## Good to Great; Quadrant Analysis

Recently I was performing some Quadrant Analysis at the Leicester Viz Club (see www.vizclubs.com) when the guys asked if I’d do a blog on how it was done as they’d find it useful. So without further ado, here’s my latest video talking you through how to create the viz below in Tableau.

See the video below to see details on how to perform this analysis.

Tableau Good to Great: Quadrant Analysis

Nottingham, UK

### 8 thoughts on “Good to Great; Quadrant Analysis”

1. Hi Chris,

A very nice post. Thank you.

Towards the end of your video it occurred to me that, with this particular data, since these marks form a very distinct diagonal pattern, that a trend line with confidence bands might be a nice way highlight the marks for further examination.

Those would be the ones either towards the top or the bottom of that diagonal, which have increased or decreased in profit from year to year. My thought was that using confidence bands will give us a modest statistical method for identifying which marks in particular warrant investigation (because they’ve deviated sufficiently outside of the confidence interval from the overall trend).

To be fancy, I experimented with using a polynomial ^2 trend model (instead of linear), which was fun to do.

Again experimenting, I also found that it is quite nice to shade the quartiles. This helps us to more statistically identify which marks warrant attention for being exceptional over-all winners or losers.

This I did from the analytics pane by dragging “median with quartiles” onto the view, and then hiding the median reference line to leave only the quartile shading.

Curious, both of the median lines land at the ~52 percentile (instead of landing closer to 50% exactly) ?

Anyway, just thoughts and experiments! Thank you for your continued dedication to producing these great videos!

2. Deb says:

Excellent presentation

3. Mike Kainrath says:

I have version 10 and am new to Tableau. When I type in the colour calculated field

IF [Familiar Percentile] <=0.5 and [CC] 0.5 and [CC] <=0.5 then 'Bottom Left'
ELSEIF [Familiar Percentile] 0.5 then ‘Top Left’
ELSEIF [Familiar Percentile] >0.5 and [CC] >0.5 then ‘Top Right’
ELSE ‘Error’
END

I get an error message that says, “Cannot mix aggregate and non-aggregate arguments with this function” and it underscores the word “and” in the calculation.

Any idea of where I’m going wrong?

4. Mike Kainrath says:

Of course after sending this I figured it out. I was not using [CC Percentile] in place of [CC]

5. Václav Hofman says:

Thank you for excellent presentation!

I have a question – is it possible create variable “Colour”, if I have a dynamic axis X?
I want when I change axis X (by year for example), “Colour” wil be changed too – in all quadrants stay circles only with relevant colour.