July 25, 2015
I decided to take a look at the specific activities I was performing. I ended up grouping them in more general categories to improve the visualisation; I prefer a bit more simplicity in my life as well as my vizzes. I used several methods for data collection this week: Swarm, Moves, Fitbit, Runkeeper, IFTTT, Sunrise Calendar. From there, I looked at a few specific categories:
- Overall rate of tracking - I was curious to see how much of my time I was actually able to account for.
- Relocation - I knew I was blowing off packing and the like, mostly because I hate it. The data proved this out.
- Sleep - Was the way I was feeling overall possibly due to a lack of sleep? I probably should have looked at sleep quality as well, but I didn’t include that data.
- Family time - Was I spending enough time with my family? This is always a huge priorty for me.
- Running - I was smack in the middle of marathon training. Was I completing my training? Was that impacting anything else?
- Work - Tom knew I wouldn’t be working much this week, but would I get ANY work done?
Given this set of goals, I explored the data in Tableau and created a few key stories. The last two tab in the story are images of the postcard.
Click on the image below to explore the story. Enjoy!
July 21, 2015
So this week, I wanted to introduce you to a simple way to create sparkline indicators. Basically all I'm doing is adding a dot onto the end of the sparkline and color-coding it for a call to action.
Click on the image below to go watch the video and to download the workbook.
For more information, here's a blog post I wrote in 2013 for how to create these as well, but I think the video above is simpler.
July 20, 2015
What struck me most about this choice of chart is how difficult it is to see the trend. Because of its design, your eyes have to move from quarter to quarter, match up the bar colors, then once you hit Q4, go all the way back to the left. This is way too much work for the reader. On top of that, you have to calculate the rate of change in your head.
I created this simple version, very much in the spirit of what I leared at one of Cole Nussbaumer’s workshops.
What I’ve done is:
- Visualize the entire time period as one continuous line
- Highlight the beginning and end of the time period
- Include the rate of change from the start to the end
This view makes the story much, much simpler to see and adds the necessary context. You can download the data here and the Tableau workbook here.
July 17, 2015
I manually created a spreadsheet of our total connections each and our connections in common. I then created random points on a graph in order to display them as a hub & spoke diagram. I referred back to my own blog post for help in shaping the data correctly. This was a bit trickier, though, as I wanted to show Jeff on the left, our common connections in the middle, and my connections on the right. You can download the Excel spreadsheet here.
Some explanations about the viz:
- You can click on any combination of things in the bar chart to update the network diagrams.
- LinkedIn only shows me when we hit 500 connections, so I totally made up the total connections.
- I chose blue for me because it's my favorite colour and red for Jeffrey because he lives in Cincinnati and that's the home of the Reds.
July 14, 2015
Week 8 was supposed to cover the period from May 25-31, but my data collection had a major fail. What I’ve done instead was use IFTTT to capture all pictures that I like on Instagram and log them to a Google Sheet. Note that IFTTT records the date that the picture was taken, not the date that I liked the photo, so the dataset reflects photo dates. Good enough for me!
I then exported the data from Google Sheets into Excel and did some date manipulation before importing the data into Tableau. Once I had the data in Tableau, I began to explore the data to see if any patterns emerged, focusing primarily on whether I liked running or non-running pictures the most. The story points below reflect my thought process. Enjoy!
I showed them again on Friday during the Data School weekly presentations, but apparently there were actually some people that didn’t watch. So I recorded them again this morning for this week’s blog post. Enjoy!
You can download the workbook here. Click on the image below to watch the video.
July 13, 2015
Yikes!... @VizWizBI this needs a little TLC. Makeover Monday Candidate 7000. #FriendsDontLetFriendsUsePieCharts https://t.co/xQdolJjYK6— Tim Messar (@TimMessar) July 13, 2015
Tim linked to this tweet from Kirk Borne promoting an exploding 3D pie chart.
