Lessons Learned – Part Two

Self-reflection is a great way to make improvements.  If you are not aware of what is going on as you do it, you can keep making the same mistakes. After my first two cases, I commented on them in my first Lessons Learned, which was based on How India Eats and The Bitcoin Economy.  Those lessons learned were:

  • Keep it Simple – A simple graphic is better than a complex one
  • Simple is Hard – Simple and informative is hard to achieve
  • No one tries to make bad visualizations – Trade-offs often drive what we decide to visualize.
  • Design is Relative – Everyone will look at the graphic differently.  People have different intuitions and preferences that will impact how they understand things.
  • It’s all about understanding – Ultimately, we need to create visualizations that are as interesting to the right audiences as possible in order to improve overall understanding.

I internalized those lessons on my next three cases studies (Baby Boomers, Solar Eclipse, and U.S. Tourism).  Those case studies led me to learn a few new things and where I should focus my future efforts.

Lots of ways to do this

A good data visualizer has a wide range of tools available to him or her.  In the course of creating the visualizations for the five different case studies, I used HTML, CSS, D3 Javascript, Excel, R, Python, Gephi, Paint, and Photoshop.  There is no one way to create a data visualization.  Each tools has it advantages and disadvantages.  As a Data Scientist, knowing what those are for each tool and then being able to mix and match those tools is an essential skill.  It’s almost like conducting a visual symphony.

In the course of this project, I even came on a couple of different tools that could have come in handy, such as Tableau and Bootstrap, that I have never had exposure to.  So, not only is it important to understand how to use everything in my tool belt, a data visualizer needs to keep expanding and deepening their toolkit.  Not only do we have to know all of our instruments in the symphony, we have to keep learning new ones in order to keep developing as data visualizer.

Sometimes, you surprise yourself

The one image I am most proud in this project is the Baby Boomers data visualization:

I made this?

This was the first time I created a data visualization that I could actually see as something that I could stumble across in article.  The colors are crisp, it lets the data tell a coherent story, and looks professionally done.  This was the first time that all of the different tools and insights I have about data visualization came together to create something worth publishing and beyond what I thought my skill set was.

No Such Thing as the Perfect Visualization

I love my visualization that I created for the Solar Eclipses.

Ain’t it pretty?

This visualization was completed right after I did the Baby Boomer visualization above.  It was another example of me using everything I have and reaching beyond that to improve my end product into something publishable.

I can also point out things that I don’t like about it, or what I might want to I can find all sorts of things wrong with my improved visualizations.  For example, the two eclipses look like eyes to me, and it looks like the left eye is winking at me! Is it, I’m not sure, but it’s distracting!

When I did the network graph for the U.S. Tourism case study:

Sketch 3 – Final – Labels and Titles Help

I struggled with how best to show the quantity of people traveling to the United States from each country.  I’m still not entirely happy with how it came out.  Despite me thinking of them as completed data visualizations, they are forever imperfect.  I do think I came up with the best solution for now, but that won’t stop me from thinking of better ways this can be highlighted for future visualizations.

Where to Next?  Keep on improving

When, I started this, I did not think I would write over 16,000 words on making data visualizations better.  The most interesting thing, is that as much as I have learned about what makes for good data visualization, I have learned what I need to do become better at data visualization.  It’s a long hill to climb.  To begin that journey, I identified two focus areas on where to go next with my data visualization abilities.

First, I’m going to focus my efforts on improving my d3 and javascript programming skills.  I was disappointed when I realized that creating something like the Avenger visualization was out of my range of my skills.

Awesome looking Avengers network graph.

There’s always room for improvement.  If we can think of something, that should be incentive for us to figure out how to make that happen.

Next, keep making visualizations.  I have the basic tools.  Now, I just need to keep practicing.  If I don’t make visualizations, not only when I not get better at it, but my skill set will decay.  I have to admit, I currently enjoy being able to write SVG code in d3 now without having to look at a previous example.  That’s called learning by practicing.  I don’t want to lose those skills that I have fought for all semester long.

Thank you! Please let me know your thoughts in the comments below.

Acknowledgments

Lastly, I’d like to thank the Professor Ahn and the teaching assistants for encouraging and supporting this semester project.

References

  1. http://bl.ocks.org/nbremer/864b11eb83aac3a1f6a2