S. Few Renounces Dual-Axis Graphs; Juice Ups Ante
By Zach Gemignani
April 1, 2008
Find more about:
charts
design
humor
After deep introspection, Stephen Few has determined that graphs with dual-scaled axes are fundamentally flawed. Rather than risk the potential for confusion, he believes that there are superior graphing approaches for situations where related data series have different units or magnitudes. His measured and thorough analysis concludes:
“It is inappropriate to use more than one quantitative scale on a single axis, because, to some degree, this encourages people to compare magnitudes of values between them, but this is meaningless.”
I commend Stephen for the courage to start down this path, but he hasn’t gone far enough. Here at Juice, we must often take controversial positions. You may remember that we were among the first to criticize Microsoft’s “databars”, the first to take on the powerful Dashboard Gauge lobby, and the first to challenge the applicability of Tom Davenport’s “Competing on Analytics” sales machine.
While it is true that the second axis can be deceptive, let’s not let the first axis off without asking some tough questions. It is the confusion—nay, the collusion—of the two that causes trouble—who is to say which is the bad seed? We must ask ourselves, do not axes belong in the “Axis of Evil”?
The problem is broader than Stephen suggests: axes are just the tip of the iceberg when it comes to graphic bling that potentially distract or confuse readers:
Take data labels, for example. They encourage users to consider specific values rather than focusing on relative sizes or placement of graph lines or bars.
Legends draw the reader’s eye away from the central storyline of a graphic.
Gridlines… please don’t waste my time with these flat faux-series. One wouldn’t put pinstripping on a Ferrari.
Place your graph in proper context and titles become redundant.
Minimalism is in. Extraneous graph decoration is out. Look no further than Tufte’s sparkline: no excessive graph decoration there.
The world cries out for a new charting aesthetic. One that champions elegance and casts down gaudiness. Let us evoke the pure visual essence of the data. Let us find a pure form to evoke the emotion and hidden meaning of the data. Now is the time for Naked graphs—stripped to the essentials (TM).
Our argument is simple: the visualization of information is the message. The data is but an intermediary form of that visualization. Therefore, any residue from the raw data should be scrubbed from your final graph. Only when you achieve this unadulterated state will the meaning of the graphic burn its way into your consciousness.
Here’s an example of an analysis that casts light on both the relationship of the Fed to hedge funds while simultaneously answering your question about what happened with last month’s sales in the Newark division.
Truly here we see the words of Mark 9:43 made real:
If your hand causes you to stumble, cut it off; it is better for you to enter life crippled, than, having your two hands, to go into hell, into the unquenchable fire.
Gaze in awe, viewers, and find wisdom on this very foolish day.
The Colbert Bump is Real, Colbert’s Nation Not What He Thinks it is
By Chris Gemignani
January 31, 2008
Find more about:
charts
economics
humor
infographics
Stephen Colbert has mentioned that he’s having trouble getting guests during the writer’s stike. We find this puzzling, given the supposed benefits of the Colbert Bump. Does being on the Colbert Show really provide a bump—a critical leap that vaults a writer, or a politician to superstardom?
We know that Colbert isn’t a big fan of “facts,” and only needs his gut to tell him the Colbert Bump is real. At Juice, we let the data decide what’s real or not, so our apologies to Stephen for not taking his word for it. Intrigued, Juice Analytics set out to find out the truth. We gathered data about Amazon sales rank for 20 authors that appeared on his show in recent months. How did those ranks change in the days immediately before and after the authors’ appearance on the show?
Hmmm, there might be something there but those sales ranks don’t tell us much. Fortunately for Stephen, some “eggheads” have worked out roughly how Amazon sales rank corresponds to actual book sales. We calculated the sales, and normalized the data so that the week prior to appearing on the Colbert Report was equal to 1.0. Here’s a picture.
That looks like a bump, Conan. In fact, being on the Colbert Report increases sales by 10 times on average. That bump doesn't last forever, but, let's face it, what does?
We also wanted to know, what kinds of books are Colbert’s audience going crazy for? After all, Colbert is well known as a rock-solid conservative. He’s tight with the Bush Administration. Even though he debates a few liberal (“pinko”) authors now and then, most of his guests are writers of pop-intellectual studies of the Gladwellian persuasion.
