Beware the Data Visualization Trap
May 30, 2017 | By: Drew Lessard | Category: Workforce Analytics
Data visualization tools are increasingly being used by human capital managers and analysts. Making the move from Excel to interactive visualization tools reveals insights and information that was previously trapped in dozens of spreadsheets . The awesome power and value of dashboards and visualization tools is how well they present interactive summaries of your data. These tools are helping human capital professionals manage their workforce and plan for the future.
Visualization tools are great for their intended purpose – visualization. But they are often trumpeted as solutions to all of your data and analysis problems – mostly by their sales teams. “Just add visualization,” they say, “and you’ll suddenly see your data with clarity, revealing insights you never imagined.” But don’t be fooled. Visualization tools are only as sophisticated as the data fed into them.
A visualization of historical data may reveal some insights. But it is no replacement for real analysis. Dragging and dropping and drilling and filtering simply cannot compete with statistical and machine learning algorithms. Humans have a tough time considering even a dozen correlated variables, let alone hundreds or thousands. Data visualization tools are great but don’t be fooled into thinking that they can solve predictive analytics problems.
As an example, consider creating a workforce plan. Visualization tools can help uncover potential hot spots for succession planning when viewing the organization by department to identify which teams have an unusually high average age or turnover. This is a great way to use current data to help make a more informed decision about how to invest recruiting resources in the future.
But visualization tools cannot be used to create a forecast of how the workforce will evolve over time. You could visualize historical data related to retirements, promotions, and attrition. But this is hardly a sophisticated forecast. This is more like a simple “thumb in the air” assumption.
A better way to approach more complex human capital problems is with true predictive analytics – using a data-driven statistical approach that leverages historical data to project the behavior of the workforce over time. Predictive analytics can provide detailed projections for workforce planning down to the location, occupation, and grade level. It can help to provide specific answers to questions, like:
- How close will we get to a reduced workforce target through natural attrition?
- What occupations and pay grades will we have the largest gap in our workforce as the result of a hiring freeze?
- At what level do we need to recruit for most aggressively to account for likely future promotions?
Predictive analytics is like the oracle that provides a glimpse into the future. Adding visualization to predictive analytics can provide insight into results and help to communicate findings. But visualization is a not a replacement for analysis. When you combine these tools, you begin to unlock the real decision-making support that human capital leaders need.