J O E   D O N A H O E

Twitter   Github   Tableau Public

2018.09.07 :: Worldwide R&D Spending

This is another post from Mr. Kriebel's blog. It started with a horribly chart-junky viz from another blog, and Andy focused on the US portion of that. He used dots, which I thought was a little too cool. Instead, I used bars to show that global spending is rising, and US spending is mostly flat. Because comparing the US portion to the global portion is visually difficult, I added the US percentage of the whole as a line.

Sometimes the data doesn't require much visual trickery. A lot can be done with just bars and lines. But this wasn't standard Tableau formatting. I wanted the viz a certain width for the header text and source attribution. That made the bars spaced too far (I felt) to be easily visually compared. So I played with the bar width to make sure my brain didn't work too hard seeing the slight increases in US spending.

I would like the full data set, but this was enough for today's exercise. The ugly-viz blog got their data from UNESCO, and the full data is there. That may be a project for another day. Check out the UNESCO--the animations are unnecessary, but the ultimate visualizations are pretty good.

2018.09.05 :: NBA Salaries Against the Cap (VizWiz Makeover Monday)

This is from an older Makeover Monday post of Andy Kriebel's. His original post is here. The original visualization from What's The Cap was actually pretty nice, if stylistically old. I thought they were missing an important element--does the cap overage help win? That's the point of going over the salary cap, right? The [m/b]illionaire owners gamble on the extra salary and the financial penalties, if they think it will bring them championship glory.

So I added winning percentage and whether the team won the championship for a handful of the years. The data comes from Basketball-Reference.com. Each year's data was pulled separately, then compiled in an Excel workbook. I also modified the season nomenclature (I would say "2017-2018", Andy says "2018", and Basketball-Reference.com says "2017").

I really like the simplicity of this, along with the clear resulting message. Between 2010 and 2018, big spending didn't really help you win in the regular season, or win the championship. But you did need to spend at least a little to win more than 70% of your games or the championship (see all those blue dots on the bottom of the graph).

Then here is this more complicated visualization.

You can click on a team on the left, and find them in the mess of lines on the right. Either of these visualzations are mediocre by themselves (the left side is simplistic, and lacks the time element; the right side is too much data to use). But the combination through actions makes the data more understandable.

2018.09.04 :: VizWiz Monday Makeover - Where Nike Makes Stuff

This is a response to a this. This is a fun little exercise. The data is by factory/country/city, with some percentages at the record level. I used those percentages to get totals, then used those totals to aggregate up. The by-factory data was a little interesting (example, one Vietnamese factory has a lot more migrant workers than the others). It was also tempting to use a map with the geographic data (news flash: Nike has a lot of factories in China).

But the best simply rendering of the data was below. By setting the axis to the same scale (default in 9.1, not in 2018.2), we can see where the workers are. I would like a little more time with this, but today's time is up. In conclusion, I think mine isn't that great. Maybe a good one to rework someday.

2018.09.04 :: Get Git Set Up

I want each of these exercises to be done in one hour. I was planning on giving myself 30 minutes to get my Git environment setup and I'm 29 minutes in, and still haven't documented my Git commands or pushed back up to the server.

Here is how I got started:

  1. Open Git Bash. After installing Git, of course.
  2. Go to the local folder where you want to do your work. I did mine in simply c:\git
  3. Run git init. Then I did git config --global user.name "joedonahoe" and git remote add origin https://github.com/joedonahoe/joedonahoe.github.io. I had these in an earlier document, but I'm not exactly sure why I needed them.
  4. Now to pull the existing objects down. Run git clone https://github.com/joedonahoe/joedonahoe.github.io.
  5. And run git status. Nothing has changed. You can also run "dir -all" to see the files' current state.
  6. Make some changes. For me, I used "rm *" to throw away all of the old stuff. Then I created a new "index.html" in the folder.
  7. Run git status, and I have a lot of changes to send to the server.
  8. Then I ran git commit -a -m "2018 Reboot" and git push origin master.

This is intended as a minimal Git reference. I highly recommend git - the simple guide and My Git Workflow for further reading. This little bit was supposed to take me 30 minutes, and it took my 49 minutes. Oh well.

2018.09.04 :: Intro

This is my practice space for Tableau. Maybe I'll get other tools here (JS, Python, R). But just Tableau for now. And I'll give myself a time limit so I don't have much time to play and polish things. I will mostly work off of Andy Kriebel's Makeover Monday columns over on VizWiz. If you have a moment, I would recommend reading Mr. Kriebel's blog. This little site is served up on github.io, and built using handwritten HTML in Visual Studio Code.