data science

data science


Christophe De Greift shares an article that contradicts popular culture and redefines the image of the data scientist. 

The ‘sexiest profession of the 21st century,’ named in 2012 by the Harvard Business Review to define the Data Scientist, has gone through various emotional states lately, depending on its field of action. Those who support understanding and predicting the health situation are heroes, whose analyzes define life and death questions such as where to send new artificial respirators. On the other hand, those responsible for predictive models for product demand, credit risk or customer churn, for example, may feel frustrated when seeing them lose precision and relevance with the sudden changes in demand and supply caused by the quarantine. How will the Data Scientists of Peruvian airlines feel about the challenge of forecasting passenger demand in the coming weeks, the next few months?

Many corporate decisions that until February this year were made based on predictive artificial intelligence models are now made subjectively by people, as was done before the ‘big data’ era.

Is returning to pure intuition the solution? Moment … We can talk about a new reality, but humans are still irrational, biased and even blind, that is, intrinsically limited to make decisions. Quantitative models, with all their limitations, can be great companions for making good decisions, if we know how to use them.


Points covered in this article include:

  • The data scientist
  • Artificial intelligence
  • The decision scientist


Read the full article, The Sexiest Profession Post COVID, on Christophe’s website.


Umbrex is pleased to welcome Mike Mortensen with Tandem Analytics. Mike has advised business leaders at the intersection of strategy and technology for more than a decade. His experience in business transformation comes from three perspectives: business strategy at McKinsey & Company, machine learning implementation at IBM, and as a business executive responsible for growth at a global conglomerate.

Prior to Tandem Analytics, Mike led teams of business consultants and data scientists to support IBM clients in developing analytics and AI transformation programs. He partnered with clients from concept to realization, including algorithm development, pilot program design, technical integration support, and overall program management. Mike has advised business leaders on AI and analytics strategy across industries, including telecommunications, finance, industrials, retail, and health care.

Before joining IBM, Mike was a Director of Strategy & Innovation at Wolters Kluwer, where he led digital transformation for a B2B portfolio company. To improve customer centricity, he launched a portfolio of initiatives, including machine learning for segmentation and customer behavior insights. Early successes with pricing strategy and customer segmentation fueled transformation efforts across digital marketing, as well as increased personalization of sales and service.

Data scientist and psychologist Tobias Baer (who recently published a book on algorithmic bias) is giving a talk on how to prevent algorithmic bias in the U.K. on Tuesday, 11 February 2020. 

Algorithmic bias can affect us everywhere, from minor trivia such as our social media feeds to critical decisions where bias can wreak havoc with a person’s life dream or a company’s survival. Sources of algrorithmic bias are manifold – some, such as biased data and overfitting, sit squarely in the domain of data scientists themselves, while others only can be tackled by the business users and government agencies who use algorithms, be it through carefully crafted experiments that generate truly unbiased data or through deliberate tweaks of the decision-making process.

 The discussion will include:

  1. The psychological and statistical sources of bias
  2. What business users and data scientists can do respectively to manage and prevent algorithmic bias.
  3. How regulators should think about algorithmic bias

To learn more about the event, visit: