Scroll Top

Does your company need a Data Scientist?

Does your company need a Data Scientist?

TAKE NOTE (Insights into SAP solutions and Emerging Technology)

It wasn’t long ago that a company’s competitive advantage was based on its data.  The more information it could gather about its business, its customers, and its competitors, the more successful it could become.  But today, small businesses as well as large corporations are drowning in so much data that it is difficult for anyone to make sense of it.  It’s only going to get more challenging.  According to Gartner Research, over the next five years, the world’s available data will increase by 800 percent, and only 20 percent of it will be structured data from transactions.  The remaining 80 percent will consist of unstructured data, such as e-mails and comments on social networks like Facebook.

As a result, competitive advantage has shifted to advanced data analytics.  This is the capability that lets a business consistently make profitable decisions based on big data sets.  For a typical large company, a big data set is on the order of many terabytes in size.

According to Jean Paul Isson and Jesse Harriott, the authors of the book “Win with Advanced Business Analytics“,big data, as defined by IBM, has three characteristics:

  1. It has volume.  Big data consists of massive amounts of data, beyond what any individual can comprehend or any conventional tool can process.
  2. It has velocity.  Big data is constantly flowing into your organization at a very high speed.
  3. It has variety.  Big data sources can be structured or unstructured, and can include everything from keywords searched on a Web site to data from GPS devices.

To profit from big data, the data needs to be identified, structured, collected, and analyzed.  An obvious example of a company that uses big data strategically is Google.  Google leverages that data to constantly improve its algorithms for search and ad placement.  But many other Web-based businesses depend on Big Data. For instance, Zynga uses it to enhance games like “Words with Friends” so people will keep playing the game and paying for upgrades. Netflix uses big data to recommend movies to users. PayPal recently used analytics to eliminate counterproductive “introductory pricing promotions,” which were losing money because most customers closed their accounts when the introductory discount ended. LinkedIn used data its users had entered into their profiles to recommend three people who had attended the same schools or worked at the same companies at the same time.  These “People You May Know” ads generated a 30 percent higher click-through rate than any other efforts LinkedIn had ever attempted, leading to millions of additional page views and a surge in the site’s growth.

I can hear you saying…dah of course, it’s no surprise that Internet firms like LinkedIn and Google are constantly collecting data, analyzing it, and using it to drive decisions.  But many companies in more mature industries are also benefiting from big data.

  • A data scientist at Target Corporation created a “pregnancy predictor” model that could identify customers who were likely to be pregnant based only on their shopping activities.  For example, if a customer purchased large quantities of unscented lotion, followed by nutritional supplements like calcium and zinc, and then jumbo bags of cotton balls, the model determined it was highly likely she was pregnant, and the company would send the customer e-mails for products targeted at pregnant women.
  • A major airline discovered that its pilots’ estimated arrival times for incoming flights were often wrong.  As a result, passengers and flight crews spent too much time on the ground, waiting for planes, which meant the airline was forced to schedule fewer flights.  To improve the accuracy of the estimates, the airline hired PASSUR Aerospace, which offers decision-support technologies such as RightETA.  Using weather reports, flight schedules, and data from a network of its own satellites that monitor the location of every plane near the airport, RightETA accurately predicts when each plane will reach the gate, saving the airline several million dollars.

In each of these cases, data scientists helped businesses use data to improve the company’s performance or create more value for customers, which led to increased revenues or lower costs.

What is a data scientist?  As explained by Thomas Davenport and D.J. Patil in a Harvard Business Review article called, “Data Scientist:  The Sexiest Job of the 21st Century,” this is a new breed of IT professional who mines vast amounts of data to discover insights.  Data scientists combine different sources of big data, structure the analyses, and present the outputs for executives and product managers to act upon.

To be effective, data scientists must possess certain key skills and qualities.  They must know how to write code.  They need to be able to explain what the data means to decision makers.  They must be able to develop a sense of trust so they are seen as valued advisers. They must also be driven by a powerful curiosity to want to know what is causing a problem, and to develop a theory that can be tested.

Since there are, as of yet, no formal degrees in data science offered at universities, the people in this field often come from such backgrounds as math, economics, computer science, systems biology, and astrophysics.

If you determine that your company can benefit from analyzing big data, you’ll need to have data scientists at your disposal.  Davenport and Patil have found the following guidelines to be effective:

“To find data scientists, you can recruit them at the university level.  The most promising places are Stanford, MIT, UC Berkeley, Harvard, Carnegie Mellon, North Carolina State, UC Santa Cruz, the University of Maryland, the University of Washington, and UT Austin.”

“You can also find data scientists on LinkedIn, or among the members of user groups that revolve around data science tools.  These include the R User Groups, which focus on an open-source statistical tool that many data scientists use, and Python Interest Groups.”

UNDER DEVELOPMENT (Information for ABAP Developers)

In the next few installments of this blog I will be inviting a good friend of mine William Craig to teach us about ABAP Object Oriented coding and custom workflow development. Here is a little bit about Bill….

Bill Craig is the president of ASAP Consulting Inc., located in Centennial, Colorado.  Bill has over 10 years of experience with SAP ECC systems and integration with SRM.  His areas of expertise include SAP Business Workflow and ABAP development.  He made his way into programming as a COBOL programmer, switching to SAP when the company he worked for migrated from mainframe programming to SAP.  Prior to programming, Bill ran an upholstery shop in Denver, where he managed the workforce, payroll, sales, scheduling and delivery.  You can reach him at [email protected]

Using ABAP Object Oriented Coding in your Custom Workflow – Part 1

SAP Business Workflow was built on an approximation of object orientated programming called the Business Object Repository or BOR.  It uses object oriented techniques which allowed developers the ability to create copies of standard BOR objects, modify them and through inheritance allow them to be used in place of the standard.


Q&A (Your Questions answered)

Q. Is there an easy way to DEBUG SAP Background Jobs?

A. Yes. You can debug a background job whether it’s running, finished, or not started yet by using the same selection-screen parameters as when it was scheduled. There are different ways to DEBUG based on the candidate job’s status. You can debug background jobs in Transaction SM37 (Job Overview) if the job has either finished or not started or Transaction SM50 (Process Overview) if the job is currently executing.

Access Transaction SM37 and select the relevant job as shown below. Choose Extras • Debug Job to start the debugging process or (Transaction code JDBG).

An ABAP program opens in a debugger session. This is not the ABAP program that you actually wanted to debug, so you must step through a few times (you can use ) to reach the code you need.

If the background job that you want to debug is running, you can debug it with Transaction SM37 using the Capture Active Job function. Select the relevant job in Transaction SM37 and choose Job • Capture Active Job to start the debugging process.

There’s an alternative way of debugging running jobs in the Process Overview transaction (Transaction SM50). Process Overview allows you to see the details of processes running on the application server that you are logged on to. You can see both background and foreground processes in the list.

Select the job that you want to debug from the list as shown below, and use the following menu path to start the debugging process: Administration • Program • Debugging

Easy Enough!

Pin It on Pinterest

Share This

If you enjoyed this post, why not share it with your friends!