It’s my last week at Oxford University and as a Data Studio blogger. From next week, I cross the divide and become a Tableau employee. So far, my Tableau journey has been a blast, and I wanted to look back at my life as a Tableau customer.

I’ve been at Oxford 4 years, and looking through my notebooks, I found my first written note about Tableau. Turns out this was in prep of a business case to convince my managers to purchase one license of Desktop. Below is what I wrote. The key points are still totally applicable to anyone looking for a great visual analytics tool – speed, good quality visuals, ease of use, and sharability. Looks to me like Tableau has stayed true to its mission for the past four years (click the image for a slightly bigger version):

A page from my work note book, sometime in 2007

Since those early days, we evolved pretty quickly and expanded across the university. I soon realised that I wanted to become more involved in the user community and organised the first UK Tableau User Group in July 2009. 40 people turned up and this event seems to have gone down in folklore as the first ever Tableau User Group. Andy Kriebel claims he organised a meeting of Tableau users prior to that one! Either way, Andy K and I were early user champions, and it’s been a thrill to be at the front of the tidal wave of Tableau growth in that time.

By November 2009, Tableau was beginning to become a hobby as well as work – I was finding data online and using Tableau to visualise this. In that monthI made my first comment on the Guardian Data blog (these ancient days were Before Tableau Public).

I had also started putting vizzes online at a posterous blog. This blog never really took off but after meeting Mel at the user group and other Tableau events, I started blogging here at The Data Studio. The floodgates opened and over the last 12 months it’s been an absolute pleasure coming up with new ways of using Tableau and doing exciting things with public data. I’ve managed 50+ posts, about one a week.

What have been my highlights? A few things stand out.

Cycling to work is a great way to generate ideas, and it was on rides to work that the ideas for bar charts in tooltips and lollipop charts (three posts!) came together. I love seeing them out in the wild.

Pouncing on topical public data is great fun, and I was delighted that I got my viz about London’s Cycle Hire scheme embedded on the Guardian Data Blog. It was a tactical dashboard design – I don’t think it’s the best dashboard one could have made, but the data was hot and many people pounced on it the day it came out: being first was possibly more important than being best in terms of exposure for this blog and for Tableau.

Along the way I’ve also had a bit of fun making 3-d pie charts and Christmas trees

For the more complete list of my blog posts, click here. Did you have any favourites?

Leaving the University of Oxford and this blog behind was a tough decision, but I’ve now got the opportunity to continue doing what I love all day every day – I can’t wait. Talking of my Oxford colleagues, I had a leaving presentation yesterday and they presented me with a card decorated with the following dashboard – clearly I have taught them well! Click the image for a larger view:

I also need to thank you all for reading  – your feedback has been fantastic and it’s been a hoot to meet you online or at conferences and geek out over data viz. Rest assured, that will continue once I’m embedded in Tableau!


Croydon Cyclone 2011

This weekend saw the Croydon Cyclone disc golf tournament. It was the biggest UK tour event for a long time, and the Croydon club created a tough course. Here’s the analysis of each hole’s results. As you can see, it was a tough event. Click on the holes in the upper chart to see score distributions in the lower section.

What do I take away from this?

This was a tough course. Check out how many holes had more bogeys than birdies. For the Open, it’s about half of the holes. For the other divisions, just about all holes were bogeyed more than birdied. In fact, the Int Ams only managed 28 birdies in the hole tournament.

What to make of the new and amended holes?

Hole 7, with an OB “lake” right in front of the tee seemed too easy for the top Open players. I think this dashboard proves that. The Open Division managed, on this hole: 36 birdies, 30 pars and only 10 bogies. The Advanced Ams were evenly split (14,42,20), so it played well for them. The Int Ams struggled (1,28,28).

Hole 9 saw proportionally more Open player bogey than Adv Ams, proving that we have too many open players who overestimate their driving and approaching accuracy!

I will leave you to explore the rest of the holes. Feedback welcome…

Panel charts in Tableau

Over on the forum, Jeremy asked about making trellis charts. For more info on trellis/panel charts, John Peltier has a great article about them. Consider a case where you want to visualise two measures (eg sales and time) for a Dimension with many members, such as State. There’s no pretty way to do this in Tableau by default. Because there are lots of members in the State dimension (up to 50), you’ll end up with one of four things:

  1. A crowded line chart
  2. A very wide chart
  3. A very tall chart
  4. Something using filters to show just a few states at a time.

None of the above are ideal. Instead, you can use the principles of small multiples to arrange your dimension members into a grid.  The workbook below shows the results.

Tableau loves small multiples. I’ve blogged previously about my most satisfying creation, which used small multiples. Tableau likes small multiples that are arranged with separate dimensions on the Rows and Columns. What we’re trying to do here, though, is arrange one dimension’s members arbitrarily on a grid. The data itself contains no information that tells Tableau which row or column it should appear in. Fortunately, we can use table calculations to create that information.

