Data Viz Done Right

October 24, 2016

Makeover Monday: How big is America's debt?

This week's Makeover Monday looks at this infographic from Visual Capitalist:

What works well?

  • The author is at least making an attempt, though a poor one, at putting the US debt into context.
  • Overall, the infographic is visually pleasing.

What doesn't work well?

  • The author uses green for the US in the pie chart, but black everywhere else. This should be consistent as it could lead to confusing the two.
  • The pie chart is 3D and appears to have an extra little white slice that doesn't mean anything.
  • All of the comparisons except the S&P 500 seem to be a real stretch.

When I created the data set for this week, I included only two records. I did this because I wanted to challenge people to work with small data. I'll be writing later this week (hopefully) about how people actually handled it, many not very well.

My initial idea was to create a unit chart that included 1 million dots, but Tableau wouldn't draw them. I then went back to Alteryx and reduced it to 10,000 records. Here's my workflow:

I then created two food themed vizzes. First, I created this candy dots chart. If you don't know what candy dots are, click here.

I think this shows the context of the US vs. the rest of the world well, but I don't love it. Ok, how about a waffle chart:

I like how this is really long, but when I showed it to my son Oscar on the plane, he told me it wasn't very good. Nothing like being told the harsh truth by a 14-year old. 

I had food on my mind, and most importantly, SIMPLICITY! I was making this too complicated. Back to the original data of just two records I went. My only goal was to communicate very clear, very simple message. 

Yes, I know donut charts aren't "best practice". I like them, though, when I only have two segments and I can use the hole to communicate the message.

October 21, 2016

Fix It Friday: Ten Alternatives Methods for Presenting Alcohol Consumption in OECD Countries

If this post turns into a bit of rant, bear with me. Let's start with the Tweet that got me worked up:

You might think "It's just a chart Andy, relax!" True. It's a chart. It's not changing the world or anything. There are several things that have me a bit upset:

  1. Paul Kirby calls the chart "interesting" and maybe the CONTENT is interesting, but the chart is terrible.
  2. He says "Austrians drink twice as much as Italians", a fact that is simply not true. They drink 61% more than Italians. You can't just spout facts like that.
  3. Paul is visiting professor at the London School of Economics. I can only assume that his students follow him on Twitter. When he tweets things like this, his student will assume that this is how charts should be made, which only proliferates the number of poor charts we'll continue to see.

The chart itself has its own set of problems:

  1. It's too dark overall. The dark red bars and dark bottles are hard to see against the blue background.
  2. The flags are unnecessary. What value do they add?
  3. The bottles are cute, but unnecessary decoration.
  4. The legend is in reverse order.
  5. Do the bottle extend beyond the bars or do they start from the same baseline?
  6. It has a weak title. What's the story?

This is chart junk at its best. Don't create charts like this. I went to the OECD website and downloaded the data. Below I present ten alternative charts that all work better than the original. You can download the Tableau workbook with all of these charts here.

October 19, 2016

Tableau Tip Tuesday: How to Create a Diverging Bar Chart with One Measure

No comments
This week's tip came about via a question from Bethany Fox of the Data School during our client project. I had previously posted a tip about how to create a diverging bar chart with two measures, but she wanted to create a diverging bar chart based on only one measure.

With a cheeky use of the INDEX table calculation, this was quite straightforward. In the video below, you'll see that the middle of the charts aren't lined up. I fixed this by using the INDEX calc again. You can download the workbook to see how I got it to work.

October 17, 2016

Makeover Monday: A State by State Look at Trump vs. Clinton

No comments
I had an idea as I headed into work this morning for another view for the election data for Makeover Monday week 42. I wanted to look at the latest forecast (i.e., Oct 12) and look at the discrepancy by State. I also really enjoyed this because it went from idea to created in about 15 minutes.

I think this view helps show much better than my last version that gap between Clinton and Trump. I also included a sorting option so you can look at it from different perspectives.

October 16, 2016

Makeover Monday - Trump vs. Clinton: A Race for the Presidency

No comments
This week for Makeover Monday we look at election forecasting data by Drew Linzer and visualised both on his website ( and on Daily Kos Elections. I first met Drew back in 2012 when I asked him to come speak at Facebook. When I tell people who Drew is, I say that he does the same kind of work as Nate Silver, except he's transparent about how his models work.

Given that we're nearing the end of the election cycle (thank god!), we thought it would be a good time to see how the races are stacking up. First, let's look at the viz from Daily Kos:

What works well?

  • The interactivity is amazing!
  • Nice summary on the left
  • The dots for the polls add nice context
  • Simple and easy to understand
  • Great overall design
  • Great use of color

What could be improved?

