See also:

• History of a Chart 1. “Good Times, Bad Times”
• History of a Chart 3. “Where Did All the Rockstars Go?”
• History of a Chart 4. “This is the New Flesh”
• History of a Chart 5. “Guess What?”

(click to see full interactive chart)

Idea

The very first chart I created with this data was a line graph that showed ALL of the genres (more in those early stages before I merged some of them together). The idea was effective but the chart was incredibly cluttered.

A problem that emerged time and time again was one of scale. Pop and rock were such HUGE genres, with large data sets, yet the smaller genres, still important, disappeared among other small genres at the base of the chart.

I then decided to brush up on my Flash (something I had wanted to avoid, as I was hoping to create a larger static visualization) and create an interactive chart that allowed the user to select the genre they wanted, and also be able to see them all together in context.

Method

A chart for each genre was created in Tableau (using the same axis and scale to maintain accuracy)

These were then exported into Illustrator (for re-colouring), and into Flash to build into a rollover chart.

Evolution

The chart above was the result of much manipulation and “playing” in Tableau. Here are some of the more interesting developments during that process.

Inspiration: n/a

Idea: This was one of the first charts I created in Tableau with my data set. I wanted to get a simple overview of the charts and judge the best way to show it. The line version of this remains the simplest way to show the chart.

Method: With Year on the X axis, and Measure Value on Y, and colour set to mark GENRE, this plotted out the values of each genre per year.

Why Ditched: This chart has no real value  it is near impossible to distinguish any flow as there is so much data crammed into a small space.   I reverted to the line version.

Inspiration: n/a

Idea: to show a definite sense of trend over time

Method: To use the POLYGON setting in Tableau to show the overall trends between the various genres over time.

Why Ditched: I am still unconvinced about the benefits of the POLYGON chart. Whilst you can quickly see the trend, it simply shows a line from the start of the chart to the end, and highlights the above and below points. Whilst this may be useful for a definite data set, this is very much a part-data set. I chose to begin in 1970, but that is in no way the start of music. This chart would look very different were I to all music over a different time period. It does highlight a moment where rock and pop switched over in terms of dominance, but overall I decided it was overly simplistic and better suited to financial static data.

Inspiration: DNA charts (example: DNA chart)

Idea: to use a simple heat map to show the presence, and strength of presence, of each genre over time. 3 shades of colour would be given to each genre – not present (grey for each) then dark and light to show the high and low values.

Method: This  was created with separate charts for each genre, to account for the vastly different scales. So it marked the relative values of each genre separately. I then exported the layers, and placed them together in Illustrator.

Why Ditched: I was pleased with this chart and how it looked (I later have doubts about it’s validity – see next chart) but at this stage I was interested to see it displayed in a circle. I hoped I would be able to widen the data areas to make it less cramped.

Inspiration: Jake Kennedy Hobart – The Rain Project,

Idea: I wanted to use the chart above to create a circular, or semi circular chart. I hoped I would be able to stretch out the size of each unit, so it would appear less cluttered.

Method: Using the table created above (and removing any excess white at each side), I used the Envelope Distort function in Illustrator to bend it into a 100% circle.

Why Ditched: I realized that the original method of shading was giving a false impression of the chart. I was comparing each music genre on its own scale, yet placing them next to each other, giving the impression that they were on the same marked scale.  This only shows that pop and rock were a lot more present in the charts than the others, not by how much. A simpler and more accurate chart would have show simple presence, or no presence – colour and no colour. This lighter colour is deceiving.

Inspiration: n/a

Idea: Despite the issues with the colour on the last chart, I was keen to attempt another circular chart using the same process (ie bending a rectangular chart into a curve). Again, I hoped that chart that formerly suffered for lack of space would benefit from a circular shape.

Method: I created a line chart, showing all of the genres (distinguished by colour) on a black background (it showed the lines up clearer than white). I then concluded that I would have to break the chart into quarters, to enable a full 360deg circle to be created (I later solved this problem). For demonstration purposes I created one 90deg bend, and simply copied it 3 more times, and placed the quarters into a circle.

Why Ditched: I was concerned that, as with the simple line chart, the smaller genres (classical, spoken word etc) were being lost at the centre of the chart and therefore were unreadable – rendering the chart useless.

Inspiration: as above

Idea: as above but I hoped  – by reversing the chart and having the smaller values on the outside, wider part of the circle, they would be clearer

Method: I simply turned the chart upside down before creating the circle shape – hence the larger values were on the inside

Why Ditched: Although this did make the smaller values easier to see, it went against logic to have the larger values at the centre of the chart. Naturally the eye would treat the middle of the circle as the bottom of the line chart, hence reading the values upside down. The chart was not useful in any form .

Inspiration: C. Van Vleck | Information Taking Shape,

Idea: Still interested in pursuing the idea of a circular chart, I decided to solve the problem of concealed and cluttered data, by attempting an interactive rollover chart in Flash. This means that even if the data is hidden, the user can make a selection and see it in isolation as well as in relation to the other values.

Method:  I created a full stacked bar chart, with each colour block showing the percentage of albums each genre represented per year. Exporting this as a PDF into Illustrator I was able to recolour the chart (to help distinguish the smaller genres). I then created 8 versions of the chart, each of them with one colour highlighted and the rest “greyed out”. In Flash I converted  the menu boxes into a series of rollover buttons and each showed a version of the chart with the respective colour highlighted. Roll off, and the whole chart was shown.

Why Ditched: I was essentially happy with this chart. It was a little fuzzy about showing the smaller genres and all in all it was difficult to compare respective values (a design fault of the stacked bar chart and its transformation into a circle). However, the main reason for scrapping this chart was the discovering of a major error in the data that would have required redrawing the chart from scratch at a late stage. I took a judgement call that it was not strong enough a visualization to construct again from scratch.