FRED Data Visualization

Create indexed comparison charts from Federal Reserve Economic Data (FRED) series.

Quick Start

For a simple two-series comparison:

python3 scripts/fred_chart.py --config chart_config.json --output chart.png

Workflow

1. Identify the FRED series

Find series IDs at fred.stlouisfed.org:

2. Create config JSON

{
  "title": "USA: GDP vs Wages (1959 = 100)",
  "series": [
    {
      "id": "A939RX0Q048SBEA",
      "label": "Real GDP Per Capita",
      "color": "#2E86AB"
    },
    {
      "id": "COMPRNFB",
      "label": "Real Compensation Per Hour",
      "color": "#F18F01"
    }
  ],
  "start_date": "1959-01-01",
  "end_date": "2026-12-31",
  "fill_gap": true,
  "annotations": [
    {
      "date": "1971-08-15",
      "label": "Nixon Shock",
      "position": "top",
      "y": 140
    }
  ]
}

Config fields:

• title: Chart title
• series[].id: FRED series ID
• series[].label: Legend label
• series[].color: Optional hex color
• start_date/end_date: Date range filter
• fill_gap: Shade area between first two series (default: true)
• annotations[].date: Event date (YYYY-MM-DD)
• annotations[].label: Event text (supports newlines with \n)
• annotations[].position: "top" or "bottom"
• annotations[].y: Vertical position for label placement

3. Generate chart

python3 scripts/fred_chart.py --config my_chart.json --output my_chart.png

Advanced: GDP Per Capita Calculation

When FRED lacks a direct GDP per capita series, compute it from GDP (GDPC1) and population (POPTHM):

# gdp_pc = (GDP in billions * 1e9) / (Population in thousands * 1e3)
# Or use the pre-computed series: A939RX0Q048SBEA (available from 1979)
# For earlier dates, manual calculation required

Tips

• Indexing: All series are automatically indexed to their first observation (set to 100) for fair comparison
• Annotations: Use \n for multi-line labels. Position alternates top/bottom to avoid overlap
• Zooming: Use start_date/end_date to focus on specific eras (e.g., 1959-1985)
• Colors: Choose contrasting colors. Blue (#2E86AB) and orange (#F18F01) work well
• Multiple series: Up to 5 series supported. Gap fill only applies to first two

Common Chart Patterns