Here are the authors and how we categorized them:
Pinkos: Jessica Valenti, Full Frontal Feminism: A Young Woman’s Guide to Why Feminism Matters, Wesley K. Clark, A Time to Lead: For Duty, Honor and Country, Robert Shrum, No Excuses: Concessions of a Serial Campaigner
‘Publicans: Tom DeLay, No Retreat, No Surrender: One American’s Fight
Pop Essayists: Daniel Gilbert, Stumbling on Happiness, Daniel B. Smith, Muses, Madmen, and Prophets: Rethinking the History, Science, and Meaning of Auditory Hallucination, Michael Gershon, The Second Brain: A Groundbreaking New Understanding of Nervous Disorders of the Stomach and Intestine, John J. Mearsheimer, The Israel Lobby and U.S. Foreign Policy, Thomas L. Friedman, The World Is Flat: A Brief History of the Twenty-first Century, Frank J. Sulloway, Born to Rebel: Birth Order, Family Dynamics, and Creative Lives, Jared Diamond, Guns, Germs, and Steel: The Fates of Human Societies, Nassim Nicholas Taleb, The Black Swan: The Impact of the Highly Improbable, Richard Preston, The Wild Trees: A Story of Passion and Daring, Malcolm Gladwell, Blink: The Power of Thinking Without Thinking, Bjorn Lomberg, Cool It: The Skeptical Environmentalist’s Guide to Global Warming, Andrew Keen, The Cult of the Amateur: How Today’s Internet is Killing Our Culture, Michael Wallis, The Lincoln Highway: Coast to Coast from Times Square to the Golden Gate
Popular: Stephen Colbert, I Am America (And So Can You!), John Grisham, Playing For Pizza: A Novel, Tina Brown, The Diana Chronicles
How much of a bump did each of these groups receive?
It’s a shock! Liberals and high-minded eggheads do better than popular or conservative books. I’m not sure if Colbert knows this, but his audience isn’t who he thinks they are.
Here are all the authors and their normalized sales around the time of their appearance on the Colbert Report.
This post was a collaborative effort of the entire Juice team. Pete Skomoroch concocted the idea, wrote copy, and found the study linking Amazon Sales Rank to actual sales. Zach data mined. David May whipped up elegant, instant visualizations. Sal Uryasev munged data.
25 comments | Show all comments only the last 5 are shown
Freccia said:
Great idea...
Valenti, Shrum, Gershon and Smith all had big bumps... did they appear on any other shows at the same time, or was the impact all-Colbert?
As a member of the Colbert Nation, I personally don't watch any other shows, but I've heard that they exist.
Hadley said:
What delta did you use? It isn't listed in the paper you reference.
David said:
Wow - I hope that you realize that Colbert is only pretending to be a 'publican. He is actually a pinko and is very aware that his audience is too. It's the 'publican leaning authors that may be confused.
Stephen said:
It seems clear that everyone has gotten the joke - except you
Rob said:
1) Perhaps you should watch the show - you would then realize his whole act is 'playing' a conservative.
2) Umm.. Don't book authors go on all the chat shows they can usually within a short period. You would need to look at all the other main TV shows each author was on too if this data wee going to make any sense at all.
Chris Gemignani said:
David/Stephen/Rob: Colbert's not the only one who can play a double game. Did you _really_ think we thought Stephen's White House correspondent's dinner roast was evidence that he's "tight" with the White House?
Andrew said:
To be fair to David, Stephen, and Rob, there are people out there whose genes contain a specific mutation that stops their minds from comprehending sarcasm in witty writing. Keep up the good work--it's interesting to see the numbers behind the phenomenon that I already believed was real. There are at least 3 books that I was influenced to buy after seeing the author on the Colbert Report, and I know I can't be the only one watching...
david said:
Quite the analysis!
It will be interesting to watch global trends in book publishing, to see if e-books ever steal the thunder from print books, the way e-music seems to be cutting into "printed" music on CD's.
There's an interesting list of sources of statistics on book sales at Google Answers:
http://answers.google.com/answers/threadview?id=246739
<a href="http://answers.google.com/answers/threadview?id=246739">Global Book Sales</a>
Worth a look.
