Discussion


Note: Images may be inserted into your messages by uploading a file attachment (see "Manage Attachments"). Even though it doesn't appear when previewed, the image will appear at the end of your message once it is posted.
Register Latest Topics
 
 
 


Reply
  Author   Comment  
sfew

Moderator
Registered:
Posts: 803
Reply with quote  #1 
My friend Howard Spielman says that business people today must be competent in three important areas to succeed: (1) literacy (the ability to reason and communicate verbally and in writing), (2) numeracy (the ability to reason and communicate quantitatively), and (3) graphicacy (the ability to reason and communicate graphically, that is, with images). I believe that businesses are lacking in all three competencies today.

While literacy and numeracy have been part of educational curricula forever, ability in these areas seems to be declining in America today. I believe that this is partly due to the way schools have been hamstrung by the "No child left behind" requirements that force them to teach to the test, rather than teach in ways that best support real learning.

Graphicacy has never been a part of educational curricula, but now needs to be. Using visual representations of data interactively to help us think about information and to find meaning in it, and using graphics to communicate meaningful patterns, trends, and exceptions in the data to those who need that knowledge to make informed decisions, is an important competency that deserves a place in every student's educational diet. Human beings possess powerful visual perception. When information is presented effectively in visual form, insights can result that would be difficult or impossible to achieve otherwise.

The U.S. government has recently recognized this potential and is funding research in information visualization through a program called the National Visual Analytics Center. Businesses are becoming more and more aware of this potential also, but have become prey to software and techniques that look impressive on the surface, with lots of flash and dazzle, but often fail to support effective thinking and communication. People need to learn the potential of visual analysis and communication and to recognize the difference between visualizations and techniques that really work and those that are designed to impress but yield no results.

How can we improve graphicacy? Let's think it through together.


__________________
Stephen Few
camoesjo

Registered:
Posts: 32
Reply with quote  #2 
You cannot improve graphicacy when, for most users, graphicacy means knowing how to create a chart in Excel. Probably 99% of Excel training courses teaches you this, without any warning on how bad designs should not be used, in spite of its availability in the options. If someone teaches you how to use a pencil, is that enough to write a novel? Or even a short sentence?

This is a side effect of bringing specialized knowledge to the masses. You can see this in various fields. We live in some kind of template culture, where the message is "you don't really need to know those boring details. Just fill in the blanks and, voila!, you will get a wonderful [put something here]".

This is not necessarily bad. In the short term, it will raise awareness. In the long run the overall quality will improve.

Or not.
sfew

Moderator
Registered:
Posts: 803
Reply with quote  #3 
Camoesjo,

The points you've made are very much in line with the message that I often preach in presentations, classes, and articles. I believe that one of the side effects of the PC is that people now tend to believe that if they know the basic mechanics of using software to do something, they know how to do that something that the software was designed to help them do. Unfortunately, however, knowing the basic mechanics PowerPoint does not make one a good presenter, knowing Word does not make one a good writer, and knowing how to make a chart appear in Excel doesn't prove graphicacy. In fact, most of the charts that people produce with Excel demonstrate the opposite--the fact that they don't know how to communicate graphically. I find that a basic level of graphicacy can be learned quite easily by most people given the right resources, but learn it they must, for graphicacy, much like numeracy, is not intuitive.

Thanks for contributing your thoughts to this forum.


__________________
Stephen Few
PSu

Registered:
Posts: 32
Reply with quote  #4 
Stephen:

Being a relatively new member and absorbing all stuff regarding graphicacy recently, I would like to add some reflections on the use of Excel.  I cannot agree more with camoesjo's statement that Excel users don't seem to possess in fact graphicacal (is this English?) skills.  But is Excel to blame?

Making a straight-forward chart is easy, but to most people the result is not particularly appealing. Indeed, I agree that choosing the wrong type of chart or charts with distracting clutter seems to happen more often than not.  In discussion fora one can see that questions are mainly technical: how to do this or that (sometimes straight-forward, but often requiring some chart tricks), where the responding community almost never asks why this is needed, or suggests an alternative path to bring across the message (e.g. a broken y-axis can be done technically but isn't recommendable when using column charts).

In my view Excel is a powerful tool for analysis.  Charts are just a way of producing outputs and to "show the numbers".  Many Excel users have only limited knowledge of the tool; and anything Microsoft can do to make things easier should be done.  This holds particularly true for charting.  In this respect I agree with Stephen Few that Excel 2007 isn't particularly an improvement.  More help should be provided to the average user how to bring their message across effectively, rather than providing an overwhelming amount of "beautifying" possibilities.  But I believe that Microsoft just responds to what they see as what the market wants.  They see dashboard charts now becoming popular, a booming business indeed.  And this market segment is highly driven by eye-candid software, where Excel (2003) cannot deliver.  Hence, their focus on beautifying, adding glare, shadowing, transparency layers, etc can be explained.

