Data Viz Done Right

April 16, 2018

Makeover Monday: The Seasonality of Confirmed Malaria Cases in Zambia Southern Province

No comments
For Makeover Monday week 16, Eva and I are hosting a #MakeoverMonday Live at Tableau HQ in London. The data and viz this week were provided by Jonathan Drummey and the Visualize No Malaria project.

What works well?

  • The colors are distinct from each other.
  • The seasonality is very evident.
  • The title is simple and tells us what theviz is about.

What could be improved?

  • Are the colors stacked or is one behind the other?
  • The overall decline is harder to see than necessary.
  • What happened at the spikes? Adding some annotations would be helpful.
  • Why is the data split between health facilities and health workers?

My Goals

  • Can I show the overall decline more effectively?
  • What does the viz look like when I combine the health facilities and health workers?
  • Are there colors that will work more effectively?
  • How can I make the seasonality more evident?

With those goals in mind, here is my Makeover Monday week 16. If this looks somewhat familiar, I created a very similar viz with a very similar data set for Makeover Monday week 34 2016.

April 12, 2018

Workout Wednesday Part 2: Total Products by Sub-Category

No comments
Ok, part 2 of this week's challenge, the Jedi version, really sucked (requirements here). It took me FOREVER! I used table calcs for all of the calculations and getting them just right took a really long time and a lot of experimenting. Surely Tableau can make this easier for us.

Some thoughts:

  1. Getting the subcategories to layout correctly in a trellis plot was easy.
  2. Getting the labels above each grid was easy.
  3. Getting 10 dots across each pane was easy.
  4. Getting the stacking of the dots in rows was a pain!
  5. Luke has an evil side.

But I absolutely loved the challenge. I'm really enjoying these! My advice for everyone is to keep at it until you get it. Even if you're stuck, don't cheat and download the solution. That doesn't help your learning. If it's too hard, then consider skipping it; it might not be the right level for you yet. You can always come back and do it later.

April 11, 2018

Workout Wednesday Part 1: Top 5 Subcategories with the Most Products

No comments
Luke is back and he has posted two challenges this week. First, create a pictogram of the top 5 subcategories that have sold the most products. They should be represented in 100 dot sections and colored depending on the section they are in. Get the requirements here.

There were two requires that I didn't need to use to make it work:

  1. Set the minimum and maximum values on the columns axis (x-axis) to -3 and 12, respectively.
  2. Set the minimum and maximum values on the rows axis (y-axis) to -1 and 32, respectively.

I'm not sure what the purpose of these would be, but I suspect it's some sort of spacing. I didn't need them, so I ignored these requirements. Here's my version and now I'll get to work on part 2.

April 9, 2018

Makeover Monday: Arctic Sea Ice is Disappearing Fastest in Summer Months

No comments
I'm writing this having just finished a bike tour of Rome with my family in an absolute monsoon. Global warming is proven to cause unusual volatility in the weather, including hotter summers, extreme winter storms, and changing warm water patterns around the earth. This warming is most evident near the Arctic, where ice levels are at all time lows and the cycle of melting is accelerating year upon year.

So when I found this visualization by the National Snow & Ice Data Center, it seemed an appropriate topic for Makeover Monday. One of the most fun elements of this data set is that it includes only two columns: date and sea ice extent.

What works well?

  • Without even trying, it tells a compelling story.
  • The interactivity is fabulous. I really like being able to simply click on an item on the legend to have it added or removed as a highlighted line.
  • Including the 1981-2010 median along with the IQR and IDR provides great context.
  • Defaulting the view to show 2012 (the previously worst year for arctic ice) to 2018 helps show how 2018 is looking to surpass 2012 (in a bad way) by a lot.
  • Subtitle explains what sea ice extent means
  • Good use of simple colors
  • Great example of using highlighting for context

What could be improved?

  • The x-axis could be simpler by only showing the month names and removing the word "Date" from the axis title.
  • Make the title more impactful

My Goals

  • First, I wanted to rebuild the original and see if I could make it any better. I couldn't.
  • Second, build a spiral diagram that shows the months around the outside, but this only worked well when it was animated.
  • Finally, I settled on a different take on the metric that swaps the months and year on the original. That is, put the year on the x-axis and month on each line. This gave me only 12 lines which looked less busy and helped me see patterns for each month.
  • Next, I included a line that is the average of each year (black line).
  • I then decided to look at how each year of each month changed compared to 1979. I went with a percent change because I think that provides more context.
  • Lastly, I included a highlighter for the months and included some BANs of the actual values for comparison.