  • I wish the most recent results would stay on the line chart as I hover over another date. Yes, I know they are on the left, but then my eyes have to move back and forth.

For my version, I wanted to learn how to create small multiple tile maps, so I went straight to Matt Chamber's blog. I added a couple bells and whistles to it too:

  • I added a reference line on each state at 50% to help show if one of the candidates has more than half the vote.
  • I included bar charts in the tooltips.

I really like the line chart by Daily Kos, so I rebuilt that as well. I decided to use a parameter for alternative date because this allowed me to address the issue of not seeing the latest forecast as well. I color coded the lines based on whether they are above or below the date picked by the user.

Once again, I've learned a ton working on Makeover Monday. Click on the image for the interactive version.

October 14, 2016

Join us at #RunData16!

No comments

For the past few Tableau Conferences, members of the Tableau Community have come together for some early morning running. Last year, as you can see above, we had an incredible turnout and we expect this year to be much more of the same.


  • Distances: 5K and 10K (easy enough to make longer or shorter if you'd like)

Yes, a 5:30am start time is early, but in our experience, you HAVE TO start this early if you want time to make the keynotes. Running people are a weird bunch anyway, so 5:30 is never too early for us!

The run leaders will have high visibility bibs and torches for everyone's safety. If you have lights, bring them along. I've run these routes many times and it's a beautiful trail along the river. There are always tons of runners out and about.

NOTE: You may see a meetup list on the conference website, but it doesn't start until 6:30. If you go to that one, you have very, very little chance of making the keynotes. Plus, most of the runners will be at our run. And we'll be done before they even start!

Tableau has informed us that they will not be supporting us this year. However, keep up with the #RunData16 hashtag on Twitter for all of the latest information. See you in Austin!

October 11, 2016

Tableau Tip Tuesday: How to Limit the Number of Marks in a View

In this week’s tip, I show you how to limit the number of marks that are displayed in a visualisation. The example I show will help prevent your users from creating line charts with too many different lines.

October 9, 2016

Makeover Monday: How satisfied are people with public transportation in some of Europe's biggest cities?

No comments
This week for #MakeoverMonday, we look at this simple stacked bar chart of public transportation satisfaction survey results from the Financial Times.

I first saw this survey in print at Gatwick airport on my way to Prague, then it appeared in feedly. I know from speaking to John Burn-Murdoch that the print and online graphics standards are different. The print version I actually found easier to understand because it used blue for negative sentiment.

What works well?

  • Clear sorting by very satisfied
  • Sticks to their color guidelines
  • Simple title

What could be done differently?
  • Use different colours for the negative and positive sentiment
  • Add an overall score (like net promoter score)
  • Include 2012 for comparison so that you can see which of these cities improved
  • Add a more descriptive title so it's even more clear what the audience is looking at

I used a few resources to help me create my final visualisation:

First, I recreated the FT viz, but with different colors for negative and positive sentiment. I also included bar charts in the tooltips.

Next, I included 2012 and labeled the bars where they fit.

I don't particularly like the labels on the bars, so I've removed them from the final version. I also changed the bars to a Likert scale, which moves the negative to the left and positive to the right, and helps shows the discrepancy better. I also included the net promoter score.

Last, I added a slope graph to help show the change and included a more descriptive title and subtitle. You can click on any bar and it'll highlight in both places.

October 7, 2016

Tracking Hurricane Matthew

No comments
It's back to Data School week for DS2 and today we were running through the new features of Tableau 10. I was demonstrating the Google Sheets connector and given the terrible storm that is currently battering the Southeast US, we created a connection to the storm path data from Weather Underground and built this storm tracker. We then also looked at Device Designer, which means there a nice mobile version of the viz too.

What's even better is that Tableau Public will refresh this data daily, so we can look back and see how the storm has progressed.

October 6, 2016

Progress for Sci-fi Reviews by Women

Last night I attended the London Data+Women meetup and Emma Cosh gave a fantastic presentation about a project she worked on with Strange Horizons about the change in women reviewers and reviewers in sci-fi magazines. The visualisations she presented worked well because she was there to explain them. On the way home, I got to thinking...would these work well without her explaining what they mean? Maybe, but I think they could be better.

This also got me thinking about the work by Stephanie Evergreen and her focus on effective, impactful titles. I recommend to people that when they are creating visualisations, assume the audience will see a static image. If they can't understand it, then it should be changed.

