December 3, 2016
Friday at the Data School, we had Rhiannon Fox in to run a workshop about her design process and how the Data School can apply that to their work. If Rhiannon’s name is familiar, it’s because we used her great Bermuda population growth infographic for week 30 of Makeover Monday.
We like to expose the team to multiple viewpoints throughout their training. Not every process resonates with every person, so providing them broad opinions helps them find their own way and their own process. Personally, I really enjoyed the workshop with Rhiannon as it opened my eyes to a whole new way of researching ideas and getting inspiration.
As a bit of context, the first day of every Data School cohort starts with me teaching them about data viz best practices, making choices, and critiquing visualisations. We use a lot of A2 paper and crayons! Then, about a month into their training, we have Caroline Beavon come in run a workshop. Caroline’s workshop is based more on infographics and storytelling as she comes from a journalism background. Her workshop builds upon mine. Rhiannon’s then builds upon Caroline’s, however, her approach is more from a graphic designer’s perspective.
I‘ve learned a ton from both of them, but don’t want to give away all of their secret sauce. Contact either or both of them if you’d like to hire them to run a workshop for you. Both of the workshops are absolutely amazing!
In Rhiannon’s workshop, I ended up missing out on a lot of the sketching parts due to meetings and the team presented some really amazing drawings they did for their project. The hardest part was NOT using Tableau to explore the data. In fact, the data exploration is done through the drawings themselves. Since I was not able to catch up, I cheated and used Tableau. Here’s what I built in about an hour or so. Click on the image for the interactive version.
December 1, 2016
Makeover Monday participant Will Chen created this chart for week 48 that compares the share of wealth of different groups to the immigrant population. If you’re not following Will, you should. He’s been creating videos of his Makeover Monday process.
Instantly, I saw a more compelling story and decided to take his viz and iterate. In just a few steps, I changed his viz into a story about the how the change in wealth of the top 0.5% seems to trend along with the percentage of immigrants in the U.S.
Correlation does not mean causation, however, it does lead me to think…will President-elect Trump’s stance on immigration lead to less wealth for him and his peers? I guess we’ll find out in a few years. The Tableau version is embedded below.
November 30, 2016
Yesterday I wrote about how much I liked a World Series viz created by Business Insider. One of my favourite ways to learn Tableau, and one I highly recommend to everyone, is to reproduce work that inspires me.
What was most fun about creating this viz is that it’s built completely with ASCII squares. Yes, I use a measure for the axis, but the measure is merely a placeholder. I learned a lot creating this viz this way; basically you can easily create a unit chart without having to densify the data by using a simple calculation that trims the ASCII squares instead. I also included bar charts in tooltips.
Download the workbook to see how I did it. In the meantime, here’s my take on the frequency of teams appearing in the World Series.
November 29, 2016
Our goal going into the session was to see if we could get to 50 tips in 50 minutes. Right before we started, we decided to let our wonderful judges and the audience count the tips. Why? Because Jeff and I have been using Tableau for a long time and sometimes we do things that come second nature to us that others didn't know.
This was BY FAR the most fun presentation I've ever given or been a part of. Thanks Jeff for being a great partner!!
If you'd like to follow along, you can download the unedited workbook of my tips here and my final workbook here.
My advice, pause the video along the way and try to reproduce the things I do using the starter workbook. Have fun!
November 28, 2016
A connected scatterplot is great for visualising paired time series data. In this case, the pair is the bottom 90% vs. the top N% as picked by the user. The line is colored by the difference between the bottom 90% and the top N%. I added dots on the ends of the lines to make the start and end easier to find.
What works well?
- Line chart is an appropriate chart choice as we're comparing two values over time
- Title is clear and simple
- Good sourcing and footnotes
- Tells a simple story effectively
- Only one axis is needed
- Title is a bit misleading as the values aren't actually equivalent
- Color choices imply democrat vs. republican
- Feels like there's a bit of extra visual clutter
- Difference could be accentuated more
This week, I again recorded all of my work along the way. In 45 minutes, I created 175 images. But this doesn't include parameters, filters and all the work done inside the dashboard, otherwise it would probably be twice as many.
You'll see in my final version that I put a lot of focus on the difference between the lines. I also used a parameter so the user can pick their own comparison. I also have a dynamic subtitle that updates based on the values picked in the parameter.
November 24, 2016
Like most runners, I love my running data! My watch syncs to Tom Tom, Runkeeper, Strava and Nike+. Why all of them? Well, why not? Naturally, I wanted to see how my training went. Was it effective? How'd I do in my long runs? How often did I run? What was my average pace? The questions are endless.
I'm also in the middle of testing a new Web Data Connector for Strava that brings back A TON of information about each run. Mix all of this together and you get a dashboard of my marathon training.
Click on the image for the interactive version (it's too wide for my blog). Once you're there, you can click on any activity and see the map update with the route of each run. The activity will also be highlighted across all of the charts.
And yes, I got the data from Strava yet I'm using Runkeeper colors. I simply like their colors better. Enjoy!
November 23, 2016
Very, very early this morning, I had the pleasure of speaking to the first Sydney Tableau User Group. My good friend Fi Gordon asked me to join them to talk about Makeover Monday, how I approach makeovers, why I do them, examples of work I’ve done, and the encourage people to get involved.
It was bit of a different atmosphere as I couldn’t hear or see anyone. Fi, being the genius she is, was messaging me on Twitter throughout to keep me updated.
Without further ado, here’s my presentation. Enjoy!
November 22, 2016
November 21, 2016
Let's start with Kelly Martin's incredible airplane/wildlife collisions dashboard:
What works well?