While Excel 2007 maybe disappointing from a data visualization point of view, critique will fall on deaf ears (or even worse: it may be regarded as "cursing in church") with the majority of people.  From a marketing point of view, I believe Excel 2007 is an opportunity: the need for good advice will be even higher than before (more clutter/options = more need for guidance).  I am convinced there is a need for "Effective (Dashboard) Charting in Excel 2007" where your messages on effective data visualization are combined with Excel charting tricks.  Providing Excel templates for your bullet graph and spark line concepts (e.g., make instructions how to make them downloadable) to make it easy to use them, would also be helpful to support your mission to increase graphicacy among people.  (Note: I made bullet graphs according to your design ideas on Excel 2003 already; it's doable but not that obvious).
I would be happy to collaborate with others on integrating your ideas into sensible Excel use (v 2003 / 2007).  If this idea sparks some interest, please let me know.

Vbr,


__________________
Henk
sfew

Moderator
Registered:
Posts: 803
Reply with quote  #5 
Henk,

Given your interest in data visualization and your expertise in Excel, there are some folks that I'd like to put you in touch with who share these interests. The first in an Excel afficianado named Charley Kyd, who authored a book on Excel-based dashboards. You can learn about Charley and is work at http://www.ExcelUser.com. The others are two brothers, Zack and Chris Gemignani, of Juice Analytics. They have a passion for data visualization and have done some interesting things with Excel. You can find out more about them at http://www.JuiceAnalytics.com.

Take care,

Steve


__________________
Stephen Few
PSu

Registered:
Posts: 32
Reply with quote  #6 

Stephen:

I am not particularly an expert Excel user, although above average I guess.  Thank you for your suggestions.  I have bought Charley's book some time ago and indeed he is pretty far in producing sensible dashboards with Excel.  Good idea to contact him.

The brothers Gemignani I didn't hear about until you mentioned them.  I will contact them too, and see whether we can join forces.

 

Best regards,


__________________
Henk
PSu

Registered:
Posts: 32
Reply with quote  #7 

Folks:

Today I spoke to a German guy who disliked Radar Plots as do so many of us.  Interestingly, he mentioned that he took cognitive psychology classes "ages ago" where he referred to research which confirmed that bar charts are far more effective than radar charts in data visualization.  He mentioned that the research involved presenting people similar data displayed in radar plots and displayed in bar graphs. Data was presented to two groups (one for bar chart, one for radar chart) and afterwards both groups were asked what they remembered as most significant.  The results showed the bar charts message was significantly better remembered than the same data in a radar plot. 

 

Although I don't have the actual research findings at my disposal (I am scientific enough to treat this "hear say" with appropriate care and cautiousness), the findings do support the view that bar charts are better than radar charts.  This triggers me: are there more of this kind of research findings that substantiate what Stephen and others teaches us?  It will make the job of convincing people to drop their wrong habits of producing chart junk, pies, etc.  Anyone?



__________________
Henk
camoesjo

Registered:
Posts: 32
Reply with quote  #8 
Henk

There is a lot of research. In Google Scholar, try some names, like Lohse, Shah, Tversky, Lewandowsky or Scaife (put also "graphs" or similar words in the search string). And of course you have authors like Cleveland, Kosslyn or Bertin. And, now for something completely different, you should try Nigel Holmes also, just for fun. Not all research supports Few's or Tufte's views.

What I find interesting in Tufte is that he builds a theory of graphs based (among some other things) on some very strong aesthetic principles from minimalism (Mies van der Rohe). In the end, the result is very coherent and it gives you a strict and normative way of looking at graphs ("maximize data/ink ratio"; "maximize data density"; "no chart junk", etc.). But design is, by its very nature, a subjective answer. You can't always back it with scientific evidence. You can't anticipate all the logical consequences of your design theory and some of them will collide with scientific findings, sooner or later. And, of course, usually you can't convince the business community with some "academic" papers ("well, what do they know about business, anyway?").

This is again a problem of "graphicacy". If you don't know how to use graphs for data analysis, you must have some training on graphs but also on data analysis. Then you will understand what a graph is useful for and how it should be designed.

Look at art. If you don't have the specific "codes", you can't "read" the painting, or the music. So, you will like Vivaldi, but not Karlheinz Stockhausen. You will like Monet, but not Pollock. You will rely on your experience to decide if you like them or not. But if you spent all your life with Monet and Vivaldi, the winner is obvious. If you spent all your life with Powerpoint presentations what will you choose?

Without formal training (or no training at all), the charts you see in the media are the ones you will try to mimic. They are your role model and they are eye-catching. If you're not a designer, you'll select a template that looks like the ones you see in the media or you try something from scratch, usually with catastrophic results. But, with or without training, if you are presenting something to your peers (let's say, product managers) in your company, and you know your audience, and your boss is there also, you know they are expecting something in some specific format. So, you must be strongly convinced (and strongly convincing) that any alternative format will be clearly understood and welcomed. If you don't, of course you're running into troubles.