Pete Skomoroch said:
For some more anecdotal evidence of the Colbert Bump in Amazon rank (including a brief interview with Stephen Himself), check out this video:
http://asap.ap.org/data/interactives/_entertainment/colbert/
Mark said:
Heck, I keep trying to talk my local book stores into adding a 'Daily Show \ Colbert Report' shelf, along with their new releases, top sellers, etc.
Someone needs to get on that. And then also pay me.
dogintub said:
We already knew it was real because Stephen told us, really what you have done is simply confirm that data analysis can detect a force of nature lol
Well done!
Kyle Wegner said:
I'd like to refer back to one of my favorite sayings when looking at data analysis: correlation does not imply causation. There are plenty of factors outside of being on the Colbert show that would "bump" the ratings of some of these authors. To cite a few:
As was stated before, many of these authors were probably on multiple other shows within a short period of time, meaning that it is more of a media bump than a Colbert bump.
I would venture to guess that the majority of these authors are out pimping brand new books right around their release time, which obviously means their rating will shoot up around the same time they are out selling their wares...this is the first time it is possible for them to see an increase.
These are just the first 2 examples that came to mind, but they seem entirely realistic and probably counteract the "Colbert bump" almost completely.
Diego said:
Kyle's right. You could probably include a comparison group to go around the first of his points, perhaps with authors who were on media tours around the same time but did not appear on the Colbert show.
Jason said:
From the second figure, it looks like the median "bump" is by a factor of about 1 (i.e. not a bump), and certainly not more than a factor of 2.
Hamilton said:
Re: Diego: Building a control group would be difficult, because, well, it wouldn't exactly be random.
I will say this, though; my aunt and uncle are musicians who make obscure but NPR-friendly music. They tend not to do ANY sort of publicity for their work, except on NPR. When they do, though, their Amazon sales rank goes through the roof. This even happens when their record has been out for a year (or two). Just a small anecdote.
Chris Gemignani said:
@Jason: The bump is an increase of sales by a factor of about 10. While it's certainly not there for all books, I think it's worth calling it a bump.
More interesting, but not shown, would be data about the quantity of daily sales over the bump period. While we can only estimate it, some of these books entered the Colbert Report with quite low numbers of daily sales (~10 per day on Amazon). The bump is great, but it's only adding a couple thousand extra dollars in these author's royalty checks.
This isn't Oprah's book club.
James Hanley said:
Um, I'm not real familiar with this site, so I'm not sure if you're being tounge-in-cheek when you call Colbert a "rock-solid conservative"? You are, I hope? Or are you satire-deaf?
Ross said:
I'D LIKE TO SEE THE COLBERT BUMP HE HAS MADE ON DORITOS SALES. GOT ANY STATS ON THAT? I BET WHOEVER THE DORITOS DECISION-MAKER WAS HAS BEEN PROMOTED.
Adele said:
I have to echo James Hanley's question...you were joking, right? Colbert's show is a satire on conservative news programs, its no wonder that liberals enjoy it. And Colbert knows exactly who is audience is.
Chris Gemignani said:
Adele/James: Colbert's not the only one who can play a double game. Did you _really_ think we thought Stephen's White House correspondent's dinner roast was evidence that he's "tight" with the White House?
Adele said:
ahhso.
guilty as charged, haha.
jeff said:
since many of these authors appear on colbert the same week/day that they appear on other shows (often/usually as part of a promotional tour for the book/product they're schilling), can this bump truly be ascribed to colbert and his nation?
mike said:
thats a good question, correlation does not necessarily mean causation :)
mike said:
oops already mentioned, perhaps the suggestion of a control group would be best, comparing a media blitz without Colbert Report to those that appear on the show. it would be difficult to separate out the other factors though, like maybe someone that chooses to go on the CR is also more effective in their other promotions. possibly if there were enough data points, then other effects would be insignificant?? ;)
or maybe find someone that ONLY goes on the Colbert Report, a clean sample sort of :D
Aaron Deyfer said:
great article!
one question: how did you manage to get the historical sales rank data? Did you gather the data "manually" using AWS over time or do you use another service?