Top 20 Data posts 2015-to-date: @Dataversity http://t.co/U5lzHGuYh6 #BigData #Analytics #DataScience #MachineLearning pic.twitter.com/yBh9OCb4hT— Kirk Borne (@KirkDBorne) July 12, 2015
My first thought was back to the great Darkhorse Analytics post entitled “Salvaging the Pie”. I thought I’d follow a similar process with a combination of Excel and Tableau. I started by reproducing the 3D pie chart in Excel, which looked remarkably like the original, including the default colors. Follow through the story points for the step-by-step process of the makeover.
You can download the Excel file here and the Tableau workbook here.
July 2, 2015
Tableau On Tour comes to London next week. This will be my first conference as a partner, so I’m interested to see how the experience is different. You’ll likely be able to find me hanging around The Information Lab booth, which you won’t be able to miss in all of its orangeness.
First, I know I won’t be able to compete in the Information Lab Speed Challenge.
But you can, and you can win a drone. This year’s challenge include a bit of Alteryx as well. If you’re not an Alteryx user yet, ask us for help. Don’t worry, the Alteryx piece will be simple enough for anyone to follow along, even a brand new user.
I’m very excited to be hosting a breakfast Wednesday morning with founder, boss and friend Tom Brown. We’re going to be talking about the Data School and how the experts we’re building will be instrumental in the world of data analytics. If you think you’re company would like to take on one of these consultants for a 6-month engagement after their training is complete or if you’d like to have the School do a 1-week project for free with your company’s data, email me and I’ll get you into the breakfast. Trust me, this isn’t a group of talent you want to miss out on.
The consultants at the School will be very visible throughout the conference. If you see one of them, stop them and ask how things are going. Ask them to show you what they’re working on. Find out how much they’ve learned in only two weeks of training.
Lastly, I thought I’d share my agenda. I’m hoping I get to go to this many sessions, but I’m fairly sure things will change. I’d always rather talk shop in the hallways than go to sessions. These types of conversations are what the Tableau community and Tableau conferences are all about; these are how you get value from them.
See you at The Brewery Monday!
June 30, 2015
The basic reasons behind this project were twofold:
- I was learning Alteryx and wanted a use case to apply what I was learning.
- I'm training for my first marathon and wanted a better way to see all of my runs in one place.
June 29, 2015
I found an infographic on Twitter this week from the Wall Street Journal that described the average American's day.
What Americans did with their time last year: more work, more sleep, more TV, less school http://t.co/MhLZkOBBdU pic.twitter.com/TIJB6FR4y1— Nick Timiraos (@NickTimiraos) June 25, 2015
There are several problems with this infographic:
- It's really hard to see easily what American's spend most and least of their time doing
- It's difficult to compare the years - the colour encoding helps, but you have to work out the actual change in your head
- Your eyes have to look around a lot to get the whole picture
- There's a lot of reading involved
- The squares equal one minute - but it's hard to compare values using area I had a go at making an alternative.
I have to admit I couldn't find the same data set that the Wall Street Journal used. I downloaded this one from the Bureau of Labor Stats (hence my numbers don't exactly match).
In this version I have:
- Changed the infographic to a bullet chart - the bars show the 2014 values and the reference line 2004 values.
- Sorted the bars by the 2014 value
- Coloured the bars by the change, to let you easily spot increases / decreases
- Added tool-tips to show the actual change (rather than doing the Math in your head)
You can download my viz from Tableau Public here.
Would you do anything differently?
June 25, 2015
As part of this exercise, we were building a dot plot and Laszlo Zsom asked how to connect two dots on the same row. I hadn't ever done it before, so I used a Gantt chart to connect them. Then Chris Love suggested using lines.
In this week's tip (two days late as it is), I demonstrate both of these methods. Click on the image below to enjoy the video.
June 22, 2015
A few weeks ago, the Guardian Datablog published this series of circular heatmaps to represent monthly rainfall across a 20 year period in three Australian cities.