Click on the image for the interactive version.

April 4, 2018

Workout Wednesday: Frequency Matrix

No comments
This week Ann Jackson stepped in to provide everyone the challenge. I think she's been seeking some revenge and she sure dished it out. Find all of the requirements here.

Easiest Parts

  • Use sub-categories (HINT: Since it's a frequency matrix, there's a trick here.)
  • Dashboard size is 1000 x 900; tiled; 1 sheet
  • Distinctly count the number of orders that have purchases from both sub-categories
  • Sort the categories from highest to lowest frequency
  • White out when the sub-category matches and include the number of orders
  • Match formatting & tooltips – special emphasis on tooltip verbiage

Hardest Parts

  • Calculate the average sales per order for each sub-category
  • Identify in the tooltip the highest average spend per sub-category (see Phones & Tables)
  • If it’s the highest average spend for both sub-categories, identify with a dot in the square

I messed around with the average sales per order calc for quite a while. I have every number the same except for a couple in paper and binders. I have had a look at Ann's solution and I think mine is more accurate, so I'm sticking with it. :-)

Excellent challenge this week! Give it a try, but give it a REAL try before looking at someone's solution; you will absolutely learn more. Click on the image for my interactive version.

April 1, 2018

Makeover Monday: World Wine Production

This week Eva provided a simple data set and a simple viz from the International Organisation of Vine and Wine.

What works well?

  • It's a simple line chart, which makes it easy to understand.
  • The red line stands out well against the white background without being too bright.
  • The units on the axis are labeled.
  • The title tells is what the line represents.

What could be improved?

  • The subtitle could be moved to a caption below the image.
  • The axis has a strange scale. Why does it start at 180?
  • Adding the drop lines makes it look like the length of those lines is important, but if you compare the length of the lines, then that could be misleading due to the axis starting at 180. I'd remove those lines.
  • The year labels are diagonal.
  • Each year doesn't need a label.
  • Why doesn't the source document contain data for all of the years?

My Goals

  • It's Easter and I have basically no time to work on this, so do something quick.
  • Mimic what we created for Workout Wednesday week 33.
  • Focus on the relative change from a chosen period instead of the absolute change. For me, this is more meaningful if you want to see how much a country has changed and it normalizes all of the countries.

That's it. All done!

March 28, 2018

Workout Wednesday: Color and Ordering

This week Rody challenged us to create a tabular view that has three columns, only one of which is colored and the user needs to be able to sort by a given year and metric. Find the full requirements here.

I had done something like this previously for the color. What tripped me up was the sorting calculation. It's not overly complicated. The only hint I'll provide is that you can't sort the subcategory by this calculation. Click on the image below for the interactive version.

March 26, 2018

Tableau Tip Tuesday: How to Customize the Tooltip of a Reference Line

Tableau (as of this writing) does not allow you to customizing reference line tooltips. So how can you work around that? Simple! Check out this video and download the workbook to see how it's done.


March 25, 2018

Makeover Monday: What is the UK's Favorite Chocolate Bar?

No comments
This coming Sunday is Easter so I thought we'd look at the UK's favorite chocolate. As someone new to the UK, the Brits can be quite snobby about chocolate, particularly despising American chocolate. I mean, how can someone NOT love Reese's?

The chart we're making over this week is from CDA.

What works well?

  • The bump chart is a very nice visual display for ranked data.
  • Including the rank as a number at each point.
  • The lines are easy to follow.
  • Labeling both the left and right side so that you don't have to trace the line back to the start when you get to the end.
  • Using a different mark type when the chocolate is not ranked.
  • Simple title and subtitle.

What could be improved?

  • This is a LOT of colors and some of them are very close to each other.
  • Why are there age bands missing?

My Objectives

  • Split each of the age groups out rather than connecting them and then include a total, which is the average across the age groups. I'm making the assumption here that the same number of people were surveyed in each age group. 
  • Display the data as a dot plot along a scale from 0-10 for each chocolate bar for each age group
  • Use a brown theme to go with the connotative color of chocolate
  • Color the values using a brown scale

March 21, 2018

Workout Wednesday: Sub-Category Sales Change in the Last Two Periods

No comments
Another Workout Wednesday in the books. This week, Curtis Harris created the challenge in Luke's absence. Why didn't I think of subs last year?