Add a comment
Analytics Roundup: TIps for showing, sharing, communicating
By Chris Gemignani
December 6, 2007
Find more about:
Business_Intelligence
analytics
business
charts
excel
google
graphics
graphs
powerpoint
presentation
- Developer's Guide - Google Chart API - Google Code
- Beautiful stuff, particularly the Venn diagram.
- Align Journal - BI Worst Practices
- We often see articles on BI "Best Practices" here is an article telling us what NOT to do.
- flot - Google Code
- Attractive Javascript plotting for jQuery.
- ongoing · On Communication
- Interesting blog post about how different forms of communication rank for immediacy, lifespan, and audience reached.
- The Excel Magician: 70+ Excel Tips and Shortcuts to help you make Excel Magic : Codswallop
- SlideShare
- Source for presentation ideas.
Introducing Chart Chooser
By Zach Gemignani
November 20, 2007
Find more about:
charts
excel
powerpoint
tools
Find and Download Great-Looking Excel and PowerPoint Charts
Chart Chooser is an online tool that answers two questions we commonly get:
- What type of chart should I use to show my data?
- How can I make good looking Excel or PowerPoint charts?
Chart Chooser is easy:
- Check the boxes on the left that best describe your objective
- Select the chart that you want to use
- Choose from Excel or PowerPoint downloads to get a formatted chart template
A few notes about Chart Chooser:
- Thanks to Andrew Abela of Extreme Presentations for inspiring Chart Chooser with his “Choosing a Good Chart” post and for working with us to put this tool together.
- We’ve tried to make the charts both Tufte-compliant (i.e. minimal chart-junk) and visually attractive (thanks to Google for the color scheme).
- Feel free to suggest other types of charts that you’d like to see in the Chart Chooser. Send an example to chartchooser@juiceanalytics.com.
- If you’d like a customized version of Chart Chooser for your organization, write us at chartchooser@juiceanalytics.com or call me at 202.251.7750.
23 comments | Show all comments only the last 5 are shown
Clint said:
Not bad for a v1 guys - I especially like the waterfall chart as a funnel visualization. The comparative column and bar charts seem a bit noisy though - might be as simple as using thicker bars and columns, then again might not.
aaron said:
very cool!
why did you exclude the basic line chart from the 'relationship' category but include the two-axis line-column chart?
Chris Gemignani said:
Clint: Sounds like someone's volunteering...
Aaron: In my experience, the basic line charts are used to show the performance a bunch of similar series over time. The line-column is typically used in business to show two aspects of the same thing over time, sort of like showing prices and volumes in stock charts.
MikeW said:
Excellent tool, thanks guys! I've sent it round the office so now everyone can create clear, uncluttered charts. RIP grey background!
grossu said:
Awesome. It would be great if you add Numbers for ac format for uploading.
Tony said:
Guys, great tool much like the Chart Cleaner! I also understand that you aren't going to be able to please everyone.
The good: Excel downloads that show the table templates (specifically for the waterfall chart, which many people have trouble with), color schemes and formating are excellent, interactive selecting is great, inclusion of bullet charts and pushing less chartjunk is a big plus.
Opportunity: I see some downfalls with the options that are presented to me. Much like some of the dashboard tools, it gives me the option of what's available, but does not indicate what is optimal. For example, by selecting Composition, I get everything from bar charts to waterfall charts, to pie charts, to tables. It may be helpful to possibly rank which are the most effective. Why use a pie chart when a bar chart is a better choice. Or, use a stacked bar chart (I find very ineffective) when a line graph is probably better.
I think the objective here is to show what's possible and appeal to the masses. I just question your design when you have previously voiced that some of these break the fundamental rules of data visualization (pie charts).
Jesse Robbins said:
Outstanding work!
Joe said:
This is awesome! Great time saver. Thanks for the great posts and tips.
zaxl said:
Great tool! With Chart Cleaner you open mi mind to a whole new world of graphics. This is an excel-ent addition to my arsenal. Many thanks!
govi said:
Great!