There are several problems with using radial heatmaps:
- They are not too dissimilar from geographical heatmaps in that they tend to skew towards the segments that cover the most surface area, in this case 2015.
- It’s difficult to compare across years, across months, and across months and years.
- Your eyes are drawn all over the place.
- There’s very little sense for trends.
- Like a pie chart, you’re trying to compare the angles of the slices, which is nearly impossible.
There are some other issues with this particular implementation:
- The hover does not work once you get to the smaller segments.
- The labels are quite hard to read.
- When I hover, all I get is the value. This means that I have to look back to the labels to see which month and year it refers to.
Given these problems, I created this alternative version.
In this version, I have:
- Taken the radial heatmap and flattened it out. I liked their idea of using a heatmap, but needed it to be easier to read.
- Added two trend charts: (1) cumulative rainfall, (2) monthly rainfall
- Added a selector for the city
- Added a highlight action on the year
- Included informative tooltips
- Improved the title
You can download the viz from Tableau Public here. What would you do differently? What could be improved in my version?
June 17, 2015
A very common question and feature request I see on the Tableau Forums is to show the axis above a chart rather than below, as Tableau does by default.
When I was working on solving this, I started by looking at the XML for a workbook and there is a bit of code that controls whether an axis displays. I built a dual-axis chart to see if I could change the display in the code.
In the <style> section, note the value=‘false’ setting. This is what is hiding the axis. I changed only the top axis by setting it to ‘true', but when I opened the workbook again, both axes were displayed. Back to the drawing board I went.
Here are the steps to follow to display the axis above the bar chart and not below (sort of):
Step 1: Create a bar chart.
Step 2: Duplicate the measure that is shown in the chart by right-clicking on the measure in the Measures list and choosing Duplicate.
Step 3: Drag the duplicated measure onto the secondary axis.
You should now have a chart that looks like this:
Step 4: Change the mark type for both measures to Bar.
Step 5: Remove the Measure Names field from the Color shelf. This isn’t required, but you really don’t need to have two colors for the same measure.
Step 6: Right-click on the bottom axis and choose Format. Change the Font to white and set the Ticks to ‘None’. This doesn’t hide the axis (hence the reason I said ’sort of’ above), but it will give the appearance that the axis isn’t there.
Note: The reason that you have to duplicate the measure has to do with how formatting works. If I used the same measure again for the secondary axis (i.e., Sales in this case), then when I format Sales, it applies to both axes. However, when you duplicate the field, Tableau treats it like a totally separate measure with its own formatting.
Step 7: Double click on each of the axes and remove the titles. This isn’t required, but it makes the axis narrower.
That’s it! While this isn’t a perfect solution, it works. Also, be sure to clean up the tooltips since they will show the measure twice.
You can download the workbook used to create this viz here.
June 4, 2015
For week 3 Jeffrey sent me a donut chart by accident (so he says…) and while I was thinking about my ideas for this week, I started connecting some data points with lines and ended up with a radar chart. (See the draft version in the story points.) It’s funny how he and I both have gone against what we would consider best practices. What does that mean??
I decided to go with a clock them this week and split the data up between morning and afternoon. From there, I plotted each day going outward from the centre for that hour. For example, at 12am, Monday is closest to the middle and Sunday is farthest from the middle. This helped me see which hours were cumulatively the most frustrating for me for the week. Each dot is separated by the frustration level. If the frustration level was three, then the dot would be 6mm from the previous dot. I then sized the dots by the frustration level so as to double encode the values.
It’s no surprise that my sleeping hours were generally the least frustrating, except for 5am when jetlag kicked in. Overall, 9am was my worst hour in the morning and 1pm was the worst in the afternoon. The story points viz below goes into more of the explanations. Click on the image to read through the story.
June 2, 2015
Click on the image to launch the dashboard that contains the video.
