Data thinking: Design Thinking for Data Science

Written by Rizwan Asif
Account Manager

Becoming data-driven is challenging, albeit imperative. An organization can be inundated with data without any competitive advantage. The challenge is not to put data in use, but to leverage data without hurting the core business.

Gathering, storing, and managing data is becoming easier. Organizations find themselves with silos of data with momentary perceived value. This perception manifests as a crises or urgent need of information; for example, “costs are rising, list the services we haven’t used”; “we need more income; check the inactive clients from last year”; “material quality is not as good; who’s been our best supplier?”. Numbers dusted, crunched, and siloed.

Otherwise, continuous integration and utilization of this data is expensive and does not guarantee success. Specifically in large distributed and federated organizations, data is setup in smaller databases (silos) which makes coordination difficult. Moreover, according to this report, 1 out of 10 data science projects do not make it past the pilot phase in Finland alone. Thus, a dilemma for business leaders if being data-driven is even worth it.

Data thinking culture

Being data driven is not about adopting tools but nourishing data thinking culture. Ray Dalio in his book Principles, suggests that successful business decisions are meritocratic rather than democratic. Therefore, data-centric decisions must have merit in a data thinking culture.

At Huld, we understand this is easier said than done. A call for cultural shift should have roots in numbers rather than hype. Additionally, an organization wide movement takes time and effort. Our data (design) puzzle methodology breaks down the stages required to meet your data driven organization goals. A plan is formulated based on the strategic goals, complexity, cost, and magnitude of benefits.

Building data thinking teams

Investments in a data thinking culture must be driven by a progressive growth of personnel. An effective data science team may consist of the following roles.

  • Data Scientists
  • Machine Learning Engineers
  • Data Engineers
  • Data Visualization Engineers
  • Software Developers
  • Analytics Translators
  • Data Analysts
  • Data Hackers

Building such a team is a gradual task; our data puzzle manages your team size by prioritizing feasibilities and impacts. Team members are added quarterly or yearly based on progress.

The data puzzle consists of a few workshops conducted remotely or face-to-face. We deliver a comprehensive data thinking roadmap divided into stages, with measurable impact of each stage.

Data thinking at Huld

A misconception is that data thinking relates to only computer science. However, domain expertise and mathematics are equally components. Huld having a team of exceptional people from diverse domains and a vast industrial clientele; we consolidated our efforts in applying design thinking to organization-wide data problems.