Using it right now for my scorecards!
arun said:
I really appreciate your work... the charts are really cool... I am going to use them at work... thanks guys
Ed O'Loughlin said:
I'm sure it's a great tool, but http://chartchooser.juiceanalytics.com/ makes Firefox on XP fall over.
Stef said:
Hmmm... is it only available for US users? Doesn't work from my place - Switzerland. Neither in Safari nor in Firefox. Just some kind of legal note appears...
Chris Gemignani said:
Stef: The problems you're having relate the DNS propagation. We only set up the chartchooser.juiceanalytics.com address earlier this week and it takes a while for all the far corners of the Internet to know about that address. Not that Switzerland is that far away these days.
If you wait a few days, it should work for you.
Ed: I'll give it a try on Firefox with XP.
Michael Vu said:
thank you for this! it's going to be very, very useful for entrepreneurs and business owners.
Kelly O'Day said:
Nice job!
The line chart uses a legend to identify the 4 data series in the example. Legends add an extra step for chart readers, they have to move their eye back and forth between the lines and the legend. This can interfere with quick, easy chart interpretation.
Why not use series labels to make it easier for the reader? I've modified your line chart file to add a procedure which adds series labels instead of the legend.
You can see it <ahref="http://processtrends.com/toc_chart_doctor.htm#Replace_Legend_with_Series_labels"> here</a>.
Kelly
Kelly O'Day said:
Here's the link again. It didn't work when I used <a href="... <.a>
http://processtrends.com/toc_chart_doctor.htm#Replace_Legend_with_Series_labels
Tony said:
Kelly - Nice job! The link in your post above didn't work, so I just went to your site to find the example. I am a big fan of your changes. I would always opt for series labels versus a legend. They take up less space and make it more visually appealing.
Kelly O'Day said:
This is my 3rd and hopefully final try at getting the link to work.
http://processtrends.com/toc_chart_doctor.htm
Stef said:
Hey there, the website is still not visible from Switzerland! Gush....
Mike said:
Hi guys!
This is fu&%$ awesome! Thank you very much for this!
Tom said:
I love your site and have used several graphs to make myself 'look good' at work. Thanks.
I want to use the Waterfall chart but for the life of me I can not figure out how you remove/hide the color fill from the data points after the first one and leave it in for this one.
Thanks.
Priya said:
Hey thanks for this useful site... I was wondering if there is a write up for different type of charts displayed here, as in what type of data or steps / FAQs etc.
If I am missing something here, let me know
Add a comment
Recreating the NY Times Cancer Graph
By Chris Gemignani
July 29, 2007
Find more about:
charts
excel
screencast
squarepie
tutorial
This New York Times cancer graph is a beautiful piece of work.
I wanted to see if we could reproduce it with everyday tools.
Click here to watch a screencast showing how it was done. Warning the screencast is a little long—14 minutes—and a little unpolished. One cut, no retakes, banzai analytics!
Derek raised an interesting question about how to find the fonts used by the New York Times. While I don't think you can find a high quality free version of these fonts (Helvetica Neue, Univers?), Microsoft has made some very good new fonts for Vista and these are also available to Microsoft Office users through a compatibility pack. Here's a link or google for "microsoft office compatibility pack". I recommend using these fonts.
Here's a version of the graph with these new fonts and more emphasis on getting the typography right.
35 comments | Show all comments only the last 5 are shown
Daniel Waisberg said:
Hi Chris,
the screencast is really amazing. I learned some very helpful tips. Please do others like this one.
Now, there is only one thing that you did not include in your graph (which is a long time doubt I have): the gridlines crossing OVER the bars. I have seen several graphs by Tufte and Phew that show white gridlines crossing over grey bars; I love that! Isn't it possible to do it on Excel?
As for the graph itself, there is one thing I did not like. IMHO, there should be one additional piece of information, something like "probability of death". Not sure if this one would make. My problem is that the graph might be used to say, for example, that prostate cancer has a worrying number of new cases, and deaths might increase. However, since we do not have information regarding old cases to compare to deaths, the simple design might lead us to think deaths will highly increase as a consequence of high number of new cases. Sounds reasonable?
Anyway, thank you very for the screencast.
Chris Gemignani said:
Daniel,
The built-in gridlines pass _under_ the chart elements in Excel rather than over, so it's not possible to recreate the Times' treatment.
