The aspect of David’s offset bars that I was concerned with was the implied vertical passage of time. I wondered whether time should increase downwards as in David’s chart, or upwards as I had visualized myself while first looking at the data. I’m not too concerned about a vertical time axis, though. The measurements are discretely spaced, so it really isn’t confusing, at least not to me.

Generally I think I like the clustered column chart better for showing two series (Democrat vs. Republican by county). As soon as more series are added to the chart, as in the charts shown in my updates, then line charts work best, and I didn’t even consider clustered or stacked columns.

I saw your sparkline chart, but it didn’t really speak to me. Usually sparklines tell a story nicely, but this one was confusing. Red and blue black both refer to one series, even though red and blue are conventionally used for the two series (red=Republican). Also, although the two pieces add to 100%, somehow in this case I’m dissatisfied not to see both values in the chart.

I took a different approach and suggest a sparkline to display the democrat results around the 50% mark.

The line chart does a better job of showing the change in each party’s county numbers.

Yes, I thought of a few adjustments like these, but then it would have taken me ten minutes instead of five to prepare this post! Thanks for the reminders, though.

Update: I had to redo the charts to correct a typo in the chart title, so I lightened the axes and removed the X axis ticks while I was at it.

- Removing the major tick marks on the category axis
- the axis could have some lighter Grey
- potentially you could even remove the value axis

Andreas