The Journal The Authority on Global Business in Japan

As a leader of the Information, Communications, and Technology (ICT) Committee, I participate in many Chamber events and enjoy meeting members and guests from a variety of industries. On these occasions, I am often asked certain questions about data science, so I am pleased to have the opportunity in this issue to answer some of them.

I am often asked: What does a data scientist do? While the mix of tasks varies on any given day among companies and industries, the key elements are:

  • Using and developing machine- learning algorithms to understand data faster and in more detail
  • Applying statistical methods to frame business questions in ways the available data can answer
  • Data exploration to find new opportunities and insights
  • Designing data architecture to collect the right information and access it when needed
  • Communicating results through data visualization

Since this is a broad range of duties across intersecting fields that require technical knowledge, the next question I am frequently asked is: Where can I find a data scientist? This can be challenging, especially because the field has only recently appeared as a formal discipline, and most data scientists have backgrounds in other fields and different job titles.

Luckily, as the value of data science is leveraged across industries, many tools are emerging that make the job easier. The widespread use of application program interfaces (APIs) and flexible programming languages supports collaboration between data scientists and software developers. They simplify combining different types of data and technologies. New tools include software to enhance data visualization, packaged solutions with large catalogs of machine learning algorithms and recommended settings, and frameworks or services for managing large volumes of data.

The concept of Big Data has become a buzzword around the world, but is not always well defined. In fact, many people wonder: Is Big Data still important? Big Data is not the only resource in data science. Tiny data can be just as powerful when the right data are identified and effectively combined.

More important than volume is collecting data before they are needed. As my piano teacher would say about practice: “It is not the sudden gush of water in the morning that stains the porcelain; rather it is the steady drip, drip, drip.” Data collected over time can show useful trends.

Looking at trends in the field, Which industries will benefit from data science? In Japan and globally, some applications are well known, such as fraud detection, pricing, and product recommendations, while others are still in the early stages of development. Rapidly emerging applications like language processing and image recognition have broad utility. Working now in real estate and retail, I see a big potential for any industry that involves risk or decision making.

Data science provides both methods to assess risk and the opportunity to reduce it, by measuring similarity and finding things that are different enough to build non-obvious recommendations and robust portfolios.

Decision making is supported by classification algorithms. By systematically describing and learning from past results, every success or failure becomes a data point to improve future decisions. After training a model to recognize the patterns, the same process can be used to quickly make one decision or many, where humans would need to think through each individually.
In short, like ICT in general, all industries can benefit. I hope we will see breakthrough innovations in Japan.

Try to distinguish Big Data terminology from Pokémon at

Dr. Imai Jen-La Plante is vice-chair of the ACCJ Information, Communications, and Technology Committee and data scientist at Locarise K.K.
The concept of Big Data has become a buzzword around the world, but is not always well defined.