When I think Visualization and Excel there are two names that come into my mind: Jorge Camoes and Jon Peltier. If you want to do serious data visualization with Excel, stop here, they are the names. Since I was more familiar with Jorge’s work and had more opportunities to discuss with him I decided to interview him to cover the Excel part of this series, but you can give a look to Jon’s web site if you have any additional questions.
Jorge had been developing visualization in Excel for a long time now and I still remember the time when I saw one of his dashboards in Excel: “Wow, can Excel do that?” Give a look with your eyes to his dashboard courses. Pretty amazing isn’t it?
I have been following Jorge’s blog for a long time now and I often enjoyed his short and catchy blog posts. If you are not following him, give it a try. It’s worth it.
How did you start using Excel?
I started my professional career as a desk researcher. I had to create information products with lots of charts using market and socio-demographic data, and Microsoft Office was the only tool available. Like everyone else, I had no data visualization training, so you can imagine how bad those charts were. On the one hand, that’s very depressing, from a personal point of view. On the other hand, this proves that data visualization skills are easily acquired, once you become aware of what data visualization is all about.
What’s the best and worst aspect of Excel?
We must emphasize that Excel is not a data visualization tool, so you cannot directly compare it to other tools. That said, you can learn and practice sound data visualization principles using Excel. Its chart gallery is poor, but you can make new charts using some more or less clever tricks. Check Jon Peltier‘s site to see how you can extend the Excel chart gallery. So, its flexibility and general availability are the best aspects.
Unfortunately, defaults that emphasize marketing and sales pitch are responsible for a generation of users that don’t really know what a chart is. That’s the worst aspect of Excel. Also, because of it’s flexibility, many users do not recognize that they need stronger data management skills.
How is the learning curve vs. return-on-investment of Excel?
Most business users have access to Excel training. They just need to be brainwashed to remove all they think they know about charts. :) Corporate data visualization culture is so poor that applying simple rules can greatly improve insights and ROI, and you can do it using Excel.
Ok, I am a beginner and I want to learn Excel, where do I start?
What other tools would you recommend other than Excel?
90% of all charts you need in a business environment can be done in Excel. But if it takes a full day to code a chart that you can do in minutes using a different tool you have a good argument to make the switch. I would recommend Tableau, Qlikview or Spotfire. They are well-aligned with currently accepted data visualization best practices and they force you to learn more about structuring your data.
Some comments from Jorge … and my answers:
- J: Business users hate programming. You can’t explain a product manager that a simple recorded Excel macro can make all the difference. You can’t tell them that they need programming skills to make a chart.
E: I think this is totally fine and probably a reason behind the big success of Tableau.
- J: If it can be done in Excel managers will not spend more money getting a new tool.
E: Ok, but then they have to be ready to pay someone to let Excel do the job right? I see a great potential for consultants here.
- J: But managers are becoming aware that they need a serious (visual) reporting tool; Tableau is one of the options; traditional BI tools are moving fast. If a BI tool supports sparklines that’s a good starting point.
E: I think managers will feel more and more pressure as visualization becomes mainstream. I think we just have to wait a little to see some stuff flourishing. I am not too pessimistic.
- J: I believe tools matter, and matter a lot. Tools are not neutral (Tufte says that regarding Powerpoint). If you have to fight them they’ll make your life miserable. Try to apply Tufte’s principles to Crystal Xcelsius. I already wrote about this in my blog using fable about the scorpion and the frog (“it’s in my nature”).
E: Sure, tools matter. Especially if you know how to switch from one to another according to your needs. Nonetheless, I still believe principles come first. And in order to select the “right” tool and understand its limitations you have to have a clearer idea of what you want to achieve.
- J: Life is short: I would argue that it’s better to learn about perception, statistics, data management and graphic design. Delegate the programming part. I’ve been making some dashboards and I spend more time programming than exploring better ways to show the data. Hate that.
E: Cannot agree more. Eve though I think we are still in a phase where it’s really really hard to split between the designer and the implementer. The two things are so intertwined that trying to outsource the implementation may very easily lead to unsatisfactory results. But sure, the real skill is in the design IMO.
- J: If you don’t include R in your list you’ll get into troubles :)
E: Sure! It’s in the pipeline :-)