As organizations implement processes that harness the power of analytical tools to extract insight from their data, it’s imperative that best practices are followed to ensure those processes result in smart decisions being made for the business.
An analyst can come to a faulty conclusion, for instance, that could result in huge losses for an organization by failing to take into account the characteristics of a dataset. For example, from 1999 to 2009, the “number of people who drowned by falling into a swimming pool” might roughly correlate to the “number of films that Nicolas Cage appeared in.” An uninformed analyst may see this data and determine that swimming pool deaths can be eliminated by reducing the number of Nicolas Cage films. Thankfully for fans of National Treasure, that’s not a solution to the problem. And although that’s an extreme example of the phenomenon, it does demonstrate that even though a relationship may be found when comparing datasets, it doesn’t necessarily mean there’s a causal relationship between the variables.
To avoid these types of pitfalls, one needs to take a step back and evaluate the data in two stages: 1) what data is being accumulated, and 2) what conclusions can be made based on the analysis. Optimally, the first stage should be performed early on, before substantial work is performed with the data. Identifying inconsistencies at this stage can help prevent having to restart the analysis down the line.
The second stage needs to be a continual process throughout the analysis. Not only does one need to ask “What does this mean?” but also, “Does this make sense?” So in the example noted above, although the analysis may answer the first question, it takes asking the second question to determine that a conclusion shouldn’t be made based on the analysis.
By using these best practices, organizations can efficiently and effectively evaluate their data and deliver accurate and appropriate insights.
For more information on Schneider Downs’ data analytic service capabilities, contact Joel Rosenthal at 412.697.5387 or email@example.com, or Andrew Trettel at 412.697.5436 or firstname.lastname@example.org.