This blog was offered by Joe Yacura, CEO and founder of the International Association of Data Quality, Governance, and Analytics.
Today’s supply chains are being re-engineered aggressively as a result of the COVID pandemic impacts on business. Decisions need to be made at speeds previously not every seen before. Humans need to be able to rapidly assess vast amounts of data, and quickly render an actionable decision. Data visualization has made significant progress in recent years in helping with this problem. Managers used to have to look at matrices of data typically limited to two (2) dimensions at a time, to try to gain insights as to what had transpired or estimates of what might occur, then make an educated guess about what to do.
Visualization soon became the norm, but they were usually limited in the information they could convey through the use of relatively basic graphics. As one analytics expert said at a conference- “can we do more than simple pie charts?” While these basic visualizations of the data helped to convert these matrices into a new level of interpretation/information and insight, there was still a lot to be desired.
Today, well defined and easy to navigate user interfaces on several visualization tools such as Tableau, Klik, and PowerBI have led to a rapid adoption of these tools. The power of these tools has also increased dramatically in their ability to represent data in multiple dimensions simultaneously. This new level of data visualization of multiple dimensions of information greatly assists humans in their decision making process.
But wait! It also leads to the next set of questions, related to how humans and machines engage and interact once these visualizations are produced in real time. These questions are profound and serve as an important set of guidelines that lead to multiple organizational design and talent management implications.
- To what extent do we want human intervention in decision making in the future?
- Can tactical, operational decisions, be made autonomously by advanced machine learning?
- Should human intervention in decision making be migrated more towards strategic decisions making given the decision latency associated with human decision making?
- Given the rapid changes in today’s world, can we afford to wait for human decisions and still be competitive?
- Is data visualization just a transitional decision aid until we gain greater trust and confidence in the data used to make decisions?
- Should we continue to commit resources to advance data visualization in our companies or should we transition directly to advanced/autonomous data driven decision making?
- Is data visualization a transitional decision aid that needs to be viewed and assessed from a long-term strategic solution perspective?
Advances in AI, ML and inference engine technologies have greatly enhanced autonomous data drive decision making utilizing vast amounts of structured and unstructured data in real time. Todays’ systems can analyze millions of data points on an unlimited number of dimensions and either conform to previously defined decision rules and/or learn from the data in either a supervised or unsupervised learning environment. The conversion and representation of this information into visualization for human consumption seems to be of diminishing value based on its cost and time consumption associated with presenting this visualization to humans for review to assist in a decision.
While some decisions may still require human intervention, most operational decisions can be machine derived eliminating the need for visualization. There is no right or wrong answer, but these questions should be a starting point for discussions and debates in organizations seeking digital transformation.