1. Create an index on the Dimension that you are arranging into panels

This is nothing more than an INDEX() function, set to compute using [Customer State]. It gives each State a numerical value, rising alphabetically.

2. Create a parameter to determin the number of columns

In the view above, you can change the number of column via the Parameter slider.

3. Create a calc for the Column value

To work out the correct column for each State, the calculation uses modular arithmetic.

The calculation has to check if the modular result is zero – if it is, then that member actually needs to be in the right-most column.

4. Create a calc for the Row value

For this calculation, we don’t use modular arithmetic, but use a ceiling function based on the [index]/[columns count]. Tableau doesn’t have a native CEILING function but Alex Kerin and Joe Mako came up with some on the Forum.
That’s all we need. You can then build the worksheet as follows:

I’ve done a bit of tweaking by hiding the Row/Col value headers – they are relevant to you as the report writer, but not the viewer. I also labelled the line ends to enable identification of each panel.

Critiquing other people’s Tableau visualisations

Tableau dashboard with and without design

Tableau’s sample files wouldn’t look so good if they weren’t well designed

Alistair Knock posted a great question by Moose Peterson on Google+  about feedback on user photographs on the web:why do most people just make anodyne comments along the lines of “Nice!” when they probably have some constructive criticism they are holding back?

This got me thinking. How do I respond to seeing Tableau Public workbooks?

The more I use Tableau Public, the higher and higher I believe the bar has to be set for a successful Tableau Public visualisation. You must think like a web designer when publishing to Tableau Public. Layout, fonts, interactivity, colour, etc: these consume vast tracts of web designers’ time for a reason. Just as a badly designed web page is a disastrous advert for your brand, so too is a weak Tableau Public viz. I’ll be exploring these issues in more detail in my session, “Expose yourself with Tableau Public”, at the Tableau Customer Conference in Las Vegas in October.

An anodyne retweet of my own

What is my own reaction to seeing Tableau Public visualisations? If I like it, I’ll comment on the blog, or tweet it. Good work deserves to be praised. What about my critical thoughts? Believe me, I see areas for improvement in almost all the visualisations I see; some recent popular Tableau Public vizzes by well known members of the Tableau community have been, in my opinion, really badly designed. Have I told them? No.

If I think your workbook needs redesigning or tweaking, generally, I say nothing. Why is that?

Because I appreciate effort – at least the author is using Tableau, and having a go. I wouldn’t want them to stop that. I think most authors just want to get their data out there. They aren’t interested in someone they probably don’t know dissing what might have taken them hours to create.

Because I question my own opinions: on whose authority am I an expert in Tableau Public? What if my opinions are wrong? I don’t get much criticism of my own work; I get almost universal praise. This is brilliant, but what if everyone else is also just hiding their critical thoughts about what I do? I think I’m pretty good at this Tableau-data-viz thing (actually, I’m quite confident I am good at it), but I don’t know for sure. Stephen Few is famous for his approach of calling things exactly as he sees them. If he has a problem with your work, he’ll share that with you and with the world, and justify those opinions. As a result he is widely read and widely respected. He has also stimulated many great debates as both his supporters and detractors explore the issues he raises.

Because I’m not sure I can express my opinions in a constructive manner. Stephen Few writes excellently, justifying his opinions. I am not sure I can critique well enough to cause the author to reflect rather than get angry.

Reading through the many thoughtful comments on Moose Peterson’s G+ post, I think I should rethink my approach.

What do you think? Would you be comfortable critiquing others’ Tableau Public workbooks? Would you behave differently on the internet as you do face-to-face? Would your freedom to critique others change when you become a Tableau employee (as I will be in 6 weeks)? How would you react if I started saying “nice viz, but your dashboard layout’s all wrong”?

JISC: how an offensive tweet helped the data viz revolution

Back in November, JISC InfoNet published the results of its survey of BI solutions in Higher- and Further- education. A snaphot is below.

Being someone who is passionate about good quality visualisations, I rattled off a couple of less than complimentary tweets about this. Here’s one:

JISC got in touch with me privately and they were rather upset at my offensive comments. I apologised on twitter the following day, but behind the scenes, we explained why 3-D exploded pies are not a good idea, pointed them to some books (Stephen Few, Tufte, etc) and suggested that Tableau could achieve better, cleaner results. That cleared the air, and was the last I heard about it.

Until a couple of days ago.

JISC got back in touch to say they are now using Tableau for their visualisations (click here to see their first workbooks) I am absolutely delighted about this result. They have come up with some great visualisations, and have even blogged about the experience here. I feel totally relieved, too, that what started with a dodgy tweet has ended up with more converts to the ways of better data visualisation!

