Data scientist Susara van den Heever describes herself as “someone who likes to solve problems using math.” As a practice leader within the IBM Data Science Elite Team, van den Heever gets lots of opportunities to solve problems for clients using her particular expertise in decision optimization, a key facet of prescriptive analytics.
Take a look at this video to hear her explain it in her own words:
Using decision optimization to solve client problems
In an earlier post, van den Heever’s colleague Carlo Appugliese explained the overall structure of Data Science Elite Team engagements, which are available at no cost to IBM data science customers. In her own work with clients, van den Heever applies her training in operations research (she has a Ph.D. in chemical engineering) to business problems that require smart decisions in the face of tough constraints.
Her working process typically goes like this:
- Talk to members of the business to understand pain points, whether that means overflowing warehouses or getting the right e-mails to the right customers at the right time
- Determine a measurable business goal, for example increased revenue, that can be translated into mathematical terms
- Write down the problem in ordinary business language
- Transform the natural-language problem into math so it can be addressed by data science algorithms and software
Van den Heever’s specialty in decision optimization allows her to help client companies better manage constrained resources of all types — people, money, machines, warehouses, and so on. In some cases, the work involves the creation of custom code, but van den Heever and her team routinely use IBM Decision Optimization, IBM Watson Studio, and other IBM data science solutions to deliver their results.
There’s also a human side. As van den Heever points out, the point of her work is to turn prescriptive analytics into action that changes outcomes for the client company; achieving that requires showing business stakeholders the seriousness of the problem, and then getting them involved in executing the solution.
As she puts it, “My career has been about using math to make the world a better place, to make businesses and people run more efficiently, save money, save time — it’s always been about that.”
To hear more of van den Heever’s thoughts on her work, including the importance of clean data and how prescriptive and predictive analytics complement each other, listen to this episode of the Making Data Simple podcast from August 2018.
Data Science Elite Team sessions at Think 2019
You also have a golden opportunity to learn more from van den Heever in person by attending her Think 2019 session, “Make Better, Faster, Smarter Decisions by Combining Machine Learning and Decision Optimization” (session 2920A). Here’s the description for her session:
What if you could reduce your planning process from a week to an hour, or from an hour to a second? What if you could, at the click of a button, improve your bottom line by double digits? Come learn to do just that by leveraging IBM’s powerful machine learning (ML) and decision optimization (DO) technologies together. You will learn the differences and complementary strengths of ML and DO, learn best practices, and see examples of combining these technologies to achieve financial gains and efficiencies from a selection of IBM Data Science Elite client use cases. The discussion includes demos of combining ML and DO in IBM Watson Studio.
And don’t miss these other sessions from members of the Data Science Elite Team, starting with Seth Dobrin, VP and Chief Data Officer of IBM Analytics:
- One Year of IBM Data Science Elite Client Engagements: Best Practices and Lessons Learned (session 6860A)
A year ago, IBM created a Data Science Elite team—a group of data scientists with a mission to ensure that IBM clients achieve early and measurable success with their data science endeavors. In this session, you’ll learn about Data Science Elite projects completed during the team’s first year of existence, across industries such as banking, energy, healthcare, and retail, as well as hear about associated best practices and lessons learned. You’ll learn how to select a use case for a first data science project, and hear tips on getting buy-in from the business, potential pitfalls and how to avoid them, and how to combine different data science techniques to get more measurable value from your data science investments.
- IBM’s Data Science Elite Help Improve Experian’s Operational Efficiency (session 6869A)
Experian collects data files on consumers from banks and credit card companies. A small amount of this data is not correct. Currently Experian uses a set of about 400 rules to try and determine if a file should be loaded or rejected, triggering many false positives that require manual intervention. A solution based on IBM ML technology helped reduce the number of false positives. The new system also can adapt as new failures are observed, and automates a formerly manual process using AI.
- Banco Macro Uses AI to Anticipate Client Needs in a Dynamic World (session 7480A)
Banco Macro in Argentina, in an effort to serve their customers better, wanted to anticipate their needs to offer them the right products and services in a highly complex economic and political environment. In this session, you will learn how Banco Macro, leveraging the expertise of IBM’s Data Science Elite team and IBM’s Watson Studio, developed, deployed and maintains their machine learning models and integrates the results into their daily workflow.
- Turbocharge Your Data Science Practice on Watson Studio—Centralized, Distributed and Hybrid (session 6972A)
To win with AI, you need not only data scientists but also people with titles like data engineer, business analyst, DevOps engineer and application developer. Is your company struggling to find the right resources and enable them to produce better results along the data science lifecycle? Come and learn from an experienced data scientist from IBM’s data science elite team on how to collaborate, share assets and drive learning to produce results with a variety of data science organizational designs. We will discuss centralized, distributed and hybrid data science practice models and show with a live demo how IBM Watson Studio can meet varying requirements including deep learning, visual recognition and natural language processing.
- Improving Call Center Experience with Machine Learning at DBS Bank (session 6866A)
DBS Bank sought to improve customer experience, reduce transfers and reduce call handling time at their call centers, by understanding intent and change in intent in customer interactions. DBS partnering with IBM Analytics Data Science Elite team to co-create a solution (still in development). It will use machine learning and text analytics to determine caller intent and feed this to customer service representatives. They can offer direct solutions more quickly for a better customer experience. The DSE method of co-creation offers side-by-side sharing of expertise between IBM’s expert data scientists and the client’s data scientists.
Learn more about decision optimization and the Data Science Elite Team
Have you registered for Think? If not, don’t delay — follow this link to register now. We look forward to seeing you in San Francisco on February 12–15!
Meanwhile, find out more about how IBM helps its data science clients through engagements led by expert problem-solvers like Susara van den Heever at the Data Science Elite Team page.