Sports Viz Sunday

I have been a fan of sport since I was a little boy standing on a milk crate on the terraces at Portman Road (when I was a bit older a family friend made me a wooden box he had painted blue and white to stand on – writing this makes me feel very old).

So when I saw that Spencer Baucke, Simon Beaumont and James Smith had created an initiative called SportsVizSunday I was keen to throw my hat in the ring.  As well as encouraging people to share sporting vizzes on and twitter (using hashtag #SportsVizSunday), the team also supply a dataset each month with a sporting theme.

In this post I will showcase my contributions to the project, with a design or technical tip thrown in for good measure.  For a perspective on why sport is such a good topic for visualisations, I highly recommend this blog post by Neil Richards.

March – Formula One

To be honest, I find Formula One intensely dull (unfortunately it is the only sport my partner watches), but Simon gave me a nudge at the Midlands Tableau User Group so I thought I would give it a go!

My viz focuses on individual race winners, highlighting a select band of drivers who have dominated different eras:

Formula One Grand Prix Winners

Viz tip – I wanted to provide a way of users to quickly highlight winning drivers that weren’t in this select band.  Here are the steps I took to do this (I have used variations of this in a number of vizzes and it is a great way of making your viz configurable).

  1. Created a parameter ‘Select driver’ using the field containing winning drivers (allowable values = list, add from field)
  2. Created a calculated field ‘Selected driver’ with a value of 1 for the driver selected using the parameter dropdown (IF [full name] = [Select driver] THEN 1 ELSE 0 END)
  3. Included the value for ‘Selected driver’ in the calculated fields for colour and size, so the selected driver is represented by a large, orange dot.



April – Masters Golf

Golf is another sport I don’t know much about, but I had seen Rob Radburn’s viz and had been intrigued by just how many tournaments some players had been involved in (50 or more for Gary Player and Arnold Palmer).

I hadn’t known that previous winners are invited back to play every year, so my viz focuses on this aspect and on how their performance deteriorates over the years (Jack Nicklaus is an interesting outlier here).

Masters Golf

Viz tip – I took great care over the placement of the player names in this viz.  Here is the default view, with the player name as a row header:

Masters 1.png

It looks alright for the early years of the tournament, but the amount of space between the name and the marks for the more recent winners make them hard to identify.  Here is how it looks as a left-aligned label:

Masters 2

The labels are nearer the values, but they are all over the place as my data is ordered by the year of the first win, not the first tournament, for each player.  So in my final viz I adjusted this so that no label is further to the right than any one above it (I did this by manually adding a column to the dataset, although it may be possible using a calculated field).

A small decision, but I think it makes a big difference.


April – World Championship Snooker

My final viz is on snooker, one of my favourite spectator sports.  For British sport fans of a certain age the 1985 World Snooker final was an iconic moment (I remember being allowed up past midnight to watch the end of the match, which was decided by the final pot of the game after almost 15 hours of snooker).

Neil Richards created a trademark radial viz on the same game, which you can see here.

1985 The Black Ball Final

Viz tip – How to visualise each shot in the final frame was an interesting challenge.  I knew I wanted the balls potted to be represented as intuitively as possible using coloured circles:


However I wanted to make it more obvious where a player had made a break.  The answer?  A dual axis chart, with a (very thin) Gantt chart representing break duration.  Note that where Davis fouled (black diamond) then was put back in by Taylor and played a safety (light grey diamond), I purposely did not connect the two shots:

Balls with line

And the snooker table plots?  I could have created these as images, but to allow me to alter the ball positions I created these as scatter plots, with the snooker table as a background image.


Keep it clean

One of the things I like most about Tableau is that it allows you to keep your dashboards clean and uncluttered while adding detail on demand just below the surface.

In this blog post I will demonstrate four methods I have used to achieve this recently, with links to posts that walk you through each process.

You can also download and investigate any of the dashboards featured from my Tableau Public profile.

Method 1 – Viz in Tooltips

Viz in Tooltips was introduced with Tableau 10.5 and allows you to insert sheets into your tooltips as well as the traditional dimensions, measures and parameters.

For me, this is a real game changer in facilitating the addition of details on demand.

