Take your data beyond visualization - Pyramid Analytics

Visualizations are effective tools to simplify complicated business challenges. They can turn data into meaningful and actionable information, revealing previously obscured data trapped in tables and grids. Data visualizations have transformative power.

Modern analytics and BI software tools have embraced the idea: "See and understand your data! Visualize your data instantly!" Beautiful charts and dashboards are prominent features of these tools. The power of data visualization in their brands and products is undeniable.

However, when addressing real problems, issues arise. Data may be out of date or incomplete, and access often depends on others. Multiple tools may be required for modeling and analysis, making the process cumbersome. Despite the potential of data visualizations, progress can be stalled before it even begins.

Look before you show

In The Back of the Napkin, business author Dan Roam explains how humans are prewired to understand visual information. Even the most difficult business problems can be simplified using drawings, charts, and visualizations. Roam breaks the visual thinking process into four steps:

1. 1.Look – Collecting and screening information
2. 2.See – Selecting and clumping information
3. 3.Imagine – Seeing what isn’t there
4. 4.Show – Making it all clear

Roam notes that many struggle to create effective visualizations because they start with the “show” step, before truly looking, seeing, and imagining. He writes:

“This tendency to equate visual thinking with the creation of elaborate and refined drawings is just plain wrong. It approaches the process of visual thinking backward, limiting our most powerful problem-solving ability before we’ve even had a chance to really use it. That’s because showing... happens at the end of the visual thinking process, not the beginning.”

This is a common trap when using analytic tools. The visualization is seen as the artifact that will reveal everything, but legacy tools and point solutions often omit vital components. They make it difficult to collect and screen information, and their discovery toolsets are limited. While they can produce glossy charts and dynamic views, the underlying data may only tell part of the story. This problem is amplified within organizations.

Collaborative analytics: sender versus receiver

Analytics-based decision-making is most effective when collaborative. In collaborative problem-solving, individuals play different roles depending on their place in the process: preparer/sender (analyst) or receiver/audience (business user/decision maker).

If you are the preparer/sender, you cannot start with a visualization. You must begin with the data and follow a systematic, iterative process to extract meaningful insights. This requires access to all data sources, as well as modeling and discovery tools. Context must be added to fully communicate findings.

If you are the receiver, you may start by examining a chart or dashboard. But this is only the beginning. Context is crucial: What am I looking at? Does this answer my questions? Is the underlying data trustworthy? How does this relate to other data views?

From either perspective, analytics must go beyond visualization.

If tools reduce analytics to just visualization, the process is hindered. Many tools neglect critical aspects such as data access, governance, and scalability. Without these, chaos ensues: departments create views on limited datasets, and the "winner" is the one with the prettiest chart.

While you may gain some visibility into business challenges, analysis is often incomplete or invalid because BI tools can only produce limited views based on incomplete data. Most BI tools have not solved the problem of data siloes. This is why many visualizations fail to aid decision-making—they are based on outdated or incomplete data.

However, if visualizations are treated as a component of the analytics process—the result of thoughtful and systematic analysis—then the larger analytics process is honored. Building visualizations with a complete platform ensures that thoughtful decision-making, not just visualization, becomes the culmination of the analytics process.