Data science transformations: Learn from these clients at Think 2019

Creative and Editorial Lead, Data and AI, IBM

If you’re a data scientist or leading a team, Think 2019 is where you’ll want to be in February to sharpen your skills, share best practices and accelerate your journey to AI. If it’s true inspiration you’re looking for, look no further. You’ll hear success stories from clients using the IBM data science portfolio of solutions to help you lead their organizations through real digital transformation.

Learn from market leaders across healthcare, banking and automotive industries on how they’ve built out a winning data science practice.

Register for Think and find out more below about sessions scheduled at San Francisco’s Moscone Center.

BMW's approach to Conversational Support through its Watson-based enterprise platform 

Wednesday, 3:30 PM - 4:10 PM | Session ID: 1888A

Moscone South, Exhibit Level, Hall B | San Francisco Ballroom 216

Various services and processes within BMW Group rely increasingly on the quality of the natural language understanding of raw text documents of different formats. In recent years, IBM and BMW have worked towards improving BMW’s cognitive systems to cover textual input from various domains ranging from service desks to customer management. Besides the algorithms and mathematical models, the success of such cutting-edge projects relies to a great extent on the expert knowledge that guides the computer's behavior and understanding. That’s why IBM and BMW have worked toward structuring the experts' knowledge and storing it in a computer-readable format, ensuring sustainability, query ability, knowledge integration and disambiguation.


  • Markus Bönisch, BMW Group
  • Armin Rudert, IBM

 Session page

The Thing in the Middle: A natural language-generated canonical data model at ING

Thursday, 2:30 PM - 3:10 PM | Session ID: 4956A

Moscone South, Exhibit Level, Hall B | Room 301

This talk will explain the challenge and the reason behind the complexity of creating a natural language-generated, canonical data model. Language is dynamic, and the way we express a model should also cater to that. The discussion will also cover progress toward building a solution that generates models and performs extract, transform, load (ETL) "on the fly," with examples of what ING has created so far in terms of data models, APIs and events.


  • Ferd Scheepers, ING
  • Pat O'sullivan, IBM

Session page

Banco Macro uses AI to anticipate client needs in a dynamic world

Banco Macro, based in Argentina, in an effort to serve 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, learn how Banco Macro, using the expertise of the IBM Data Science Elite team and IBM Watson Studio, developed, deployed and maintains machine learning models and integrates the results into its daily workflow.


  • Carlo Appugliese, IBM
  • Nicolás Martins, Banco Macro S.A.

Session page

Nedbank Optimizes ATM Management with Watson Studio and CPLEX

Nedbank operates thousands of automatic teller machines (ATMs) across South Africa. When a machine is out of service, it impacts revenue and customer service. In addition, predicting outages can be difficult, and scheduling technicians to fix ATMs is a slow and manual process. Nedbank implemented a solution combining predictive models and decision optimization (CPLEX) to predict machine outages and the nature of issues and to schedule repairs according to either a metric of minimizing distance traveled or minimizing revenue losses. A 1 percent reduction in outage time per year will result in hundreds of thousands of dollars in savings. The new optimized scheduling also enables technicians to address the most urgent issues first.


  • Guy Taylor, Nedbank

Session page

Kaiser Permanente gets predictive

The introduction of machine learning and data science at Kaiser Permanente has the potential to drive next-generation healthcare applications to a new paradigm of predictive behavior and empirical decision support. The data captured by healthcare applications far outpaces our ability to analyze it. Only a fraction of historical data is used to improve member experiences and patient-centered care. Imagine a call center scenario where an organization needs to improve member outcomes. A large sample size of call center data would help correlate behaviors to outcomes, resulting in trained models that detect patterns, predict outcomes and identify common mistakes. We will illustrate this work with practical examples built with IBM Streams, DSX and WEX.


  • David Herring, Kaiser Permanente
  • John Thomas, IBM

Session page

Value-Based healthcare for improved outcomes with data science at Medtronic

In today's healthcare environment, a data-driven approach is vital to providing better value and outcomes in an evolving market of patients, providers and payers. From measuring clinical outcomes to spotting trends from devices, patient records and other relevant insights, a data science and analytics practice plays a pivotal role in improving patient health, reducing costs and managing risks at Medtronic. This session discusses how Medtronic puts high-performance analytics to work to drive benefits across transaction monitoring, segmentation, promoting right physician collaboration and hospital coordination.


  • Sean Larson, Medtronic
  • Julianna Delua, IBM

Session page

Register for IBM Think 2019.