First, here's the visualisation that Emma created:

Click for the interactive version

There are a few things I would change:

  1. Give it a stronger title that explains the visualisation and the key message
  2. Only label the lines that are increasing
  3. Only show the magazine name on the left label to minimize the text
  4. Make the footer legible (brown on black is too hard to read)

The workbook wasn't downloadable so I recreated all of the data manually. I made the changes noted above, which are all pretty minor. Most importantly, though, the viz now has a stronger title and gives the reader a much clearer message. As Stephanie Evergreen says:

If you do nothing else to improve a weak visualization, you’ll still seriously improve its interpretability by giving it an awesome title.
I certainly wouldn't classify Emma's viz as weak, it merely could be more effective.

October 4, 2016

Tableau Tip Tuesday: How to Label the Top N Points on a Line Chart

For this week’s tip, I go back to an old tip from 2001 and demonstrate how to use a parameter and a rank calculation to display the top N points on a line chart. It’s pretty straight forward and doesn’t require anything complicated. This method is definitely simpler than the original post, which used an INDEX table calculation.

Motor Vehicle Occupant Death Rates in the USA

I was reading through feedly this morning and saw this great viz by The Economist.

I really like this simplicity of the viz, yet the detail and insight it provides. In particular, I like the Gantt chart style they used to compare 2000 to 2013. One of the best way to learn is to recreate charts you find and like.

For my version, I used data from the CDC about motor vehicle deaths by state in the US. Overall I went with a similar Gantt bar style to compare the change in the years. I made these additional enhancements:

  • Removed the line that makes them look like candlesticks
  • Muted the gridlines
  • Moved the labels next to the bars
  • Colour-coded the bars to show whether each state has increased of decreased
  • Moved the United States average to the top to make it easier to compare to

Which version do you prefer? What else would you do differently? You can click on the image below to download the workbook from Tableau Public.

October 3, 2016

Makeover Monday: The World is Becoming Less Peaceful

Last week, the Global Peach Index, this week we fix the typo and look at the Global Peace Index from Vision of Humanity.

What works well?
  • Nice interactivity and tooltips
  • Good filtering capabilities that show additional information
  • Good slider implementation to scroll through the years
  • Dark red and dark green countries are very distinct, drawing my eye to the worst and best

What doesn’t work well?
  • It’s impossible to get a sense of any trends over the years
  • Using a filled map makes it very difficult to see smaller countries
  • It’s hard to get a feel for the overall peacefulness of the world, i.e., what’s the global average?
  • Color palette is hard for color-blind people and doesn’t supply enough range in colors

For my version, I used the text of the summary below their visualisation to help craft my story. Throughout this year, I’ve primarily been creating long vizzes, but I really like the many examples Andy C has created that are wide, so I thought I’d give that a go this week.

You can click on the image for the “live" version, though there’s not really any interactivity. The mobile version will be long though, as I find that makes for a better user experience.

September 26, 2016

Makeover Monday: China is Dominating the Global Peach Index

No comments
This week's Makeover Monday was inspired by a typo in an email exchange with Andy and me. I had thought about doing the Global Peace Index this week, but accidentally typed "Peach" instead of Peace. Andy pointed it out to me, but then I thought, I wonder if there is a viz and data about global peach production.

And thanks to FAOSTAT there is! Who knew?!? Their data set is accompanied by a series of chart. I'm going to focus on their map.

What works well?

  • It's a map, so I can easily understand that it's show geographic distribution.
  • Nice filtering capabilities

What doesn't work well?
  • The color scales don't make sense. Are they ranges? Are they precise values?
  • There are a lot of yellow countries/ What does that mean? 
  • The blue water makes it hard for the blue shading on the map to stand out.
  • The mapp wraps and repeats.
  • Comparing countries on a filled map is nearly impossible. How does China compare to Holland? If you can't answer questions like that, then a filled map is not the answer.

For my version, I wanted to focus on the top peach producers in 2012, so I created a set that only includes to top 10 countries of 2012. I started with a summary, then an view over time, followed by a heatmap to help highlight the differences.

Click on the image for the interactive version.

September 19, 2016

Makeover Monday: Data breaches are getting bigger and more frequent


Several people have recommended Makeover Monday for the Project of the Year in the Kantar Information is Beautiful Awards, which I must admit is quite stunning and flattering at the same time. The suggestion for this week’s makeover came from Andy Cotgreave. We intentionally picked something from Information is Beautiful with the hope that it gets a bit more exposure. Shameless perhaps, but what can it hurt? This viz from David McCandless certainly deserves a makeover.

What works well?
  • The viz is eye-catching and definitely draws you in. There’s something to say for that.
  • The interactivity is fantastic.
  • Good filtering, colouring and sizing options

What doesn’t work well?
  • The bubbles move all around for no apparent reason.
  • There’s way too much overlapping, making it hard to identify any insights.
  • Whether something is interesting is extremely subjective. I wouldn’t make these same choices.
  • The viz doesn’t fit in a single view, requiring too much scrolling.
  • Not all records are included. I guess this was done for artistic purposes as David is known to do, but it distorts the message.

