This is the end, beautiful friend (can you hear Jim Morrison’s baritone voice?).
This is my last blog post in my 2.5 year long Tableau blogging experiment.
The quick synopsis of this experiment follows:
- 160+ articles
- many new friends
- a whole lot of learning
- many graphics produced
- many videos produced
- many quantitative techniques explained
- a lot of laughs and good times
- a huge amount of respect gained for the Tableau and Alteryx companies, their employees, their communities, and the the software products themselves
- About 7,000 article views per month – A steady-state readership established.
With that being said, it is time for me to move on.
This article explains why my experiment has now ended and what I am going to do next. Thank you for joining me along this journey and for being a part of my experiment.
I heard a great talk on motivation by Daniel Pink at the Tableau 2015 Conference. For me personally, this hour was one of the most beneficial experiences I had during the conference. The reason for this is that I have been thinking critically for quite a while about what motivated me to create my Tableau blogging experiment. I wanted to determine what specifically created the motivation for me to spend many hundreds (thousands?) of hours over the past 2.5 years writing 160 articles focused on Tableau and Alteryx.
Why Did I do the Experiment?
I think Daniel’s talk explained why I conducted this experiment. His talk made perfect sense to me with respect to where motivation originates. Now I know why I did the experiment.
The motivation I experienced came from me pursuing my passions. I took a lifetime of work experience, training in computer science, math, and natural science and blended it all together into a big experiment to see what I could create. I allowed the blog topics to meander across time because my passions are widespread and varied. My interests can change from day to day and I tend to think critically about a lot of different topics in different subjects.
I wrote about techniques in computer science, math, and natural science and set these into a framework of investigation using Tableau and Alteryx. I solved problems, created solution techniques and I shared stories of my experiences and my life. I tried to write practical articles that would help people in their jobs and careers.
While doing this work, I met many new friends along the way. I also got better in the field of analytics because I connected to a wide group of really talented people. I loved that experience of making friends I never met. I went back to school, in the way that I used to do it, by reading, learning, teaching, writing, and creating ways to solve problems. I did these things because I was driven to learn. It was a fantastic journey and one that I will always be thankful that I took.
Where and When Did The Experiment Begin?
It was 2.5 years ago (June 2013), on a beautiful spring day in Knoxville, TN, that I decided to create my Tableau blogging experiment. This experiment has been published on WordPress.com in a blog called “3danim8’s Blog”.
I really only had one question I wanted to answer by writing this blog. I wanted to know if people would read my blog if I took the time to write it. Well, I now have that answer.
Looking back on the experiment, not only can I quantify how many readers read my blog, I can also determine how long it took them to discover my work. Additionally, I have learned what readers like, what they don’t like, and what they are indifferent about. In other words, I have learned a lot. Now it is time for me to share some more of those insights with you as I close this experiment.
The End of the Experiment is Now Here
In this article, I’m going explain a few things that I learned from conducting this 2.5 year experiment. The only way for me to have learned these things, was by doing the experiment. In some ways, I feel like I just finished another Master’s degree and this article is the defense of that work.
I think what I learned is really interesting. I hope that some of the results I show in this article will surprise you. Unfortunately I can’t tell you everything I have learned, but I will unleash some new insights that I found fascinating. If you want to see more of what I learned, go back into this blog and look for the series of articles I wrote on lessons learned as a Tableau blogger.
As with any endeavor, however, my experiment has to end. I will not be continuing to write this blog with the same intensity and commitment that I have previously given it. It is time for me to trying something new.
Will I write anything else beyond this final article? Maybe I will, or maybe I will not. One thing I can promise, however, is that if I do write any additional 3danim8’s articles, you will get only best of what I can produce.
I will use the lessons learned from this experiment to produce the best material that I can. If the situation presents itself, I will only write articles that are worth reading. With that being said, let us have a look at the answers to some questions that I originally asked myself when I created the experiment.
If You Build A Technical Blog, Will They Come?
Yes. They will come.
When I was younger, I used to produce custom furniture in my woodworking shop. Each year, I built commissioned pieces for friends of mine. The reason I did it was that (1) my house could not hold everything I created in the shop and (2) I wanted to leave a legacy behind of things I produced that were tangible.
I knew when I built a piece of furniture, people would use it and appreciate it. I didn’t get that same satisfaction when building numerical models on my day job. In some ways, building this blog was like me building another piece of furniture. I know that people can use the information and appreciate the knowledge contained within, even after I stop building the blog.
What I have learned is that if you take the time to write a blog, people will find it and some may even enjoy reading it. The most surprising finding for me is that when you have a technical blog like I created, most of the traffic to the blog will come via internet searches, rather than people coming to your blog to see what you have written lately.
In the case of 3danim8’s blog, about 80% of the traffic occurs because people are searching for information on topics that I wrote about. Since I wrote about a lot of different topics, a lot of people find this blog. In fact, during Monday through Friday, about 250 to 300 people read these articles each day. That translates to a 1500+ article reads per week, or about 7,000 per month. That amount of readership has been surprising to me and is reflective of the excitement and enthusiasm that exists for Tableau and Alteryx software.
I did not just get lucky in picking the topics I chose to write about. I happened to know that these topics needed better explanations because I discovered these problems and/or sticky points during my own usage of Tableau and Alteryx.
I also created many of my own techniques that are not written about anywhere else. I knew that these topics would be informative because I used Tableau for 7.5 years, just about every day on nearly 2,000 different data files and Alteryx on a daily basis for the past 1.5 years. In other words, I have really pushed the software to do things that went beyond normal use cases.
I also had 20+ years of continuous work experience in high performance computing, data science, database management, and geographical information systems before starting the usage of Tableau and Alteryx. That experience was a big advantage for me, including 3 years on the editorial board and 10 years as an associate editor for the journal Groundwater.
I reviewed dozens of manuscripts about topics like solving partial differential equations via analytical and numerical modeling methods. By editing technical documents like I did, learning to write effective blog articles was easier for me compared to most people. Seeing good writing and helping others communicate better is part of my experience. I used that experience to write decent 3danim8 blog articles (at least I hope I did!).
How Long Did It Take to Gain Readership?
One day I just created 3danim8’s blog without fanfare, no announcements, no support, no word-of-mouth. I intentionally did this to make the experiment as pure as possible. I didn’t broadcast my articles other than one Twitter tweet per article.
I didn’t name the blog after Tableau, I didn’t attend a Tableau conference until last month, and I was basically unknown in the Tableau community for the most part. I was (and still am to some degree) a stranger, in other words. So with that as a basis, how long did it take for people to take notice of 3danim8’s blog?
The answer to that question depends upon how you measure the term “taking notice of my blog”. In my case, I’m going to define it in a particular way via a measureable statistic.
The Definition of An Article-Day
Assume today is the first day that I wrote and published an article. That article will have a value of 1 article-day. When tomorrow comes, that article will have a value of 2 article-days, since it existed for two days. On the third day, I write another article, so that on that day, my blog will now have 4 article-days (3 from the first article and 1 from the second article). As I write more articles, I can keep summing up the article-days.
Essentially, the number of article-days is just the summation of days of each article you have written from the time it was written to the current day. Since I know when each article was written, I can easily use a couple of Tableau functions to count the days and do a running sum of all the articles to determine the article-days over time.
Next, assume that I have an expectation that for each day an article exists, one person will read it. This simply means that I expect my total readership of all my articles will be at least equal to the article-day count over time. These measures will be equal if and only if, one person reads each article each day.
During the experiment, I learned that if my total blog readership was less than my article-days, I wasn’t creating a very successful blog. If the total readership was greater than the article-days, I knew that I was gaining some readership and making progress as a blogger.
With those key measure definitions, let’s see what happened with 3danim8’s blog over time.
The Visual Analysis of 3danim8’s Blog
Figure 1 presents three pieces of information. First, the blue line indicates the number of article-days I created. Second, the brownish data points indicates actual article reads. Third, the shaded light grey (right side axis) indicates the number of articles written over time.
I included the third piece of information on this chart just to be able to visualize how often the articles were written over time since that function was non-linear and subject to inconsistency. Other than a 4 month hiatus (shown as at “Time Off”) annotation, however, I wrote fairly continuously over 2.5 years, with an average of about (160/29)= 5.5 articles per month. I can tell you that creating that many articles is not trivial.
