
Performance metrics

// D3.js example
const line = d3.line()
  .x(d => xScale(d.date))
  .y(d => yScale(d.value))
  .curve(d3.curveMonotoneX); // Smooth curves

Comparisons: Bar Charts

Use for comparing values across categories:

Department budgets
Regional statistics
Survey responses
Program outcomes

Distributions: Histograms & Box Plots

Show data spread and outliers:

Income distributions
Test score ranges
Response times
Geographic variation

Relationships: Scatter Plots

Reveal correlations:

Poverty vs. education
Investment vs. outcomes
Demographics vs. services

Geographic: Maps

Essential for location-based data:

Service coverage
Demographic patterns
Resource allocation
Environmental factors

3. Use color deliberately

Accessible color palettes

Ensure colorblind-friendly choices:

// ColorBrewer palettes for maps
const colorScale = d3.scaleQuantize()
  .domain([0, maxValue])
  .range(['#f7fbff', '#08519c']); // Blue sequential

Highlight What Matters

// Emphasize key data point
const colors = data.map(d =>
  d.isHighlighted ? '#f97316' : '#9ca3af'
);

Respect Cultural Associations

Red: Danger, deficit, negative
Green: Safety, surplus, positive
Blue: Neutral, trustworthy, government

Tools and Technologies

For Web Applications

D3.js: Maximum flexibility and control

import * as d3 from 'd3';

const svg = d3.select('#chart')
  .append('svg')
  .attr('width', width)
  .attr('height', height);

Recharts: React integration made easy

import { LineChart, Line, XAxis, YAxis } from 'recharts';

<LineChart data={data}>
  <XAxis dataKey="date" />
  <YAxis />
  <Line type="monotone" dataKey="value" stroke="#f97316" />
</LineChart>

Mapbox/Leaflet: Interactive maps

import mapboxgl from 'mapbox-gl';

const map = new mapboxgl.Map({
  container: 'map',
  style: 'mapbox://styles/mapbox/light-v10',
  center: [-74.5, 40],
  zoom: 9
});

For Static Reports

R with ggplot2: Publication-quality graphics

library(ggplot2)

ggplot(data, aes(x = year, y = value)) +
  geom_line(color = "#f97316", size = 1.5) +
  labs(title = "Budget Trend 2015-2024") +
  theme_minimal()

Python with Matplotlib/Seaborn: Data science integration

import seaborn as sns

sns.set_style("whitegrid")
sns.scatterplot(data=df, x="income", y="education")

Case Studies

Municipal Lead Portal Dashboard

Challenge: 564 municipalities reporting on lead hazards across multiple data categories.

Solution:

Interactive map showing compliance status by municipality
Time series showing reporting trends
Filters for different hazard types
Export capabilities for detailed analysis

Result: 70% reduction in time to identify compliance issues.

OPRAmachine Analytics

Challenge: Communicating 75,000+ public records requests’ impact.

Visualization Approach:

Geographic heat map of request concentration
Category breakdown of most-requested records
Response time distributions by agency
Success rate trends over time

Impact: Informed policy discussions about OPRA reform.

NJ Eviction Data

Challenge: Making eviction statistics accessible to policymakers and advocates.

Design Decisions:

County-level chloropleth maps
Time series showing eviction trends
Demographic overlays
Comparison to state/national averages

Outcome: Influenced housing policy legislation.

Common Mistakes to Avoid

1. Chartjunk

Don’t clutter with unnecessary elements:

3D effects that distort data
Excessive gridlines
Decorative elements without purpose
Too many colors

2. Misleading Scales

Always start bar charts at zero:

// ❌ Bad: Truncated y-axis exaggerates difference
yScale.domain([98, 102])

// ✅ Good: Full scale shows true proportion
yScale.domain([0, maxValue])

3. Information Overload

One chart = one insight:

// ❌ Too much in one chart
<Chart data={allData} metrics={allMetrics} />

// ✅ Focused message
<Chart
  data={filteredData}
  metric="key_indicator"
  title="Main Takeaway"
/>

4. Static When Interactive Would Help

Modern tools make interactivity easy:

<ResponsiveContainer>
  <LineChart data={data}>
    <Tooltip /> {/* Shows values on hover */}
    <Brush /> {/* Allows time range selection */}
  </LineChart>
</ResponsiveContainer>

