insights.jpg

Are You Using The Right Data Analysis Techniques?

By John Dillard

Businesses may think they must embark on a Big Data mining mission, but the truth is, they already have all they need to do meaningful data analysis.

Getting in touch with your own data is vital to operational movement. There’s an emphasis on cloud, Hadoop and business intelligence, which has given the impression that data analysis requires expensive software. But that’s not the case. You’re able to get started right now.

Every organization has information. It may be in a form that requires you to export it from accounting or other back-office software, but it is there, ready to be harvested with a variety of methods of data analysis. Excel and even Word have tools as simple as search-and-replace or parsing, which makes information available for analysis.

Once you have the data, it’s time to visualize it. Most data analysis is used to extrapolate or find patterns like correlation. Fortunately, the human mind is inherently adept at finding patterns and extrapolating simultaneously. In fact, the ability to find patterns visually is one area in which the brain specifically evolved. This ability was critical to our survival, from the point of origin of the species to walking across a busy city street.

There are a number of data analysis tools to consider. We know how to do bar charts for time series, but if you want to find patterns, try a scatter plot and see what turns up. Doing so could lead to correlation. Does it look like there may be a pattern? You don’t even need to have an in-depth knowledge of statistics to add a trend line. Excel has capabilities that let you measure how well the line fits your data and extrapolates into the future.

Now, it’s time to aggregate. If you don’t see a pattern initially, consider another method of data analysis. Even just filtering and sorting allows patterns to form inside your data. In Excel, experiment with pivot tables and use functions like SUMIF or SUMPRODUCT to look at categories inside data.

Once you’re comfortable with the manipulations you’ve created, cycle back again, visualize and look for correlations. Find ways to play with the data you’re working with. Then test your hypotheses.

Remember: The hardest part is getting started. Once you’re comfortable, there are plenty of free resources for the next step, such as Tableau, which helps you to create a visual presentation and share your data.

When presenting your data, keep the format clean. For instance, avoid pie charts. “[D]eath to pie charts,” Cole Nussbaumer declares in a post on the Storytelling with Data blog . An unintended consequence of using colored pie charts is it creates an optical illusion, with darker slices appearing bigger. “My strong opinion is that color should always be an explicit choice and should be used strategically to draw the audience’s eye,” Nussbaumer writes. “This preattentive power is being wasted here. If you must use a pie chart, at least make the slices the same color and highlight only the one or two you want to draw attention to.”

Also, don’t use clever shapes instead of bars. Using any shape other than a rectangle distorts the data because the apparent area of the shape makes bigger shapes seem to represent bigger numbers. Avoid chart clutter, too. “[I]t does make the numbers themselves more difficult to decipher,” Naomi Robbins writes in a post on Forbes.com.

For the most effective data analysis, think about the quality of the data, consider your assumptions and document your sources because all data should be attributed. It’s also crucial to make sure your sample size is big enough to support the claims you make, and make sure to use current data. You must be able to trust your data, so ensure it’s not too “rolled up,” which leads to incorrect analysis.

Your analysis is going to be used for decision-making, presentation and persuasion, so be prepared to explain where the numbers came from, or make that information clear up front. Think about how confident you are about each link in your logic chain. Doubt grows exponentially if you lack confidence in any one of these links. Using a strong data analysis methodology will ensure your information is always accurate and a good foundation for your business decisions.

Want to strengthen your organization’s data analysis processes? Call (704) 266-0015 or click on the button below to send a message:

Contact a Big Sky Expert