Jeffrey Shaffer has written two excellent posts on how to create Viz in Tooltips which include several potential case uses:

10 tips for Viz in Tooltips | Another 10 tips for Viz in Tooltips

I used the technique recently in my viz on Female Nobel Laureates for the Midlands Tableau User Group.

By importing photos of each female laureate as a custom shape I was able to create a profile that appears when you hover over the relevant mark:

Viz in Tooltips

Method 2 – Filtering

In the same visualisation I wanted to include a section showing profiles of some of the women who have been overlooked for the Nobel prize.

With dashboard space limited, my solution was to use shapes as a filter to allow the user to cycle through the different options.

For a clear walk-through of the process, see Dash Davidson’s post How to use custom shapes as filters in your dashboard


Method 3 – Sheet Swapping

Sheet swapping produces similar results through a very different method. The trick here is to create your sheets then add them to a single container, with a parameter to allow the user to choose which is visible.

To create this visualisation I followed the steps in this blog post by Hashu Shenkar:

Tableau tip: Switch between views dynamically on a dashboard

Sheet swapping1

Method 4 – Background Images

The method I used in my St.David’s Day viz was the most labour intensive – although I did have 548 images to add manually!

The section on the left is a separate sheet with each image loaded as a background image. An image is only shown when the unique ID is selected by hovering over a dot on the dragon.

For an overview of this method (using a more sensible number of images), see Shawn Wallwork’s post QT: Dynamically Switch Images Using Filter.

Happy St David's Day

Basic bars, intentional design

“Good design takes planning and thought”. Cole Knaflic, Storytelling With Data, p139

The tweet

Naomi tweeted me yesterday with a question about my Makeover Monday viz on policymaker estimates of gender equality measures:

Naomi tweet

You can see the viz – and judge for yourself – here, but I wanted to blog about four intentional design choices I made creating a much simpler viz: a basic bar chart for this month’s #SWDChallenge.

Choice 1: Labels within bars

Screenprint 1

Now Tableau is definitely my Data Visualisation tool of choice, but it does have its quirks. One example is labels on bars, where the default options are as follows:

Label options

Personally, I like to have labels right aligned to provide an additional positional comparison, but with them tucked just inside the bars (don’t ask me why – it just seems…neater).

This is possible within Tableau, but involves a workaround. I created a second bar chart using a calculated field SUM([Total])-[Label offset], where [Label offset] is a parameter control I could adjust until the labels fell just inside the bars when the axes are synchronised.

I would have been better off googling how to do it, as Andy Kriebel had already detailed two superior methods in this blog post, including one which keeps your secondary axis free.

Choice 2: Colour

Screenprint 2

I knew I wanted separate colours for swimmers and gymnasts, with other athletes a neutral colour (I had some qualms about using grey in case this was misinterpreted as denoting silver medals, but I considered the risk of this to be small in the context of the overall design).

Often I will chose my own colours by eye, but this time I decide to use Coolors, an application which generates custom colour palettes. One of the first colours suggested was Aquamarine, which seemed perfect for the swimmers. By locking this colour and trying more combinations I found a complementary yellow for the gymnasts.

Using the colours in the takeaway title helps to strengthen the association, and removes the need for a separate key.

Choice 3: Removing duplication

Screenprint 3

In my first draft I had labels for every bar, but this resulted in a significant amount of duplication (there are 3 athletes with 13 medals, 8 with 12 medals, and 5 with 11). I felt the viz would look cleaner by just showing a label for the first instance of each total.

If you are looking for a particular athlete then this may not be optimal, as you may need to move up a few rows to see the value. Having the repeated labels in light grey to de-emphasize, rather than remove, these values is another option.

Choice 4: Additional data layer

Olympic Medal Count

Interactive viz

Now here is where Tableau really comes into its own. Viz in Tooltip was recently introduced as a new feature in the 10.5 release, and I am a huge fan of how it allows you to keep a viz super clean, with details available on demand.

I decided to use Viz in Tooltip to show an additional data layer containing each athlete’s country, sport (information not shown for the grey bars) and gold, silver and bronze medal count by Olympic year.

