Building a custom Gauge Table with HTML

Published Categorized as Custom Visualisations, How To's, HTML Content Visual, SVG No Comments on Building a custom Gauge Table with HTML

A quick design for a very specific use case

I wanted to visualise scores similar to the above in a way that was quick and easy for me to read and interpret.

Step 1: The Table

Whilst it is true that SVG can be displayed in native matrix visuals, the last time I checked these rendered square and thus a little impractical for my requirement.

In need of a quick viz, I went to what I was most familiar with. I remembered seeing a worked example from Daniel Marsh-Patrick on how to build a table with the HTML Content Visual, so I moseyed on over to his site to take a look. It was a perfect example so I copied the code and set about modifying it. I added the extra columns and updated the table styling.

<HTML> Linear Gauge Table = 
    VAR SourceData =
                'KPI'[Survey Group]
            "%", FORMAT('KPI'[KPI Percentage], "0%"), "Score", 'KPI'[Gauge SVG]
    VAR TableRow =
                /* column 1 */
                & "<td style='width: 40%;'>"
                & [Survey Group]
                & "</td>"
                /* column 2 */
                & "<td><strong>"
                & [%]
                & "</strong></td>"
                /* column 3 */
                & "<td style='width: 40%;'>"
                & [Gauge SVG]
                & "</td>"
            & "</tr>"
        /* Table Styling */
            table {width: 100%;font-size: 12pt;} td {font-size: 12pt; padding: 3px; border-bottom: 1px solid #ddd;} 
        /* Table and heading row */
        & "<table>
                    <th>Survey Group</th>
        /* Data */
        & TableRow
        /* End of table */
        & " </tbody>

Step 2: Linear Gauge SVG

I wanted to show a score on a scale of -100 to 100. I thought to place a number on a sliding track filled with two diverging colours to indicate positive and negative values.

To do this I drew a rectangle 200 x 16 and filled with gradient CSS styling.
I placed a circle at the midpoint (100,10) and added the score to the x value using a transform function so that it would move left and right accordingly on the scale.

Gauge SVG = 
"<svg width='230' height='22' xmlns='' xmlns:xlink='' display: block' viewBox='-10 -1 220 21' overflow='visible'>
    <linearGradient id='linear' x1='0%' y1='0%' x2='100%' y2='0%'>
      <stop offset='0%'   stop-color='#0056A9'/>
      <stop offset='100%' stop-color='#00A569'/>
  <text text-anchor='right' font-size='10pt' x='-8' y='12.5'>-</text>
  <rect x='0' y='2' rx='3' ry='3' width='200' height='16' fill='url(#linear)'><title>Score: "&'KPI'[Score]&"</title></rect>
  <g transform = 'translate("&'KPI'[Score]&",0)'>  
    <circle cx='100' cy='10' r='10' fill='white' stroke='gray'/>
    <text text-anchor='middle' font-size='10pt' x='100' y='15'>"&'KPI'[Score]&"</text>
  <text text-anchor='left' font-size='10pt' x='202' y='12.5'>+</text>

Whilst I thought this looked nice, I found it difficult to sight the mid-point without putting a ruler to my screen.

I tried again without the gradient and placed a centre-line on the track.


Much easier for me to read the scores 🙂

Step 3: Go nuts

Whilst I had what I needed, my curiosity piqued. I wondered if I could highlight rows with low response rates.

I could.

I added a flag to the row id so that it could referenced in the CSS table styling like so:

<tr class='"&[KPIFlag]&"'>
/* Table Styling */
        table {width: 100%;font-size: 14pt;} td {font-size: 14pt; padding: 3px; border-bottom: 1px solid #ddd;}
        table tr[class='1'] td {background:#FFFFED;}
        table tr[class='2'] td {background:none;}
        table tr[class='3'] td {background:none;}

Neat, but it didn’t help my comprehension. I was then curious to see how I could draw attention to low response rates with another microchart.

I didn’t want a bar as

1) I already had a bar

2) Horizontal bars with percentage fill are often construed as progress bars.

Here I wanted to convey percentage as a whole at a point in time whilst also keeping the table compact. So I went for a pie chart.

Not too shabby.

Leave a comment

Your email address will not be published.