Green pastures

I have a new job, starting in late September. It’s with a technology startup you may just have heard of.

Yes, I am moving to Tableau.

I’m very excited and in many ways it’s the least surprising job move ever. I’ve been amazed by the product since first installing v3.1 * and have always harboured a desire to work for this fantastic, growing company. I’ll be a senior product consultant, helping grow the company and share its many benefits with the world. That is pretty much what I do already for The Data Studio and the University of Oxford, except now it’ll be my full-time job.

Between now and then, I’ll try to do a few more blog posts, in an attempt to hit 50 in total. It’s been fantastic blogging and developing the world of Tableau from the user perspective, and now I get to do the same from an employee perspective. I intend to be just as active in the social media scene, so you’ll still be hearing from me.

* incidentally, can you believe that when I started using Tableau, there was no Server product, and you couldn’t do the following:

  1. maps
  2. customise tooltips
  3. actions
  4. customise titles
  5. dual axes
  6. extract data
  7. keep your Excel file open while Tableau was connected to it
  8. maintain Print settings between sessions (oh, you still can’t do that – ah well, I will fight for it internally)

Jerome’s Tableau Viz Contest entry

Jerome Cukier has put together a great viz for the current Tableau vizualisation contest. Here it is, with my thoughts below (his original blog post and commentary are here):

The dashboard is excellent. It is technically superb and the story is clear.

Would I change anything? Not much,but I thought i’d list some comments. Don’t take the length of this list as a comment on the quality of the viz – I am trying to speak as one expert to another!

  1. Each of the panels on the right is a filter, but that’s not immediately obvious to non-Tableau users. Maybe put a Text panel saying “Filters” at the top of those four panels, and put a border around their layout container? I do something similar here.
  2. The “Switch to map view” is really nice. You could achieve that within a single dashboard using a trick that Joe Mako and Ty Aliversos have blogged about. Here’s Ty’s blog post. There’s a recent thread on the forum that explains this too but I can’t find it right now.
  3. It’s not clear what the bars in the “Year of sale” represent. I assume it’s number of sales, but there’s no tooltip or other indication of what the length shows. Using a bar chart as a filter is a great use of screen space, much more efficient than a filter and I blogged about that a while go, too.
  4. There’s no Y-axis label. I guess that’s because you can’t dynamically control the axis text. However, if you put the parameter onto the column shelf next to the measure, and then rotate it so it’s vertical, you get the same effect. I did this in my gapminder post.
  5. In order to make the story more clear, you could annotate a couple of indicative marks on the scatterplot that highlight the issue. You’re tight for space on this dashboard, so it may not be possible.
  6. Your label below the scatterplot says “Right-click….” That’s what you do in Desktop, but not in Public. You should change it to say “Click on a mark to explore”. Actually, what you really need to say is “Click on a mark, wait, then click on it again, and then you’ll hopefully get a hyperlink you can click on.” 🙂
  7. Finally, there’s nothing you can do it about it, but, boy, Tableau Public is slow (at 3pm on a Friday in the UK)

Good luck to Jerome and all the other entrants!

Next Top Model

The power of social networking has struck again. Next are seeking their next Top Model. Computer Science graduate Roland Bunce is the runaway favourite so far. I took the data of the top five entrants and created this viz to emphasize how big a lead he’s got. Go Roland! More details at the Guardian.

Commuting in the rain?


Three years ago I had a debate with someone about why they didn’t cycle to work in Oxford. “Because it’s too wet,” was their main excuse. Being a keen cyclist, I knew that it didn’t really rain that much, and even on days it rained, it was rarely at the time one commutes to work. The other person didn’t believe me: I needed evidence. I setup a Google spreadsheet and a form, and have been recording whether it rains when I commute to work every since. I now have 3 years of data, and as of today, 20 May 2011, I am in the midst of the longest ever dry spell: 82 consecutive commutes without rain. It has not rained on a commute since 28 February 2011. Below are two displays. The first shows the dry spells:

The second viz shows the summary. Headline figure: 9% of commutes are wet. That’s good odds in my book:

Many thanks to Richard Leeke of Equinox who worked out how to do the calculated fields that work out the running sum of consecutive dry commutes.

Terrific tooltip tricks: European conference session

Thanks for visiting the blog – if you are visiting here as a result of my Tableau European Customer Conference session, welcome along. The best way to keep up to date with new posts is via my twitter feed (@acotgreave).

My session was all about tooltips, and ways you can use them to enhance your visualisation and make your audience’s lives a lot easier. I have posts on various tooltip topics:

If tooltips aren’t your thing, then have a browse of the blog for other Tableau tips and tricks.

At the time of writing, the conference workbook I used isn’t yet available for download. Follow me on twitter where I will update everyone when it is put onto the Tableau Conference website. There will also be a video recording of the session should you wish to revisit some of the specific content.