Final thoughts

People will make different design choices when creating vizzes, but every element of your design should be intentional, even (in fact, especially) if your choice is to retain elements of the default view.

Your chart type may be basic, but your design choices should always be thoughtful.

Ladies and Gentlemen…Album Covers as Data Viz

This week’s Makeover Monday was another one of those I tend to struggle with – just one measure (export value in USD for drugs and medicine) and two dimensions (country and year, with full data for only four years).

As usual I started with a quick exploration of the data, looking at total exports by year then trends by country. Yet it felt like I only had half the picture (exports but not imports). I was surprised to see Germany as the largest exporter, expecting it to be the U.S.  Perhaps U.S. production is mostly for the domestic market?

There seemed like two options: to do what Mike Cisneros describes as looking for the missing dogs (a sankey showing exports and imports?) or to embrace the limitations and think creatively…

Album Covers as Data Viz

Creative up to a point, that is. My idea was to create a homage to Neil Richards’ series of albums as Data Viz, themselves homages to iconic album designs. (You can see examples by Neil here and here, or at the twitter hashtag #AlbumsCoversAsDataViz).

Given the subject matter, Ladies and Gentlemen We Are Floating in Space by Spiritualized seemed like a perfect choice (as well as being one of my favourite albums). The artwork for the album was designed by Mark Farrow to resemble prescription medicine, with the CD contained in a foil blister pack and credits resembling medical information.

The circle design lent itself to showing total drugs and medicine export value by year (there is one fewer circle, but it does mean the most recent year to the right of the chart is more prominent). The chart itself is just a circle for each year with a white border and 0% opacity, sized by total export value.


I found an article online that suggested the font used was Helvetica, used HTML Color Codes to get the right shade of blue, and chose text inspired by the original artwork which related to the subject. By importing the original album as an image to use as a template, I was able to make the final design pleasingly close to the original:

The finished image is lacking in insight, of course. However the beauty of viz in tooltips is that I could add detail to the interactive Tableau public version while keeping the design as minimal and clean as the original. So ladies and gentlemen, here is the final version, complete with tooltip for each year showing the total export value and top 10 countries:

Viz in tooltip


Makeover Monday: a different approach

There have been a few changes with Makeover Monday this year, most notably with the move from posting visualisations on rather than Twitter. In terms of tracking submissions this makes perfect sense, although I do think that interactions with other participants would be encouraged by enabling thread view in the discussions.

I have also changed my own approach slightly, and it seems to be working, with both visualisations I have submitted this year being included as a favourite in the weekly round-ups. So what am I doing differently?

It comes back to one of my Q1 aims: to find new angles in data.

Week 4 – Turkey Vultures

Tableau Public viz: How Far Do Turkey Vultures Fly?Turkey VulturesFor a real lesson on the benefits of finding an angle in your data, and using storytelling techniques to share your findings, check out this wonderful viz by Matt Francis.

My own angle was to try and create something for children. I am not sure why this came to mind as the original data is from a fairly dry academic paper. The fact that each vulture in the study was given a name was definitely a factor, and having a primary school teacher for a partner was definitely another.

(I also wasn’t the only person with this idea, I really liked Staticum’s engaging viz They Call Me Domingo).

When I found the delightful Turkey Vulture icons from Birdorable (and was given permission to use them), I knew the idea had potential. It also gave me the opportunity to experiment with mapbox, which I highly recommend as an easy way of adding to the Tableau default maps.

I even created a gamified version to share at my local Code Club, where I volunteer helping 8-12 year olds learn to code (you can play the game at Morongo the Turkey Vulture).

Week 5 – What the Most Profitable Countries Make per Second

What Apple makes in one minute

Week 5’s dataset was one of those I normally struggle with (limited data points, and a subject I am not particularly interested in). I had two main thoughts after my first look at the data: just how far ahead of the other companies Apple where, and just how enormous the company profits are.

A simple bar chart showing all 25 companies in the dataset would have worked well, but just seemed a bit…easy. I also wanted to highlight the sheer size of the profits, which comparing each value against the other doesn’t really do.