Accessibility Considerations

Provide Alternative Text

<svg role="img" aria-labelledby="chart-title chart-desc">
  <title id="chart-title">Budget Trends 2020-2024</title>
  <desc id="chart-desc">
    Line chart showing budget increasing from $10M to $15M
  </desc>
  {/* Chart elements */}
</svg>

Don’t Rely on Color Alone

Use patterns, shapes, or labels:

// Multiple indicators
<Line
  dataKey="series1"
  stroke="#f97316"
  strokeDasharray="5 5" // Dashed
/>
<Line
  dataKey="series2"
  stroke="#14b8a6"
  strokeWidth={3} // Thick
/>

Ensure interactive charts work without a mouse:

<button
  onClick={() => setFilter('category1')}
  onKeyPress={(e) => e.key === 'Enter' && setFilter('category1')}
  aria-pressed={filter === 'category1'}
>
  Filter by Category 1
</button>

Telling Stories with Data

Structure Your Narrative

Context: Why does this data matter?
Insight: What does the data show?
Implication: What should we do about it?

Use Progressive Disclosure

Start simple, add detail:

// Overview dashboard
<HighLevelMetrics />

// Click for details
{showDetails && <DetailedBreakdown />}

// Export for deep analysis
<DownloadButton data={rawData} />

Connect to Human Impact

Numbers are abstract. Make them concrete:

“500 eviction filings” → “500 families facing housing instability—equivalent to filling every seat in City Hall three times over”

Tools of the Trade

Data Preparation

Pandas (Python): Data cleaning and analysis
dplyr (R): Data manipulation
PostgreSQL: Large-scale data aggregation

Visualization Libraries

D3.js: Custom, interactive web viz
Plotly: Scientific and engineering charts
Tableau Public: Quick exploratory analysis
Mapbox GL JS: Modern mapping

Design Tools

Figma: Mockups and prototypes
ColorBrewer: Accessible color schemes
Google Fonts: Web typography

Best Practices Checklist

The actual goal

Every visualization should answer “so what?” before it goes in front of anyone. If you can’t state the insight the chart is supposed to communicate, the chart isn’t ready. The best policy visualizations I’ve seen make the point obvious — you don’t need to explain the chart, the chart explains itself.

Numbers are abstract until you anchor them. “500 eviction filings” lands differently as “500 families facing housing instability — enough to fill every seat in City Hall three times over.”

Resources

Back to all articles

Making Data Tell Stories: Visualization for Public Policy

Audience first

2. Choose the Right Chart Type

Time Series: Line Charts

Comparisons: Bar Charts

Distributions: Histograms & Box Plots

Relationships: Scatter Plots

Geographic: Maps

3. Use color deliberately

Accessible color palettes

Highlight What Matters

Respect Cultural Associations

Tools and Technologies

For Web Applications

For Static Reports

Case Studies

Municipal Lead Portal Dashboard

OPRAmachine Analytics

NJ Eviction Data

Common Mistakes to Avoid

1. Chartjunk

2. Misleading Scales

3. Information Overload

4. Static When Interactive Would Help

Accessibility Considerations

Provide Alternative Text

Don’t Rely on Color Alone

Keyboard Navigation

Telling Stories with Data

Structure Your Narrative

Use Progressive Disclosure

Connect to Human Impact

Tools of the Trade

Data Preparation

Visualization Libraries

Design Tools

Best Practices Checklist

The actual goal

Resources

Audience first

2. Choose the Right Chart Type

Time Series: Line Charts

Comparisons: Bar Charts

Distributions: Histograms & Box Plots

Relationships: Scatter Plots

Geographic: Maps

3. Use color deliberately

Accessible color palettes

Highlight What Matters

Respect Cultural Associations

Tools and Technologies

For Web Applications

For Static Reports

Case Studies

Municipal Lead Portal Dashboard

OPRAmachine Analytics

NJ Eviction Data

Common Mistakes to Avoid

1. Chartjunk

2. Misleading Scales

3. Information Overload

4. Static When Interactive Would Help

Accessibility Considerations

Provide Alternative Text

Don’t Rely on Color Alone

Keyboard Navigation

Telling Stories with Data

Structure Your Narrative

Use Progressive Disclosure

Connect to Human Impact

Tools of the Trade

Data Preparation

Visualization Libraries

Design Tools

Best Practices Checklist

The actual goal

Resources

About Gavin Rozzi

Related Articles

Media & Speaking