Probability of death is certainly an important additional measure, but so is median survival time or expected years of life. You can infer probability of death in a crude way from DEATHS / NEW CASES and I'm happy enough with that.
Cheers, Chris
Tony Rose said:
Nice job Chris. I had to laugh a few times as you were getting some push-back from Excel. Sometimes the first cut is better, or more entertaining, than a polished one... Showing how to format the number so MEN and WOMEN are labeled at zero was very helpful. The NYT did a very nice job setting this chart up. Keep up the great work!
Chris Gemignani said:
Thanks, Tony. There was a lot more mumbling under my breath than you heard. I recorded using Parallels to run Excel on my Mac and the keyboard mapping is a little wonky and the delete key doesn't work. Grrr.
It is so much easier to recreate an existing design than to create a great new design from scratch.
Rupesh Tripathi said:
Easier workaround that space guesswork (to centrally align the category labels), Chris, is to add another column where you can put in the number of spaces required; and use the formula =CONCATENATE(REPT(" ",no. of spaces,category label). Some hit and trials will give the result.For a long list / to start with something scientific before hit and trial, here is the solution:
- Figure out the length of longest label(LEN function).
- For each label, use the formula =(length of longest label- length of current label)/2
- Logically this should put in the correct number of spaces in front of each label.. however due to some reason, (may be the width of space character), the alignment seems slightly leftwards. So I add "1" to the formula above.
- In most cases, this would render center alignment. In the skewed cases, one can overwrite the formula with a higher/lower number and achieve desired result.
Wonderful screencast.
Rupesh Tripathi said:
Small correction - Apologies.
Read the formula =CONCATENATE(REPT(" ",no. of spaces,category label) as
=CONCATENATE(REPT(" ",no. of spaces),category label)
Chris Gemignani said:
Interesting idea, Rupesh. Centering those darn labels was the most time consuming part of this exercise, and they still don't look all that great. The slight leftward bias in my labels also relates to charting internal margins and other Excel esoterica. Again, Grrrr.
Michael Doan said:
Great screencast! Thanks for sharing it. I've been using Excel for awhile now but I've never had much of an opportunity to incorporate graphs. When I do, its often a frustrating experience.
James McMurry said:
Nicely done. Excel graphing masters have their own special kung-fu.
Side note: Assuming you're using a Mac notebook (sounds like it), try using the "fn" key in conjunction with "delete". This should make Windows delete
8 comments | Show all comments only the last 5 are shown
JimL said:
At long last clarity! Now excuse me while I delete several terabytes' worth of software and enterprise data—and stock up on crayons...
Dave Katz said:
What a breakthrough! It's so clear now that sales for the Newark division were actually fine, and the real problem was in Schenectady.
Dave Marcus said:
Am concerned about the the degree to which subliminal messages embedded in the sample graphic encourage terrorism and vegatarianism!
Jeff said:
A S. Few article reference, the ever gratuitous Tufte mention AND a verse from the bible - all in one article, talk about data density...
dave said:
I generally agree with your philosophy of minimizing "chart junk" but I think you may be going to the extreme here.
Most consumers of chart information are not analytics professionals and need help to interpret.
- Data labels: my users demand them. They want to know the value. I don’t think they have a negative impact on “… than focusing on relative sizes or placement of graph lines or bars”
- Legends: are you kidding? Really?
- Gridlines: absolutely necessary for bar charts. I won’t speak for you, but gridlines help my brain orient the chart information.
- Titles: Chart titles and axis labels are necessary and not at all distracting.
Is there not a comfy middle ground here?
| ^
| ^
| ^
| ^
| ^
| ^
| ^^
| ^ ^
| ^
_____________________________________
Dave
derek said:
Dave, check the date that article was posted :-)
dave said:
I'm a moron...
thanks derek!
tao said:
The root of the problem with visualization is that you are using an organ that is simply not meant to understand so much. The only solution to fully understanding data is to not visualize at all but use a direct neural implant into the brain that allows you to quickly grasp all aspects of the data. All this visualization and introspection about visualization is just trying to improve the horse buggy when the automobiles are coming.
said:
Add a comment