@Julia - Tableau and Excel are different tools, there are things that are easy for Tableau to do that are hard for Excel and vice versa. In the case of grouped bars & lines sometimes they are easily built in Tableau, sometimes not, it looks like yours is one of the "not" cases. The workbook that @Tharashasank linked to goes through different scenarios based on how your data is structured to get the desired results.
@Tharashasank - next time please link to the original post, there is more detail there: http://drawingwithnumbers.artisart.org/bars-and-lines/
I am slowly working through the very comprehensive workbook - thank you. With the size of my dataset I am not sure reworking the data is an option so am going to have to try and sell the team on stacked bars. Additionally if I duplicate the data all my other worksheets will need to be reworked although could probably manage that with filters. I now need to get my head around explaining why this is so difficult in Tableau.
1 of 1 people found this helpful
I’ve got a few explanations that I use, depending on the audience:
1) Just because Excel can do something doesn’t mean that something is an effective form of data visualization. (Effective defined as something that users can comprehend that accurately represents the data). 3D pie charts are the most egregious example, combo charts with bars and lines very often fall into this category as well. For example in the initial example you posted I can’t tell which line or bar is which and I keep having to refer back to the legend to make sense of the viz. Tableau uses data visualization research to determine what features are built into the product and what they support. Combo charts are generally less intelligible than other chart types so Tableau hasn’t put a ton of work into making them as easy to create as they can be in Excel.
2) Speaking of that ease...Excel is based on cells, so charts in Excel can be made up of arbitrary ranges of cells. This gives Excel tremendous flexibility, but at a cost of performance and maintenance headaches. Tableau uses the structure of the data as the foundation for what we can do in Tableau and that helps us get to insights faster than we can get in Excel. So based on our current data structure if we want a specific output in Tableau we may have to change that data structure.
3) You and your team are changing to some degree from Excel to Tableau. Though they can create the same output, Excel and Tableau are very different tools. To use a car analogy, Excel is like a Honda Civic and Tableau is like a Formula 1 racing machine. A Honda Civic is a very capable car but it’s designed for a different set of uses than an F1. If you are using Tableau to build all same charts that you have built in Excel that’s like using the F1 to go to the grocery store and back. You can do it, but the F1 isn’t really built for driving around parking lots. Why not use the F1 (Tableau) for what it’s good at? This often means taking some time to think about what users are really trying to get from the data and learn what Tableau can do, and then build to that.
I generally try to go with the last bit of #3. As analysts and producers of visualizations we can never go wrong by having a better understanding of our users and what they need.
I agree and enjoy positioning Tableau as the Formula 1 racing machine. In this case I believe side by side bars show what we are illustrating better than a stacked bar but am sure we can get used to that. We are also consistent with colours making the referring back to the legend a necessity if new to the viz but not if you are using it every day.
It is a slow process bringing users to Tableau but slowly and surely we are making progress. This one chart has been the thorn in my side as it is the one KPI illustrated by a chart and not just a spreadsheet with conditional formatting. I will consider a different way to show the information.
Thanks for your help,