My angle was to focus on the time element. By converting the data to profit per minute, and showing a minute pass in the visualisation, the enormity of the profit is underlined. Including selected other companies (concentrating on the most relevant comparisons) allows the viewer to see how far ahead Apple are, as well as adding some visual variety.

Now I love Tableau but I know that the pages function doesn’t work in Public, so for animated visualisations it is not my tool of choice. I don’t currently know D3 (I decided learning Python would be more useful at the moment), although I would love to recreate using D3 at some point.

So I went old school, using PowerPoint to create the initial image then creating 60 new frames, and coverting them to a gif using ScreenToGif (I also created an mp4 version for linkedin using Wondershare Filmora). If you are wondering about the second hand, the trick is to create a line (half black fill, half transparent), then rotate 6° at a time.

And if you are wondering how I found my angle, I divided 360° by 60.


A short note about annotation

One of my aims for this year is to think more like a Data Journalist: to source my own data, find new angles on existing stories, and use annotation to provide context.

So I was delighted to get the following feedback from Mike Cisneros for my most recent visualisation: When Did Happy Days Really Jump the Shark?

MIke Cisneros tweet

If you are not aware of Mike’s work (or blog) you are really missing out – aside from his technical wizardry he is a master of editorial thinking. Check out his recent #MakeoverMonday viz Deconstructing the SCMI for example:


If you are still not convinced of the importance of annotation, here is the same visualisation stripped of the written context. As the title says, Mike is deconstructing the SCMI. But it is the annotation, not the visualisation, that is doing the heavy lifting.


My own visualisation tries to do something similar: the text at the top explain the phrase in the title, argues why this did not occur in the episode in question, and proposes an alternative.

Mark labels highlight the two key episodes, and further detail in the tooltips is signposted using a direct label  Hover

All this is done as concisely as possible in line with Jorge Camoes’ advice that annotation should be “useful and accurate and should not compete for attention…a discreet and helpful whisper.”1

When Did Happy Days Really Jump the Shark?

Happy Days

Yet the importance of annotation in Data Visualisation is often overlooked. Andy Kirk describes it as “the most neglected layer of the visualisation anatomy”and with the exception of his own book it is barely covered in the literature.

Nevertheless neglecting it does the viewer a disservice – annotation acts as a bridge between them and your work (“the interface between data and communication”3), and a failure to consider it may leave the viewer stranded.



Jorge Camoes, Data At Work p339

Andy Kirk, Data Visualisation p247

Elijah Meeks, Making Annotations First-Class Citizens in Data Visualization


My #VizGoals for Q1 2018

I have been thinking about my #VizGoals for 2018 in the last few days and decided to write a short blog about them, partly to hold myself to account and partly to make a start on number 3 🙂

So here they are…I have limited myself to five goals I feel I can achieve in the first three months of the year:

  • Don’t look for a job

That’s right, don’t look for a job.  I have been on a career break since last November and have found the additional development time to be incredibly rewarding.  It also allows more time for volunteer work with Code Club and The Cinnamon Trust.

I feel privileged to be able to do this, and want to use the opportunity more wisely than my last career break in 2007 (when I spent two weeks trying – and failing – to get to the Solovetsky Islands, then took an identical job at a different organisation).

  • Think like a Data Journalist

What do I mean by this?  Every now and again I see a visualisation on Twitter that I immediately want to open on my desktop and explore, or one that gives me a completely fresh angle on something I thought I knew.

This time last year I asked Andy Kirk for advice on developing a portfolio, and his advice was to think like a Data Journalist.  By doing this – searching for data, finding new angles, researching and adding context – I aim to create some of these myself.

  • Blog frequently

I plan to blog more frequently, and not get hung up on whether I feel I have anything to add to other blogs that are out there.  Hopefully the discipline of blogging will be its own reward, and developing my voice will mean some worthwhile insights emerge.

  • Broaden my learning

I would like to broaden my learning as well as deepen my Tableau knowledge, so possibly D3, maybe R, but most likely Python (which may also help with Code Club).  I bought a discounted annual subscription to DataCamp in December, so it is time to start using it.

  • Become a Tableau Desktop Certified Professional

In December I became a Qualified Associate, so my next step is Certified Professional.  If anyone has any advice on how to successfully prepare, please let me know in the comments!