I decided to work on my makeover during my flight to Prague, thus imposing a time limit on me. I started by creating a view that simply shows the number of data breaches by year using circles. This basically flattens out the original.

While this shows the distribution nicely, I don’t love it. Next, I converted the circles to squares, hoping the result would be more visually impactful as the squares take up more space.

This is definitely better, however I don’t like how it doesn’t incorporate the records stolen in each data breach well enough for my liking. So I decided to add a dot for every breach in the data set and change the location of each dot to the number of records stolen.

Getting there…iterating is really helpful. This shows some of the outliers really well, but I feel like I’ve lost the distribution a bit. I decided to quickly open the data in Vizable and when I switch the view to records stolen by year, Vizable presented me this interesting view that shows the median and the distribution.

I really liked this so I decided to build upon it in Tableau. My final viz incorporates the view from Vizable, the distribution of each data breach and allows me to focus the story on data breaches that were hacks versus not hacks.

Click to view interactive version

I find this final view much, much easier to look at than the original and also it provides much better context. For me, context is key. Every visualisation you create should include context somehow. Why? Context makes it much easier for your audience to understand the story.

September 13, 2016

Tableau Tip Tuesday: How to Create a Combination Chart with Overlapping Bars & a Line


In this week’s tip, I look back at one of my most popular posts - 7 easy steps to create a combination chart with overlapping bars & a line. The tip hasn’t changed much, however, this time there’s a video.

September 12, 2016

Makeover Monday: Which shipping company really makes the most money?

This week’s Makeover Monday looks at the largest shipping container companies of 2016. The article includes this packed bubble chart:

What works well?

  • It’s eye-catching and draws you in.
  • The method for labelling ensures you only see the largest (as they names won’t fit otherwise).
  • The color helps identify the largest companies.

What doesn’t work?

  • Ranking is nearly impossible
  • Are the depth of color and the size of the bubble for the same metric? The viz doesn’t tell us, so we’re left to guess.
  • Is big good or bad?
  • There’s no focus or context.

I started by changing it to a bar chart, but found that to be too boring, though effective. Then I saw an example by Shawn Levin which shows looks at the TEU per ship. That adds much more context to the visualisation. Shawn compared Total TEU to TEU per ship.

For me, I thought it was more meaningful to look at TEU per ship for both the total shipments and for the ships each company owns. This led me to the slope graph you see below which tells a much more meaningful story.

September 6, 2016

Tableau Tip Tuesday: Using a Level of Detail Expression to Summarize Dimensions

This week’s tip goes back to a tip I wrote in 2012 that required table calculations. With the addition of Level of Detail expressions in Tableau 9, the need for a table calc is obsolete.

September 4, 2016

Makeover Monday: Alan Rickman - An Actor’s Life


This week’s Makeover Monday subject was sent to me by incoming Data Schooler Anna Noble. Nothing like kissing up to the coach before you even start. 😋

The viz in question is from The Slow Journalism Company:

What works well?

  • It’s obviously a timeline.
  • It’s clearly about Alan Rickman.

What doesn’t work?

  • It took me a while to figure out what the being on each side of the timeline meant.
  • Horrific colour choices; apparently the colours identify the genre. You can see that in the microscopic font on the lower left.
  • Colours with zig zagged lines are always a bad choice
  • The pointy things
  • The pie chart
  • The annotations aren’t near the data they represent.
  • You have to use the zoom feature to read anything.
  • It makes me dizzy.

For my makeover, I wanted to do something very simple. This data set calls out for something simple and clear, especially after you spend time trying to understand the original. I have no idea who Alan Rickman is as I’ve never read any Harry Potter books nor seen the movies. Yes, I’m a terrible father! I once again used 100% floating objects to create this. I’m really loving the precise control that lets me have.

Click on the image for the interactive version (but really there’s no need as there nothing more to the live version other than a mobile view).

September 2, 2016

The Toxic Twenty Five: An Analysis of Southern California Air Quality

This week I challenged The Data Duo to a #VizOff of sorts. I provided them with a data set of 8.5M ozone level readings from stations spread all throughout the U.S. I started looking at this data a few weeks ago because I was thinking about the smog in Atlanta and wondering if it had gotten any better since I left. This led me to the master data set or all cities that are measured.

Once I started exploring the data, I noticed that Southern California consistently had the most cities with high ozone levels. So I filtered the data set down to the 25 worst cities.

This helped me focus on a single story with multiple parts, as seen in the long-form visualisation below. Enjoy!