chart-forensics

Score Breakdown

Your Submission

{
  "issues": [
    {
      "chart_id": "chart-1543178255-1",
      "issue_type": "wrong_height",
      "description": "The bar for 'macOS' shows a height of ~374 (nearly matching iOS at 376) but the table value is 290. The bar is approximately 29% too tall."
    },
    {
      "chart_id": "chart-1543178255-2",
      "issue_type": "inverted_order",
      "description": "All bars have their data values reversed relative to labels. Boston shows 98 (Chicago's value), Austin shows 51 (Seattle's), Portland shows 273 (Miami's), Miami shows 105 (Portland's), Seattle shows 165 (Austin's), Chicago shows 183 (Boston's). The data appears to be plotted in reversed order."
    },
    {
      "chart_id": "chart-1543178255-3",
      "issue_type": "swapped_label",
      "description": "The labels 'Part-C' and 'Gizmo-X' are swapped. The point labeled 'Part-C' shows value 744 (Gizmo-X's table value) and the point labeled 'Gizmo-X' shows 314 (Part-C's table value)."
    },
    {
      "chart_id": "chart-1543178255-4",
      "issue_type": "missing_data",
      "description": "The table has 6 crops (Corn, Rye, Rice, Soybean, Millet, Wheat) but the chart only shows 5 — 'Soybean' is missing from the line chart."
    },
    {
      "chart_id": "chart-1543178255-5",
      "issue_type": "misleading_scale",
      "description": "The Y-axis starts at 130 kWh instead of 0, which exaggerates the visual differences between bars. Apr (206 kWh) appears very small relative to Jul (743 kWh), but the actual ratio is ~3.6x, not the ~8x visual impression."
    }
  ],
  "methodology": "For each chart, I extracted SVG rect/polyline coordinates and computed implied data values using the y-axis scale (derived from axis labels and positions). For bar charts: value = (y_bottom - y_top) * scale. For line charts: value from circle cy coordinates. Then compared each computed value against the corresponding table entry. Identified: wrong_height when value deviated >10% from table, swapped_label when two entries matched each other's values, inverted_order when all entries appeared reversed, missing_data when table had more entries than chart, misleading_scale when y-axis origin was non-zero."
}