My 2017 #DataMemories

I have enjoyed reading everyone’s #DataMemories in the last few days, so here is my tuppenceworth – five data highlights from 2017, and one lowlight!

The highlights

  • Promoting Data Visualisation within Lloyds Banking Group

One of my two main achievements in my most recent role was the Data Visualisation page I established on the Group’s internal collaboration site (Jive), which brought together all the expertise across the organisation into one place.

As well as creating engaging content including monthly Data Viz digests, a series on unusual graphs, and webinars on how to create effective infographics, I helped to establish the annual Viz of the Year competition and enable best practices to be shared across the Group.

When I left in November (more of which later), the page was in the top 50 most followed sites out of almost 7,000 and in the right hands to make it even more successful.

  • Embedding Tableau within my team, and creating new advocates

My other big achievement was to finally secure access to Tableau Desktop for my team (three years after my first attempt, albeit in a different role).

Through a mix of one-to-one coaching, team self-learning and regular Show and Tells (with Tableau swag as prizes) I was able to upskill a team of 7 analysts and develop a core of Tableau enthusiasts within the department.

It wasn’t a total success – I was still battling for Server access when I left – but it was the first big step in the transformation from providing old school, semi-manual MI to self-serve reporting and meaningful insight.

  • Becoming more involved in the Tableau community

I made more of an effort to attend Tableau User Groups and engage on Twitter in 2017 and I continue to be inspired and educated by the wider community.

Of those I met in person at either the Midlands or London User Groups I would like to thank and give a big shout out to Elena Hristozova, Ali Motion, Neil Richards, Chris Love, Neil Davidson, Sarah Bartlett and Simon Beaumont, with apologies for those I may have missed.

I even left my comfort zone to present my National Student Survey dashboard at Leicester (fuzzy photographic evidence courtesy of Neil Richards) – for more of the dashboards shown see Elena’s post NSS 2017: One data set, many approaches.


Of those I have only engaged with on Twitter, I would particularly like to thank Andy Kriebel and Eva Murray for all their hard work sourcing datasets and providing critiques for Makeover Monday, which formed the basis of over half the vizzes (25/46) I added to my portfolio this year.

  • Putting more time aside for learning, and seeing the results

At the start of the year I reduced my hours so I could spend more time developing my Tableau skills and creating a portfolio.  Then in May my manager offered to pay me for the development time as there was a clear benefit to the business – big mistake!

My good intentions to ring fence my Fridays soon went out the window and the demands of the day job muscled out my development time.

In November I left my job to take a career break, and learnt more in the last two months of the year than in the rest put together – I personally feel that there is a clear increase in quality from the start to the end of the year, with more to come in 2018.

2017 – first 5 vizzes:

Made using TurboCollage from

2017 – last 7 vizzes:

Made using TurboCollage from

I even made Viz of the Day in November for this effort, which made the Tableau Public team’s top 40 Team Picks: Notable Vizzes of 2017:

30 Sustainable Public Transport

  • Becoming a Tableau Desktop Qualified Associate

In December I took – and passed – the Tableau Desktop Qualified Associate exam, luckily submitting my answers just minutes before my internet connection went down!

As someone who is self-taught I found it very useful to learn in a more structured way and fill in some gaps in my knowledge – I found Mark Edwards’ blog post on the exam experience (and the embedded Tiny Tableau Talk by Joanna Hemingway) an excellent resource to help me prepare.

The lowlight

  • When is 74.6% higher than 75%?

The lowlight was a lost afternoon justifying to a senior colleague why my team was reporting her department as Red on a KPI which came in at 74.6%, below the target of 75%.

Her argument was that if you rounded the number up it was the same as the target (and yes, we reported it to 1 decimal place, or with greater precision if just under target).

The silver lining was that this crystallized my perception of elements of the wider culture, and how numbers were used within the area I worked in.  I had been considering – and planning financially for – a career break for over a year, and this helped give me the little nudge required to change paths.

So an exciting 2018 is in store!  Keep an eye on the blog for my New Year #VizGoals (one of which – spoiler alert – is to start blogging again more often).