Stacked Histogram Alternatives – Finding Missing Trends

Stacked bar charts and stacked bar charts are widely used in the world of information and business. I am sure that their popularity is due to the ability to show a large amount of information on one chart, while creating them is quite simple. However, I am not a fan of these diagrams and will further explain why I think these diagrams are bad and offer several alternatives.

Bad Charts

The figure above shows a stacked bar chart. I consider it a bad chart because it is completely impossible to follow the trend. The chart shows five data series and their total sum. But in reality we can only see trends in two out of six indicators! We see a trend across the data series North America and by the total amount (the height of the entire column), because them baseline is flat. In fact, their baseline is here is H with value $0. This point is very important when creating a chart that should show a development trend (trend). Our eye picks up the baseline as the beginning of each column and then compares the top of each column, moving from left to right. This approach works great for a series of data. North America and for the total amount, but it is very difficult to determine trends for the other four regions.

Stacked Histogram Alternatives - Finding Missing Trends

jagged baseline

Stacked Histogram Alternatives - Finding Missing Trends

The picture above shows how difficult fast see sales vector in Asia (row Asia) because the baseline is not straight. I emphasize – fast, because our main goal is for the chart to easily and clearly convey information and show the trend. It is much easier to look at the available numbers than to try to understand from this diagram whether the desired columns increase or decrease in different periods of time. And if we simultaneously consider all other regions, then the columns of the series Asia completely lost among the columns located above and below. A complete mess!

And here is a graph that shows the current trend for a series of data Asia. Now it’s much easier to see that sales have been declining for the last three quarters. Perhaps this will draw attention to the effectiveness of the region and prompt appropriate actions to correct the situation.

Stacked Histogram Alternatives - Finding Missing Trends

When to Use Stacked Charts

Stacked charts work great when there are significant changes in the data over time and when you want to show a trend across one or two series. It is best to highlight this row with a bright color and add text to describe the trend. This data series needs to be moved to the lowest position on the chart so that it is built from the baseline – so the user can see the series development vector.

Stacked Histogram Alternatives - Finding Missing Trends

Option #2 below shows the use of a stacked chart whose baseline is made dynamic to allow for quick trend analysis.

Solutions

Now that we have an understanding of when NOT to use stacked charts, we can start looking for alternative ways. convey information quickly and clearly. The decision will depend on what trends are found in the data and in what form we want to convey the message to the user.

Option #1: Chart Panel

A chart panel is a group of small charts that are grouped together as a panel. This method makes it possible to show the situation in each region on a separate diagram.

Stacked Histogram Alternatives - Finding Missing Trends

Here we see trends for each region. Trends we didn’t see on the stacked chart due to jagged baselines and chart scale. Sales in North America (North America) and Europe (Europe) are significantly higher than in Asia (Asia) and Africa (Africa), so these two trends are lost against the background of the first two.

Some important points are lost if you do not look at them closer. In the chart panel, we can highlight them with color, as I did in the picture above. It is now clearly visible that sales in Asia are declining, while in Africa sales have almost doubled over the past year. The amounts for some regions are small compared to the total amount of sales, and if you look only at the stacked chart, you can miss such subtle, but very important points.

Option #2: Dynamic Baseline in a Bar or Stacked Bar Chart

A good solution for when you want to keep the stacked chart, or when you don’t want to scare the audience away with something radically different from the usual charts. I call it a stacked chart with dynamic baseline, because the user can use the drop-down list to quickly select which data series to place at the bottom of the graph. Creating such a chart is very simple. You don’t need to work with macros or know VBA. Just two formulas and a combo box control.

When the data series West is at the top of the column, it is hard to see any dependency. But as soon as a row West moves to the very bottom of the column to the baseline, it immediately becomes clear that the last six months there has been a steady decline in sales.

Optionally, you can also see the relationship between data series South и West in the second half of the year, if you move the series East up the column. In a word, the data on the diagram must be arranged in such a way as to convey maximum information to the reader.

Stacked Histogram Alternatives - Finding Missing Trends

Option #3: ???

I would be very happy to hear from you what other alternatives are possible. Let them be simple or complex – as you wish, there are no right or wrong decisions. Remember that the main the goal is to clearly and quickly convey information to the user.

Thank you!

Leave a Reply