Jorge

sfew

Moderator
Registered:
Posts: 803
Reply with quote  #9 
Jorge -- You've expressed the situation thoughtfully and articulately. The situation is definitely complex.

Henk -- A huge body of research exists that is relevant to data visualization. Some of it is excellent, and some not so good. I try hard to stay on top of this work and to apply what has been learned through the best research to the principles that I teach. Although I appreciate Tufte's minimalist approach in most cases and advocate it as well, I try to remain open to useful exceptions to this approach, especially when the research suggests useful alternatives.

Stephen

__________________
Stephen Few
PSu

Registered:
Posts: 32
Reply with quote  #10 

Jorge - Thank you for your extensive answer.  I appreciate the time you took for making this clear to me (and everyone else who read it).  I do agree to it at a large extent. 

I am aware there is research going on.  What I was trying to say is that if to ban out for instance radar plots as in the example I gave, hard research evidence can hardly be beaten.  But this requires references to this work; just saying "research suggests..." or "from many research is is clear..." is too much used by politicians, ....so is usually not very persuasive.  If, on the contrary, I have a research paper in my hand proving my statement, it is stronger, and if I have more than one even the better. 

 

The more I am digging into this intriguing world of 'show me the numbers', the more I like it.  Graphicacy is largely about putting your CAPE on: Clear, Appealing, Persuasive, Educational - but I guess other people can find a better and more descriptive acronym. 


__________________
Henk
nixnut

Registered:
Posts: 69
Reply with quote  #11 

Quote:
Originally Posted by camoesjo
Without formal training (or no training at all), the charts you see in the media are the ones you will try to mimic. They are your role model and they are eye-catching. If you're not a designer, you'll select a template that looks like the ones you see in the media or you try something from scratch, usually with catastrophic results. But, with or without training, if you are presenting something to your peers (let's say, product managers) in your company, and you know your audience, and your boss is there also, you know they are expecting something in some specific format. So, you must be strongly convinced (and strongly convincing) that any alternative format will be clearly understood and welcomed. If you don't, of course you're running into troubles.

I just had a somewhat disturbing thought. What if the problem isn't one of not just knowing how to use graphs to communicate and therefore falling back to doing what everyone seems to be doing? What if there is a gap between the feeling of the reader of being informed to his satisfaction and the actual amount of information that was transmitted successfully?

What I mean is that perhaps people are inclined to feel more informed with glamorous graphs and feel more satisfied with such graphs, while in fact they retain less information from them than they think they do. And writers, editors, publishers etc may be more concerned with reader satisfaction than communication of actual information.

This could be another explanation for the rampant junk-chart production in today's media.

 

If this thought holds any truth, then the problem lies not just with educating creators of graphs (and tables), but with the readers too. Then again, properly assessing how to satisfy your readers is part of the process of communicating. So, it might just be yet another factor to take into consideration.

 

Back to graphicacy: I think design and use of tables and graphs should be taught in high-school. It can so easily be integrated in lots of classes and applied to many subjects. The basic principles are easy enough and the earlier you are introduced to them, the more you can benefit from them.  

sfew

Moderator
Registered:
Posts: 803
Reply with quote  #12 
nixnut,

You are absolutely right. Consumers of information need to be educated as well as the producers and designers of information displays. I also agree that graphicacy should be taught in high school, but believe that beginning in elementary school with the basics would produce the best results. I've spoken with educators of the young who whole-heartedly agree.

Thanks for your insightful comments.

__________________
Stephen Few
camoesjo

Registered:
Posts: 32
Reply with quote  #13 
Let me quote a study (Meyer et al., "Multiple factors that determine performance with tables and graphs," 1997): successful use of visual information depends substantially on its acceptance by the manager and the environment" and the superiority of visualization for better decision making is a "naive superiority hypothesis".

If a manager doesn't feel the need for more information than what is displayed in a pie chart, why should you try do give him more? Answer: you must show him the need. Not the need for better charts, but the need for better insights. I am using an agency market research presentation (ugly and irrelevant) to show (not tell) that, for each question, the company is loosing money because there are no insights, only some stupid charts.You must force comparisons between what they have (from the agency) and an alternative (better, hopefully) display and ask them: "which of the options would you choose to support a decision? Why?" And I ask them questions that force a global interpretation of the displays ("where would you put a new sales representative?"), something that takes for ever with the agency charts but can be answered in no time with the alternative display. And sometimes I get "wrong" answers that make sense....

I am learning a lot.

sfew

Moderator
Registered:
Posts: 803
Reply with quote  #14 
This is one of the biggest challenges that we currently face: convincing people who have preconceived notions of how data should be presented that there are better ways--ways that provide better insights , resulting in better decisions. It is easy to demonstrate by showing them both ways side by side--the bad and the good--and encouraging them to test the difference for themselves, but this only works with people who care enough to actually open their eyes and look. Much of my time is spent making this case through the use of examples.

__________________
Stephen Few
Previous Topic | Next Topic
Print
Reply

Quick Navigation: