Creative Analytics

-
by Billy Caughey

An encounter with creative analytics...

A few weeks ago, I came in contact with this term called "creative analytics". Here's the background of the situation. As a classically trained (meaning proof and theorem) statistician, I try to meet the assumptions necessary to use the method. After all, that is what I was taught to do in college. Data just seemed to perfectly fit assumptions, and if it didn't, transformations were easy enough to perform.

What happens when college education meets real world data? This is where my first encounter with creative analytics came into play. In reviewing some data, I found myself looking at what could be considered a bell curve. I found myself torn between two options:
  • Option 1: call it Normal and move on with my analysis 
  • Option 2: prove the distribution was in fact Normal and continue with my analysis 
Ultimately, I sided with option 2. I began recalling methods to prove this distribution was Normal. That's when it happened - a colleague of mine looked over my shoulder and said, "Wow, that's a nice Normal distribution." After a moment of trying to digest what he said, I explained what I was trying to accomplish. He laughed and said, paraphrasing, "It's probably close enough. It looks Normal enough that it shouldn't scare you. This falls under creative analytics.".

I was stunned by that response. My inner mathematician was horrified. My classically trained statistician was offended. Yet, I was curious by what he meant. After some inquiry, my colleague shared several stories with me about creative analytics. I realized he hadn't forgotten or dismissed any of the relevant assumptions of analytic methods at all. In fact, he was trying to adhere to many of them. He said, in his very nonchalant way, the world is made up of messy data. For some methods of analytics, the condition of "close enough" is sufficient. 

My thoughts on creative analytics since...

What is interesting to me is how creative analytics have always been part of my career. Thinking back on my career, I realize how often I have used creative analytics in my life. Mind you, I haven't always called creative analytics by the title creative analytics. Often, the naming was correcting a problem, developing a solution by the seat of your pants, or thinking on the fly. All of these are synonyms of creative analytics.

I have come to realize creative analytics has it place in the Parthenon of analytics. I would argue some careers are built on creative analytics. So, where does creative analytics sit then? It largely depends on your opinion and how you intend to use creative analytics.

In the example I previously presented, creative analytics won the day. The project was a first attempt to understand what we could see in the data. Business decisions were not going to be based on that analysis. Rather, the idea was to drum up further research questions from it. So, for that case, creative analytics were going to win the day.

If you run into a project in which business decisions will be made, there still may be a need for creative analytics. I share this example. In a research study I was part of, a response to a question was unintentionally left out. When discovering this, I met with a colleague to discuss the options. After some frustration was expressed, a solution was put together. The solution looked for a specific set of responses in combination with responses the question did have. If that solution criteria was met, we changed the response to the missed response. All surveys given after this discovery had the correct answer set to that question.

What I will caution is do not let creative analytics be a crutch. Creative analytics does not give analysts or researchers a free pass on bad methodology. In all 

Creative analytics lead to creative thinkers

I'll wrap up in this fashion: creative analytics really requires creative thinkers. I've watched a lot of college students enter the workforce and get stunned. The problems they encounter are not what they were trained on. I remember being that new analyst. My solution was I started practicing and programming with dirty data. Thinking through problems and coming up with innovative solutions helped me bridge the gap between a school data set and creative thinking.














Comments

Post a Comment

Popular posts from this blog

Introduction in under 5 minutes

-
A quick introduction There are a lot of blogs which have a long introduction. They go into a long description of covering the 5 W's and such. Well, this isn't going to be one of them. The goal is for you, the reader, spend LESS THAN 5 MINUTES reading this!  A few weeks ago, we had a conversation about our passion and enthusiasm for and about analytics. That conversation led to this blog. After some brainstorming, we came up with what our goal was: to share our enthusiasm and passion for analytics through learning, applying, and teaching.  Learning Even though we have degrees and professional certifications in analytics, we are still students of analytics. As a field, analytics includes areas like statistics, machine learning, and artificial intelligence. In a broader sense, analytics could include areas like ETL, data engineering, and data structures as well. This means there is a lot to learn!  We will review methods, papers, articles, book...
Read more

Where did I put that tool...

-
Image
by Billy Caughey Introduction In a previous blog, the analytics tool belt was discussed. Common tools were discussed and some rationale was given as to why you should have those tools. What I have come to realize is even if a tool is commonly used, there are times I still don't have it in my tool belt. I will be the first one to admit I have had to go looking for a tool in the hopes of adding it to my tool belt. In those cases where you don't have a common tool, what are you supposed to do? Where are you suppose to look to find a common tool? When do you know you've added the tool to your tool belt? Let's explore these questions a little more together. Where can you pick up the tool? I wish there was a magic answer I could give you here, but there isn't. There are too many options to list where you can pick up a tool. This shouldn't give you any stress. In fact, this should excite you. There are plenty of analytics "tool stores" for y...
Read more

A Modeling Process

-
by Billy Caughey Introduction Have you ever been called a data wizard or witch doctor? I have! Let me give you an example of this. I used to work with a domain expert who had a jokester and jovial side. On one occasion he asked for a very sophisticated data analysis. He jokingly said, "Billy, don't you just have a Stata button you can push and produce the results?". We both laughed because I knew he was joking; however, that moment has stayed with me. This domain expert had some knowledge of statistics and knew this wasn't how it worked. Unfortunately, I would argue for ever one domain expert or business leader that knows how difficult analytics can be, there is at least 10 who do not. So, what in the name of wizardry is the secret? It's really quite simple. It's knowing a modeling process that can structure your efforts. A modeling process can help put everything into perspective, define timelines, and help understand how to present results back to...
Read more