June 1, 2015
The article is trying to emphasize the change in the share of Apple revenue in China compared to the Americas. Here are some problems I have with this chart:
- It’s very hard to compare stacks in a bar chart over time because each stack is influenced by those stacks below it.
- The title of the article and the chart don’t match. The article says China vs. the US, but the chart is China vs. the Americas.
- There is not enough emphasis on the comparison. The rest of the regions should fall to the background.
- I don’t like the legend above the chart in this case because I’m constantly having to go back and forth.
Here’s what I’ve done differently:
- I changed the title to reflect the purpose of the story.
- I changed the chart to a line chart to make it easier to see the trends for each region.
- I’m only emphasising the Americas and China. The rest of the regions are in a light grey.
- I’ve added annotations to make it easier for the reader to see the values.
- I removed the color legend as it’s not necessary since I’ve labeled the end of each line.
Which version do you prefer? What would you do differently? There are so many ways to redesign charts and no single way is 100% correct.
May 29, 2015
I'm a huge quantified self data collector, which you'll likely see throughout my Dear Data Two work. I wanted to see how I could use Alteryx to help me get the data into Tableau for analysis before creating my analogue version because I feel like the best way to learn a new tool is to find a practical application. This is the first workflow I built on my own in Alteryx. It might not be the most elegant or most efficient, but I sure did learn a lot along the way. You can download this workflow here.
One of the things I have started to like the most about Alteryx is that I can push all of the complicated row level calculations that I used to do in Tableau to Alteryx, which in the end makes Tableau much faster. For example, I used to multi-row tool to calculate the distance between two geographic points recorded by my watch.
From there, I created the dashboard below to explore the data. In particular I wanted to view the maps and see the summary stats. One thing I learned is that I need to figure out a way to account for times that I paused my watch; that data doesn't appear in the GPX files.
Click on the image to interact with the dashboard.
Exploring the data Tableau helped me quantify my runs for the week, but that didn't account for all of my physical activity for the week. To capture ALL of my activity:
- I noted my total daily steps from Fitbit.
- I calculated the number of steps for my runs by taking my stride rate of 184 strides per minute from TomTom and multiplying by the minutes I ran in Strava.
- I subtracted my running steps from the total steps to get my walking steps.
- I used the time of day that I ran and roughly calculated the proportion of walking steps before and after each run each day.
May 25, 2015
I first got a demo of Alteryx from George Mathew back in San Diego at TCC12. I was working for Facebook at the time, Mike Roberts from InterWorks set up the meeting, but I didn’t see a particularly good use case immediately for Facebook. Why? Facebook Data Engineers have always (and probably always will) code their own pipelines.
The day before heading to Inspire, I was sitting with Robin Kennedy and told him that I wanted to get a headstart on my training. Low and behold, he showed me all of the fabulous training modules that are built right into Alteryx. I had no idea! I completed about 15 of these on the flight to Boston.
After watching Arsenal draw 1-1 in a drab affair Sunday morning, I headed to the first of three training sessions: Predictive Analytics for Beginners. In this class I learned how to apply different data investigation techniques to help me understand how predictive a data source may be. The instructor showed us how to use the Association Analysis, Violin Plot, and Field Summary tools.
The workflow that we created...
…resulted in this series of violin plots (apologies for the blurry image).
The regression analysis workflow we created...
…resulted in a series of tables and this chart (which shows that charting is not in the Alteryx sweet spot).
The third and final class I attended on Sunday was Intermediate Macro Development. This was a pretty simple class in which we built a workflow + macro to strip heading from a messy Excel spreadsheet.
|The Information Lab team ALWAYS has fun!|
|Nice photobomb by the TIL team!|
|The quantified self work of Tim Ngwena of TIL was a keynote highlight!|
In the end, Inspire15 was a fantastic experience for me, a new Alteryx user. I’ve already started applying what I’ve learned and am working on two blog posts. My only regret is that I didn’t start using Alteryx sooner.