The oiginal article came from the BBC and provides an easy to understand overview of the topic and the survey carried out over the years.
The chart we’re making over this week looks like this:
What I like about it:
- A very simple bar chart
- The Title is short and succinct and given the topic, it makes me intrigued
- Subtitle explains the measure displayed and gives sufficient context to help me understand the findings shown
- The colours are distinct enough to make it easy for me to differentiate them easily and they have the legend displayed prominently below the subtitle. They also work well with the images used in the article and with the second bar chart
- The datasource is stated and there is a subtle BBC logo to indicate who created the chart and authored the article
- Well, looks like a lot of stuff was done well, so is there anything to criticize?
What I don’t like about it:
- The title states that alcohol consumption drops in Britain, so it would have been nice to show the drop a bit more clearly. I think vertical bars would relate better to the concept of a ‘drop’ in something
- The subtitle, upon looking closer at the chart, is confusing me: if the survey questioned adults who drank alcohol in the week prior to the interview, why is there a category of ‘do not drink’? Maybe it would be more appropriate to say ‘% of adults questioned on whether or not they drank alcohol in the week prior to the interview’
- I don’t care much about the gridlines, I would prefer them gone and instead have the % values displayed at the end of the bar
- Not having the values for each bar displayed reduces the impact of the chart. To understand the size of each bar, my eyes need to follow the gridlines down to the x-axis, remind myself that each number represents a % figure and then I have to think back to what that % figure represents.
What I did:
- I tried to keep it simple, for real this week :-). I wanted to stick to a doing a genuine makeover rather than engineering additional storylines
- So I created a bar chart which I personally like better than the original, addressing some of the points of critique I noted above
- I also added a line chart above the bars to show how the decline in drinking happened over time
- To accentuate the first and latest years of the survey, I shaded the circles on the dual axis accordingly and carried the font colour over into the labels for the bar chart
- I used annotations to call out the type of response (one vs none vs multiple drinks in the week prior to the survey)
- Lastly, I used a ‘percent difference’ table calculation to show the difference between 2005 and 2016 while adding the actual percentage figures for both years inside the chart to show the magnitude
(Click the image to view on Tableau Public)