Find all of the requirements on the Workout Wednesday website. Essentially, you have to provide the user with a date filter and use that filter to display a barbell chart of sales up to the date selected for its month compared to the same days in the prior month.

What tripped me up most was creating the tooltips. I'm going to have to look at how Curtis did it, which I suspect was with LODs. I used LODs to get the sales based on the filter table calcs for the tooltips since all of the dimensions I needed were already in the view.

Give this one a go before downloading any of the solutions. You'll learn more by struggling through it.

March 18, 2018

Makeover Monday: What pets do people in the U.K. prefer?

No comments
For Makeover Monday week 12, Eva posted this chart from the Pet Food Manufacturers Association about the pets that people in the UK have in their homes.

What works well?

  • The pets are ranked in descending order.
  • The bar colors match the icons.
  • Including the text in the bars saves room

What could be improved?

  • Include a title that explains the data a bit more
  • While the icons are cute, are they necessary?
  • If you remove the icons, you then don't need as many colors.

My Goals

  • Practice what I learned in last week's Workout Wednesday because reinforcing the learning helps it stick for me
  • Create a better title
  • Simplify the colors
  • Create a mobile version

That's it. Simple, quick and done!

March 14, 2018

Workout Wednesday: Candy Button Small Multiples

No comments
Rody brought back a classic technique for Workout Wednesday week 11. Get all of the requirements here. Almost exactly two years ago, he wrote a guest post on this blog about how to create the unit charts he's challenging everyone to make.

One thing he mentioned in the blog is that now you can use unicode characters in 10.5, however, I've been using them for a long time. I got to this site, find the character I want to use, and paste it into the calculation or field name.

The trickiest part for me on this challenge was getting the dots in the right place, that is, starting at the bottom right. I had to swap some of the logic of the calcs around and it was done. As for the headers above each set of dots, well, I'll leave that to you to figure out. Here's my tip: don't overcomplicate it; it's quite simple.

Again, find all of Rody's requirements here. Click the image below for the interactive version.

March 11, 2018

Makeover Monday: The Twisted Nature of Irish Whiskey Sales

St. Patrick's Day is nearly upon us and we're partnering with The Information Lab Ireland and Glendalough Distillery to give away three bottles of whisky. You can find the details here.

Also, thank you to BordBia for providing the data that they collected from The IWSR. Let's have a look at the visualization to makeover from Bloomberg Businessweek:

What works well?

  • Overall the design is clean, including minimal gridlines.
  • Showing cumulative growth since 2011 makes it easy to compare the regions.
  • The region colors are easy to distinguish from one another.
  • The subtitle tells me what the chart is about.
  • The legend is prominently placed so that we know it's important.
  • Putting the axis on the right speed processing because the axis is next to the end of the lines.

What could be improved?

  • Labeling the lines directly would remove the need for the legend.
  • Labeling the ends of the lines would tell us what the growth actually has been.
  • The title isn't very relevant to the data. Why does this make the Irish lucky? A better title is needed.
  • So what? What are we supposed to get from this?

My Goals

  • There are a lot of countries and a lot more years in the data set, explore them and look for stories.
  • Consider the regional aggregations to see if there's an significance.
  • Look at both volume and change. Looking at just change doesn't factor in the volume that each country imports.
  • This is time series data, so spend time looking at line charts.

After playing with data for a while, I remembered this great viz by Emily Chen and thought it might fit well with this data set because whiskey sales within Ireland have decreased. I used her viz as my style guide to create my Makeover Monday week 11.

March 7, 2018

Workout Wednesday: Keep an Eye on Sales

No comments
Let's all start week 10 by agreeing that Luke is mean. Just kidding Luke! Workout Wednesday is supposed to be difficult and this week it sure was. This was definitely the hardest week for me so far.

We had to build this sunburst chart:

I had built sunburst charts before (link) so I opened that workbook knowing that would at least get me started. Ok, I was quickly about 75% of the way done, but my colors were all overlapping and starting at the same spot. The requirement has them essentially stacked with a gap in between.

I tried over and over again, but couldn't get the calculation right. So I went to lunch and went for a walk to get away from the computer and do some thinking. To figure it out, I thought about it as one week at a time, figuring if I get one week right, I might have it all correct.

The trick is figuring out the length of each segment within each week. Notice how they all end at the same spot. That's a pretty big hint. Then once you figure that out, you probably won't have a gap between the colors. That part is pretty simple.

Good luck!