{
  "alt": "Scatter plot of {fact:n_indicators} indicators with the weight on the first component on the horizontal axis and the weight on the second on the vertical axis. Diesel PM, traffic, hazardous-waste facilities and NO₂ sit far to the right near the middle. Storage tanks, lead paint and drinking-water violations sit high on the left or middle, and PM₂.₅ and ozone sit low.",
  "caption": "Indicators at the right weigh most on the first component, which holds {fact:pc1_var} of the variance, and indicators near the top weigh most on the second, which holds {fact:pc2_var}. A dot is an indicator, colored by group, and the data table gives the 95% intervals from a bootstrap over counties. The sign of a component is arbitrary and is set so that traffic is positive on the first and storage tanks on the second.",
  "data": [
    {
      "group": "Air",
      "indicator": "PM2.5",
      "pc1": 0.22,
      "pc1_high": 0.26,
      "pc1_low": 0.17,
      "pc2": -0.47,
      "pc2_high": -0.38,
      "pc2_low": -0.55
    },
    {
      "group": "Air",
      "indicator": "Ozone",
      "pc1": 0.25,
      "pc1_high": 0.28,
      "pc1_low": 0.22,
      "pc2": -0.31,
      "pc2_high": -0.16,
      "pc2_low": -0.38
    },
    {
      "group": "Air",
      "indicator": "Diesel PM",
      "pc1": 0.4,
      "pc1_high": 0.41,
      "pc1_low": 0.38,
      "pc2": -0.02,
      "pc2_high": 0.05,
      "pc2_low": -0.08
    },
    {
      "group": "Air",
      "indicator": "Nitrogen dioxide",
      "pc1": 0.38,
      "pc1_high": 0.39,
      "pc1_low": 0.36,
      "pc2": 0.04,
      "pc2_high": 0.11,
      "pc2_low": -0.03
    },
    {
      "group": "Air",
      "indicator": "Traffic proximity",
      "pc1": 0.4,
      "pc1_high": 0.41,
      "pc1_low": 0.39,
      "pc2": 0.02,
      "pc2_high": 0.08,
      "pc2_low": -0.05
    },
    {
      "group": "Air",
      "indicator": "Toxic releases to air",
      "pc1": 0.23,
      "pc1_high": 0.26,
      "pc1_low": 0.2,
      "pc2": 0.09,
      "pc2_high": 0.28,
      "pc2_low": -0.2
    },
    {
      "group": "Proximity and discharge",
      "indicator": "Superfund proximity",
      "pc1": 0.23,
      "pc1_high": 0.26,
      "pc1_low": 0.21,
      "pc2": 0.22,
      "pc2_high": 0.3,
      "pc2_low": 0.13
    },
    {
      "group": "Proximity and discharge",
      "indicator": "Risk-management facilities",
      "pc1": 0.26,
      "pc1_high": 0.28,
      "pc1_low": 0.21,
      "pc2": -0.08,
      "pc2_high": 0.11,
      "pc2_low": -0.33
    },
    {
      "group": "Proximity and discharge",
      "indicator": "Hazardous-waste facilities",
      "pc1": 0.39,
      "pc1_high": 0.4,
      "pc1_low": 0.37,
      "pc2": 0,
      "pc2_high": 0.07,
      "pc2_low": -0.06
    },
    {
      "group": "Proximity and discharge",
      "indicator": "Wastewater discharge",
      "pc1": 0.22,
      "pc1_high": 0.25,
      "pc1_low": 0.19,
      "pc2": -0.04,
      "pc2_high": 0.06,
      "pc2_low": -0.15
    },
    {
      "group": "Old housing and infrastructure",
      "indicator": "Lead paint (housing built before 1960)",
      "pc1": 0.14,
      "pc1_high": 0.17,
      "pc1_low": 0.1,
      "pc2": 0.45,
      "pc2_high": 0.56,
      "pc2_low": 0.33
    },
    {
      "group": "Old housing and infrastructure",
      "indicator": "Underground storage tanks",
      "pc1": 0.15,
      "pc1_high": 0.26,
      "pc1_low": 0.03,
      "pc2": 0.55,
      "pc2_high": 0.66,
      "pc2_low": 0.33
    },
    {
      "group": "Old housing and infrastructure",
      "indicator": "Drinking-water violations",
      "pc1": -0.08,
      "pc1_high": -0.02,
      "pc1_low": -0.12,
      "pc2": 0.32,
      "pc2_high": 0.43,
      "pc2_low": 0.19
    }
  ],
  "id": "pca_loadings",
  "notes": [
    "Weights are the entries of unit-length eigenvectors of the correlation matrix, so they are the weights on standardized percentiles. A population-weighted version gives almost the same components: the first holds {fact:pc1_var_w} of the variance.",
    "The first component is labeled by its largest weights and the second by its largest positive weights; the labels are descriptions, not findings."
  ],
  "source": "EPA EJScreen 2024, national percentiles of the 13 indicators; principal components of the correlation matrix, each tract once, modeled estimates",
  "spec": {
    "colors": {
      "Air": "cat1",
      "Old housing and infrastructure": "cat3",
      "Proximity and discharge": "cat2"
    },
    "group": "group",
    "highlight": [
      "PM2.5",
      "Ozone",
      "Diesel PM",
      "Nitrogen dioxide",
      "Traffic proximity",
      "Toxic releases to air",
      "Superfund proximity",
      "Risk-management facilities",
      "Hazardous-waste facilities",
      "Wastewater discharge",
      "Lead paint (housing built before 1960)",
      "Underground storage tanks",
      "Drinking-water violations"
    ],
    "label": "indicator",
    "quadrants": {
      "x": 0,
      "y": 0
    },
    "x": "pc1",
    "xDomain": [
      -0.2,
      0.5
    ],
    "xFormat": {
      "digits": 2,
      "signed": true
    },
    "xLabel": "Weight on the first component (urban-industrial)",
    "y": "pc2",
    "yDomain": [
      -0.7,
      0.7
    ],
    "yFormat": {
      "digits": 2,
      "signed": true
    },
    "yLabel": "Weight on the second component"
  },
  "subtitle": "Weights of the {fact:n_indicators} standardized EJScreen percentiles on the first two principal components, {fact:n_pca} tracts, each tract counted once",
  "title": "The first component weights diesel PM, traffic, hazardous-waste facilities and NO₂ most, and the second weights storage tanks, lead paint and drinking-water violations against PM₂.₅ and ozone",
  "type": "scatter",
  "width": "body"
}