- To put it simple, everything works well here. Kelly has a very intentional design style.
- Minimal use of color
- Great use of whitespace
- Excellent annotations
- Great use of icons
- Nice instructions
- Good interactivity
- Remove the size legend and indicate that through text
As for the advanced logging, Tableau Research asked this week that we enable a setting on Tableau that creates an image every time you do something on a worksheet. I had no idea what this would result in, but in exactly 60 minutes, I created 396 images. That's a lot of squiggles! Here's a gif of my iterations:
For my makeover, I wanted to implement a visualisation technique I learned from Tableau's Michele Tessari at #Data16 in his session "Artful Data: The Balance of Art and Analysis". My goal was to show the number of strikes by weekday and month, but also to show the values for each state.
Another fun week of playing with data. Here's my take on this data set:
November 19, 2016
This turned out to be a fantastic civic learning experience for me. And exploring the data with Tableau helped me understand how it all works. I know a lot of people, particularly those I work with in the U.K., that are confused by this election and how the popular candidate didn't win. To aid in this understanding, I've created this story points visualisation to explain how the electoral college works, how electoral votes are allocated, and how the election would have turned out if the electoral college was based strictly on population.
November 14, 2016
What works well?
- Simple title
- Good references
- Treemaps are generally a poor choice for showing the distribution of the words
- Way too many colors
- It's impossible to make any sense of the words and see any patterns
November 7, 2016
As you drill into the viz you get more and more detail, like this:
What works well?
- Super intuitive interface; everyone know how to drill in on a map
- Simple instructions
- Nice grouping of restaurants with a counter
- Nice use of symbols to represent a specific restaurant
- Good use of colors
- Takes too long to drill in. When I played with it, they could easily show more specific restaurants at once and not shown so many groups.
- I have no idea how many restaurants are good or bad.
November 2, 2016
What would I change?
- The red/green color scale might be tough for color blind folks.
- I'm confused by the colors on the ends of the bar because those don't represent the same thing as the color of the middle bar.
- There are too many controls for me at the top. Might this be better as a static image with a singular story? Pablo chose to make it interactive, which is perfectly fine. I simply might choose to do different. Neither is better.
- There are a LOTS of cities in this. I would show the top and bottom 10 for simplicity. Again, personal preference.
November 1, 2016
The purpose isn't to pick on anyone. I use this as a teaching method. How can I take an existing visualisation and show I would improve and simplify it step-by-step. I only detail what I think doesn't work about the visualisation.
This week we looked at the Scottish Index of Multiple Deprivation and Pablo Gomez create this visualisation (click on it for the interactive version):
What doesn't work?
- Packed bubbles are basically impossible to compare
- What does the size and color of the bubbles mean?
- Do the colors coordinate with the scatter plots? (The answer is no, but I didn't know that until I downloaded the workbook.)
- What do the scatterplots add? They all basically look the same.
- The massive image on the right takes up like 25% of the space. What does the flag mean?
October 31, 2016
What works well?
- Nice headers to help you know what is good and what is bad
- Alphabetical sort makes it easy to find a specific local authority
- Shading of the lowest 15% provides nice context
- Bar code representation makes spotting concentrations really easy
- Unless you read the accompanying website, you can't really make sense of the chart on its own as there is now scale and no explanation of how to read it.
- There's no way to rank the local authorities from best to worst (or vice versa).
- Interactivity would help with know which datazones are which
- None of the other ranking metrics are included; this only covers the overall rank
October 28, 2016
Today, let's take a look at this election viz from Rob Radburn.
What doesn't work for me?
- It's a time series visualisation, yet it's vertical. I prefer time series to be horizontal.
- I don't care for the grey background.
- Black text on the dark grey is hard to read
- Don't need the labels on both side of both axes
- Gridlines should be for the measure not the time series
- Too many filters
October 27, 2016
What doesn't work?
- Too many colors
- The slices aren't sorted
- The largest slice should start at 12 o'clock and then go in descending order around the clock
- What's the purpose of the inside ring?
- Needs a better, more focused title
This past week I saw this tweet from Elon Musk:
This led me to have a look at the data visualisations on NASA’s website, in particular, their viz of the global land-temperature index which reminded me a lot of all of the great work we saw for Makeover Monday week 20 - Global Warming is Spiraling Out of Control.
There’s so much to like about this visualisation. It has a great summary on the left with a massive number that is the centre piece of their story. Their intentional design of making the large number the focus make the line chart supplementary. The line chart is clear and simple, the legend is out of the way and the beacon on the end captures your attention.
The data is available right there below the viz so I downloaded it so that I could reproduce this in Tableau. I often attempt to recreate visualisations I like as a way to learn and practice. Because in the end, the only way to get better is to practice…A LOT!
I was able to reproduce everything bar the blinking dot on the end of the line. I also chose to fill in the circles on the grey line because I don’t care for the open circles. Lastly, I added a + to the beginning of the large callout number. I think that helps provide a quicker understanding of what the number means.
|Click on the image to download and interact|
October 26, 2016
Part of the discussion yesterday was about critiquing work others have submitted. I've shied away from this because I don't want to discourage anyone from participating. Fortunately, Santosh Patil has agreed to let me give his week 43 submission on US Debt a makeover. thanks Santosh!
Let's first look at his work:
For this makeover, I'm only going to focus on what doesn't work:
- Does it need a dark blue background? This makes some of the text hard to read.
- What are the candy stripes for on the donut chart? What value do they add other than decoration?
- What's the purpose of the globe in the middle other than decoration? What purpose does it serve?
- Overall, there's just too much going on for me, for what is essentially two data points.