Figure 1 shows that it took me 9 months and 56 articles before my readership starting growing more than my baseline expectation of 1 read per day per article. That surprised me a little bit because I knew I was writing good content. Those early articles included multiple detailed assessments of Tableau trend models, data preparation techniques, big data assessments, and a few Tableau fundamental articles that have turned out to be some of my most widely read work.
As I now can review the blog history, I can say that there is a lot of great information in those first 56 articles that many people have never looked at, including people that think they know my blog. So one lesson I learned was that if I had to do this again, I would save some of my really good content until my blog readership took notice of my work.
There were many other lessons I learned during this experiment and you can read about some of them here.
Also shown on Figure 1 is a reference line that indicates when I (and my family) had a major life change earlier this year. This change was when I took a new job in February 2015. This new job created an upheaval in my life that required a relocation to Austin, Texas, and then to Roswell, Georgia.
At that time of taking this new job, I didn’t know if 3danim8’s blog would continue. Luckily, I was also able to continue blogging this year due to the kindness of the Alteryx company and the support of my family and General Motors. I am very thankful for that because I needed that time to conclude the final phase of my experiment.
Now getting back to the analysis, what else is there to learn? Figure 2 shows a plot of the difference between the article reads and the article-days. I think of this plot like an airplane on the runway. The plane had to accelerate before taking off. It took 9 months of time for the plane to accelerate before being launched into the sky. There have now been about 50,000 more article reads than article-days. For any new bloggers, please realize that you have to be patient to gain an established readership. If you are diligent to produce continuous work, your readership will find you.
Figure 3 is another type of analysis I call the readership rate. This metric is simply the total article reads divided by the total article-days. This is a punishing measure as I explain in the figure caption. The reason this measure is shown as discrete points is that I only recorded total readership periodically.
For the first half of the experiment, I only recorded monthly data and then I switched to recording total reads about every 1,000 hits on the blog. As you can see, the readership rate has peaked and is asymptotically approaching a final value that would be achieved somewhere in the future. The only way for me to reverse that trend is to rapidly gain readership by writing great articles that have a broader audience than just Tableau and Alteryx.
Figure 4 shows the individual blog post hits per day over time. This measure is simply the number of total reads of an article divided by the number of days the article existed. I previously determined that my best articles have achieved a sustained rate of 4 hits per day. The size of each bubble represents the number of hits per day.
During the experiment, I applied some principles of process improvement to see if I could become a better blogger. As shown in Figure 5, I have broken the experiment down into three phases called the beginning, testing and improvement phases. In each phase, I employed the lessons learned from the previous phase.
This means that I continued to refine and focus my articles on topics that I knew would be more interesting to my readers. One of the interesting aspects of Figure 5 is that I wrote fewer low-readership articles as time went by. Also, it is very hard to write articles that get read on average over 4 times a day, over several years.
Now with this much accumulated data, I would say that any article that gets read at least 2 times a day over a long time period (>1 year) is a good article. That is a modest goal but is difficult to obtain because as an article ages, the readership rate will likely diminish because of the quantity of material now being produced.
My Tableau Influencers
This experiment was preceded by about 5 years of solid, steady work in Tableau. I burned through 1,500 different data files in a professional setting, working on data from a multitude of Fortune 500 companies. Big data, little data, ugly data and beautiful data – I’ve used it all. Every day, day after day, creating, exploring, testing, and learning. I pushed the limits of Tableau far too many times to recall.
Back in the beginning, when I started using Tableau version 3.8 in Feb 2008, there was no mapping capability. Even then, I knew that Tableau was the tool of choice. I knew that it would become what it is now. I could see the future. So I buried myself in the technology, eagerly learning as much as I could. If I could turn back the clock, I would do so to tell you how I learned what I did. Unfortunately, I can’t quite do that but I can come close.
What I decided to do to tell you this story, was to take a retrospective look at who I followed on Twitter. By going back to the beginning, when I joined Twitter in July 2008, I can tell who my Tableau influencers were over time. If Twitter recorded when we followed an individual, I would have made a time-series viz of the data. However, I do not have that data since this type of data only exists from Oct 2014 to the present, so I did the best that I could to tell you a story.
By looking at who you have followed, you can see the progression of your learning and who were your influencers. In the following video, I mention the key individuals that have helped me along the way. The video is seven minutes long but it spans about six years of time. I found this exercise to be very interesting and it helped me understand my journey a little bit better than before I did it.
I only have a few regrets about this experiment. First, I didn’t get a chance to finish some of my series, like the Wimberley, TX flash flood event from this year. I might eventually perform the science to quantify how that happened, but it is not a priority at this time. Second, I didn’t finish some articles that I started as shown in Figure 6. There are also many more topics I had hoped to write about but I didn’t have time to get to them. Third, I have never published my best work. I wish I could do it but confidentiality prohibits such an event. I dream of the day when I can show the incredible insights that I have seen in the data I interrogated. I have truly learned the power of visual analytics and that is why I have a complete love affair with Tableau and Alteryx. Lastly, I regret having to end this experiment because it has been a lot of fun.
What is Next For Me?
The good news for me is that I have another experiment to perform. That work will be called “The Jett Black Experiment”.
Whether I actually write about this experiment will depend upon the cooperation of a 4-year old. Only time will tell whether a new blog is created or whether I publish these results in 3danim8’s blog.
The mission of that work will be to educate my son Jett on how to do computer-based math, computational science, and visualization by using Excel, Alteryx, and Tableau. I want to determine and document what works best for teaching kids math, science and analysis.
My goal is to have my 7 year-old son give a presentation at both the 2018 Tableau and Alteryx conferences (Figure 7). This means that I have 3 years to perform that experiment. I know that this is an ambitious goal. I keep telling myself that there is no pressure on me. All I have to do is be a Dad. Right. I think I’m going to have to work miracles.
Both Jett and I have a lot of work to do to meet that goal. The multi-faced challenge of being a Dad and teacher to Jett is now my second priority, and it should be very interesting to see how it unfolds. I simply want to give Jett opportunities with his Dad that I never had with mine. If you want to understand what that means, read my 100th blog post by clicking here.
My first priority you ask? That priority is being the best husband that I can be by repaying Toni with unlimited love, attention and friendship that somehow got gobbled up by late night blog posts over the past few years. I need to repackage the energy, enthusiasm and passion that I sent to the blog and send it directly to my wife. She deserves it and I definitely want to give it back to her.
Saying Goodbye For Now
So for all of my readers, thanks for taking the time to read my thoughts and the lessons learned. I will continue to burn passionately for Tableau and Alteryx. I promised to continue to share my insights and experiences to future students as I continue to teach co-workers at General Motors how to use Alteryx and Tableau.
Thanks to all of you that took the time to connect with me through blog comments, Twitter, or otherwise during this experiment. Getting to connect with the best and brightest minds in the fields of quantitative and visual analytics has been amazing – so thanks to all of you.
I am especially thankful for people like Michael Mixon, Neil Sequeira, James Dunkerley, Ken Schiele, and Frederic Pinchon that have taken the time to tell me specifically how much they liked what I was doing and how my work was helping them. Those words were very meaningful to me and provided a lot of motivation for me to continue when my life became very complicated with a new job, traveling, multiple houses, etc.
How to Use This Blog As A Learning Resource
The content of this blog is extensive. Finding information is fairly easy thanks to my Tableau Public workbook that allows you to directly launch articles. Watch the video below to see how this is done.
My Favorite Goodbye Songs
Just like I began and ran this experiment on my own terms, I’m leaving it on my own terms. Each of the following songs have significant verses that explain the emotion that I feel as I close this final article. I’ve included these links because these songs say what I want to say much better than I can. As always, thanks for reading.
As I wrote on May 31, 2013:
OK. So it has been a couple of years since I last wrote to my blog. I have a lot to say, so I’ll get started with a little #Tableau training.
Today November 11, 2015:
“I have now said what I wanted to say about Tableau training.”
I guess that I didn’t expect it would take me 900 days to say what I wanted to say!