International Institute for Analytics

It’s No Game to Find and Keep Your Data Scientists - EA Battles The Market Forces for Talent

In 2013/14, EA’s voluntary turnover among data scientists was 21-22%. It is now 8%, with consistent improvements. These improvements occurred without major changes in compensation and without disproportionate change in investment in the analytics platform, which are common data scientist complaints.

Employment and Training in The Era of AI

As AI replaces some jobs and changes others, it raises questions of, “What is the role for humans in the AI world?” It is most useful to see humans and AI working together, taking advantage of the strengths of each.

The training and learning tracks will vary by role. Foundational learning will be required in all technical roles including basic software engineering, data wrangling, predictive analytics, and data visualization. Data scientists will require additional training in advanced machine learning; data engineers will require more immersion in distributed computing.

The demand for data scientists and analysts is estimated at 140,000 to 190,000. But the demand for data-savvy managers is even greater at 1.5 million. It is unlikely universities will be able to meet this demand. Universities tend to be more theoretical and less focused on practical application. Private training will be needed to fill the gaps.

Mixed Reality and Analytics

Mixed reality (XR) technology is providing quantifiable business value through multiple features and benefits, which include shared 3D context, spatial mapping, data visualization, and much more. Companies are deploying XR across multiple uses cases—and many of these use cases require the use of analytics to analyze and gain insights from massive amounts of information.

AI Hits The Barrier of Meaning

Hype about AI is not new. In 1965 experts predicted that by 1985, AI would do anything humans could. Today, many are optimistic about AI while others want to put the brakes on. So how close are we to human-level AI?

Today, the most common form of AI is deep neural networks, which can do impressive things like object detection and tracking. Classification errors have gone down and are now only 3%, while detection has improved. Combining vision with language enables systems to identify a picture and generate a caption, often with impressive results. Speech recognition and translation have improved as has the ability of machines to answer questions. Machines have shown improved reading comprehension and the ability to play video games. This progress is why some feel AI is closer to human-level intelligence.

Nike Gets Up and Running With Machine Learning and AI

Embarking on an AI journey starts with executive leadership and strategic vision. It requires alignment of the culture and capabilities. At Nike, the key elements have been business leaders wanting to be data driven, demanding deeper information, and being committed to enabling the organization.

The first step in getting up and running at Nike was unification of data science activities. For example, previous efforts were in silos by product group, such as a running app. This resulted in consumers having multiple Nike digital IDs. Having one ID per person was essential. Also important was unification of reporting so everyone at Nike was looking at the same numbers.

The Black Box: Interpretability, Reproducibility, and Responsibility

Historically, a model produced a result that was interpreted by a person who made a decision. In recent years, as the amount of data and number of decisions have grown, agency has been taken from humans and given to machines, which make decisions in a black box. Black boxes raise issues around explainability (or interpretability)—being able to explain how a decision was made—and reproducibility —being able to use the same data and model to make an identical decision.

The reality is that being able to explain complex decisions is extremely difficult, and may not be necessary. And, being able to reproduce decisions is also very challenging, as data, tools, software, models, and environments change. Any single change can have a ripple effect that changes everything. The real issues are trust, reliability, and repeatability, particularly in high-stakes decisions. Building trust starts with IT policies, governance, and infrastructure, to enable preserving history and allow for understanding and reproducing decisions. This is the key to gaining trust and scaling analytics.

Information Economy Mapping

Every organization has a naturally occurring information economy. The rules of other economies hold: there is supply and demand, supplier and buyer power, competitive alternatives, infrastructure, regulation, taxation, and more. Two important rules of thumb: demand always finds a way to get its needs met, and there are legitimate, necessary restrictions on freedom.

Roughly 80% of organizations have a Soviet-style, state-controlled information economy. In the other 20% it is a laissez-faire, demand-style information economy. In every organization it is important to know where you are and where you want to go. The right answer is always something other than a command economy or an unfettered laissez faire economy. It is analytics professionals’ job to figure out the optimal state by balancing those requirements and brokering solutions that are palatable to all.

Building The Analytics Factory at Deere

As a 180-year-old company with 65,000 employees in 30 countries, Deere is the stark opposite of a digital native. Incorporating analytics into different parts of the company has required significant transformation on both the technical and people sides. But all changes have been grounded in the company’s foundational values.

Transformation has required partnerships between the analytics function and other stakeholders, including IT, manufacturing, sales, legal, and more. Partnerships and flexibility have been necessary in reworking traditional processes to become faster and more iterative, and in revising governance and decision making.

Creating A Data Engineering Culture

Data scientists get the glory, but when they experience success it is due to a data engineering culture. In a data engineering culture, the value and importance of data engineering are recognized throughout the organization.

Agile for the Data Science Team

Agile is a methodology and a way of working, originally created to improve the speed and results of software development. It emphasizes individuals and interactions over processes and tools; working software over documentation; and collaboration with customers and responsiveness to change. Benefits include greater predictability, adaptability, transparency, and accountability. Agile is known for working in “scrums”—which are teams with well-defined roles. Scrums engage in “sprints,” which are specified work in a set time period.