Visualising Aged Care Star Ratings up to Q4 FY23-24

The latest Australian aged care star ratings have continued the upward trend. Below I show two animated histograms highlighting how star ratings have changed over time until Q4 FY23-24. The metric considered is the “unrounded overall star rating”. The overall star rating is a weighted average of the ratings in each category, which is then rounded. I simply don’t perform the final rounding step.

index_names = ["Q2 FY22-23", "Q3 FY22-23", "Q4 FY22-23", "Q1 FY23-24", "Q2 FY23-24", "Q3 FY23-24", "Q4 FY23-24"];
html`<h4 id="plot-heading">Australian Aged Care Star Ratings From Q2 FY22-23</h4>`

x_domain = d3.range(1, 5.1, 0.1).map(d => parseFloat(d.toFixed(1)));
barchart = Plot.barY(start_values, {
  x: (d, i) => 1+i/10,
  fill: (d, i) => ((i-20)/20),
  y: (d) => d,
  render: (i, s, v, d, c, next) => {
    const g = next(i, s, v, d, c);
    c.ownerSVGElement.update = (values) =>
      d3
        .select(g)
        .selectAll("rect")
        .transition()
        .duration(750)
        .attr("height", (i) => s.y(0) - s.y(values[i]))
        .attr("y", (i) => s.y(values[i]));
    return g;
  }
}).plot({
  y: { domain: [0, 0.15],
    tickFormat: (d) => `${(d * 100).toFixed(0)}%`,
    label: "Proportion of Ratings"},
  x: { ticks: [1, 2, 3, 4, 5],
      domain: x_domain,
    label: "Unrounded Overall Star Rating"},
  color: { scheme: "RdYlGn", domain: [-1, 1] }
})
data = d3.csv("stars-vis.csv", d => ({
  quarter: d.quarter,
  value: +d.value,
  n: +d.n
}));
start_values = data.filter(row => row.quarter === index_names[0]).map(row => row.n);
{
  let index=0;
  do {
    d3.select("#plot-heading").text(`Australian Aged Care Star Ratings From ${index_names[index]}`);
    d3.select("#plot-heading2").text(`${plot2_stat[index]} of homes had a rating of four or five stars in ${index_names[index]}`);
    const newValues = data.filter(row => row.quarter === index_names[index]).map(x => x.n);
    barchart.update(newValues).end();
    barchart2.update(newValues).end();
    await Promises.delay(1200);
    index+=1;
    if(index==1) {
      await Promises.delay(1000);
    }
    if(index>6) {
      await Promises.delay(1000);
      index=0;
    }
  } while (true);
}

plot2_stat = ["40.8%","47.3%","53.7%","57.8%","62.1%","66.0%","69.2%"];
html`<h4  id="plot-heading2">${plot2_stat[0]} of homes had a rating of four or five stars in ${index_names[0]}</h4>`;

barchart2 = Plot.barY(start_values, {
  x: (d, i) => 1+i/10,
  fill: (d, i) => 1+i/10<3.5?"below":"above",
  y: (d) => d,
  render: (i, s, v, d, c, next) => {
    const g = next(i, s, v, d, c);
    c.ownerSVGElement.update = (values) =>
      d3
        .select(g)
        .selectAll("rect")
        .transition()
        .duration(750)
        .attr("height", (i) => s.y(0) - s.y(values[i]))
        .attr("y", (i) => s.y(values[i]));
    return g;
  }
}).plot({
  y: { domain: [0, 0.15],
    tickFormat: (d) => `${(d * 100).toFixed(0)}%`,
    label: "Proportion of Ratings"},
  x: { ticks: [1, 2, 3, 4, 5],
    label: "Unrounded Overall Star Rating"},
  color: {
    domain: ["below", "above"],
    range: ["#00CEBA", "#432585"]
  },
})

A summary of the current number of each star rating awarded is shown in Table 1.

Table 1: Count and proportion of star ratings given based on the Q4 FY23-24 data.

Star Rating
1	2	3	4	5
Overall
0.0% (0)	0.1% (3)	30.7% (753)	65.5% (1607)	3.7% (92)
Residents' Experience
0.0% (1)	2.8% (70)	49.8% (1230)	42.1% (1039)	5.2% (128)
Compliance
0.0% (0)	0.1% (2)	3.3% (83)	36.8% (919)	59.7% (1490)
Quality Measures
3.6% (89)	7.0% (175)	43.0% (1073)	23.6% (589)	22.8% (569)
Staffing
4.4% (110)	27.0% (675)	42.6% (1063)	13.2% (329)	12.8% (320)

There has been a consistent increase in average star ratings since they were introduced in all categories other than quality measures, as seen in Figure 1.

Figure 1: Average star ratings each quarter for each sub-category until Q4 FY23-24.

Treating the star ratings as purely categorical, we still observe a convincing trend in Figure 2. The number of four and five star ratings have increased while the number of one, two, and three star ratings have decreased.