In “Thinking Like a Data Scientist Part I: Understanding Where To Start”, we examined the challenges for getting the business users to think like data scientists when contemplating where and how to leverage big data to drive business value. We introduced a “Thinking Like a Data Scientist” process that starts with identifying and understanding the organization’s top-level strategic business initiatives, then uses a “Strategic Nouns” technique to create potential business questions that were descriptive, predictive or in nature.
We will now complete this exercise by introducing two additional techniques that we can use to uncover new variables or metrics that would be excellent predictors of business performance.
Thinking Like A Data Scientist Process (Continued)
Step 4: By Analysis. “By” Analysis is a technique for leveraging a business stakeholder’s natural question and query process to uncover:
- Additional data sources
- Additional dimensional entity characteristics
- Additional areas for analytics exploration
“By” Analysis exploratory sentence format looks like the following:
- “I want to see sales and product margin by product category, store, store remodel date, day of week, store demographics, and customer demographics”
- “I want to trend hospital admissions by disease category, zip code, patient demographics, hospital size, area demographics and day of week“
- “I want to compare current versus previous maintenance issues by turbine, turbine manufacturer, date installed, last maintenance date, maintenance person and weather conditions”
Check out my blog titled “Leverage By Analysis To Expand Your Data Science Perspectives” that covered the “By” Analysis in a bit more detail.
Figure 3 shows an example of “By” Analysis for a hypothetical Foot Locker merchandising example from the perspective of the customer. We asked the business users (in a facilitated brainstorming session) to brainstorm the different dimensions and/or attributes of the strategic noun upon which they were focused. You would do this same exercise for each of your strategic nouns.
The significant number and variety of “By” dimensions and attributes that can surface in a brainstorming session can lead to incredible insight. And remember as you go through this process, all ideas are worthy of consideration; this is not the point to try to filter the creative ideas or handcuff the creative thinking process!
Step 5: Score Technique. The purpose of the “Score” technique is to look for groupings of strategic noun dimensions and attributes that can be combined to create a more predictive and actionable score. These scores are critical components of our “thinking like a data scientist” process by supporting the decisions that we are trying to make, and/or what actions or outcomes are trying to predict with respect to our targeted business initiative.
Scores are very important constructs in the world of data science, and can help to cement the business stakeholders’ buy-in to the data science process. The best familiar score example might be the FICO score, which combines a multiple questions and dimensions about a loan applicant’s finance history to create a single score that lenders use to predict a borrower’s ability to repay a loan (see Figure 4).
Scores can be created to provide predictive insights across a number of different industries and across a number of different business initiatives. Figure 5 shows some example scores from different industries.
So let’s build off of the variables and metrics that were uncovered in the “By” Analysis and see if we can integrate any of those variables or metrics into a higher-level score. In our Foot Locker example, we might want to group the Favorite sports, Favorite teams, High School sports and College sports into a score that measures that individuals “Sports Team Passion.” We might discover other potential scores around their level of current “Athletic Activity” (see Figure 6).
To be honest, this is probably the most enjoyable part of the process as you brainstorm additional data sources and metrics that can be used as part of your score. Again remember, no idea is a bad idea. Let the data science team decide via their analytic modeling which data sources and metrics are the best predictors of business performance.
Step 6: Close The Loop. The final step in the “Thinking Like A Data Scientist” exercise is “closing the loop” with respect to what analytics-driven scores or recommendations that we need to deliver to our key business stakeholders. You can use a simple “Recommendations Worksheet” that ties the decisions that our business stakeholders need to make (in support of the targeted business initiative) to the predictive and prescriptive analytics that we are going to need to build.
Last is the creation of the user-experience mockup that validates that we are building the right analytics and have a high-level understanding of where and how to deliver those scores and recommendations (e.g., management dashboards and reports, and operational systems such as the call center, procurement, sales, marketing, finance, etc.)
To get examples of these exercises, you’re going to have to enroll in my University of San Francisco “Big Data MBA” course. Sorry, got to save some homework for my students!!
Data scientists are critical to advanced analytics, and I believe that you cannot have too many data scientists. But an important challenge is to get your business users to “think like a data scientist” when contemplating data sources and metrics that might be better predictors of business performance. Having a business organization that can “think like a data scientist” will drive better collaboration with your data science team and ultimately, lead to better predictive and prescriptive results, and… value to the business.
Part I of “Thinking Like a Data Scientist” blog series can be found here.
To learn more about EMC’s unique approach to leveraging Big Data to drive business value, please check out EMC’s Big Data Vision Workshop offering.
The moniker “Dean of Big Data” may have been applied in a light-hearted spirit, but Bill’s expertise around data analytics is no joke. After being deeply immersed in the world of big data for over 20 years, he shows no signs of coming up for air. Bill speaks frequently on the use of big data, with an engaging style that has gained him many accolades. He’s presented most recently at STRATA, The Data Science Summit and TDWI, and has written several white papers and articles about the application of big data and advanced analytics to drive an organization’s key business initiatives. Prior to joining Consulting as part of EMC Global Services, Bill co-authored with Ralph Kimball a series of articles on analytic applications, and was on the faculty of TDWI teaching a course on designing analytic applications.
Bill created the EMC Big Data Vision Workshop methodology that links an organization’s strategic business initiatives with supporting data and analytic requirements, and thus helps organizations wrap their heads around this complex subject.
Bill sets the strategy and defines offerings and capabilities for the Enterprise Information Management and Analytics within EMC Consulting, Global Services. Prior to this, he was the Vice President of Advertiser Analytics at Yahoo at the dawn of the online Big Data revolution.
Bill is the author of “Big Data: Understanding How Data Powers Big Business” published by Wiley.
©Bill Schmarzo, 2015. Unauthorized use and/or duplication of this material without express and written permission from this site’s author and/or owner is strictly prohibited. Excerpts and links may be used, provided that full and clear credit is given to Bill Schmarzo and with appropriate and specific direction to the original content.