Numbers and Words: Presenting Data in Understandable Ways
senior data analyst
Math is fun. Think about all of the classic dad jokes math has made possible. For instance, why do teenagers travel in groups of three or five? Because they can’t even. But I know not everyone has the same perspective that I do when it comes to presenting data.
This means we need to think about how we present data to others in our organization. When I look at a data model I am curious to know what the concordance rate is, what the wald-chi square values are, and how the ROC curve looks. But the concordance rate is hard for me to wrap my head around and even harder to explain. Imagine what that means for those that don’t typically work with data. Imagine presenting data to those who don’t work with data.
At our user conference, a number of our customers spoke to this same point. When sharing the results of a predictive model with those that are going to actually use the information, it is important that the end users understand what the data means. We often bog them down with overly complicated validation statistics or explanations. Instead, many customers are opting for an easier route. High. Medium. Low.
By categorizing and often color coding the output of a predictive model you leave less room for misinterpretation. They can quickly look at a potential donor and see that a prospect has a high likelihood, rather than trying to interpret whether or not a probability of 63% is good. It is important to remember the goal is to provide the information to help them do their jobs more easily. What information needs to be provided can only be determined by those using the data, not the ones creating it.
In the end, the goal is to use this information to create a more data-driven culture. It is hard to do that when the end users are left confused and overwhelmed. Simplifying the output can lead to much better results and help impact your bottom line and makes presenting data easier.
To hear more about how strategies like this can help your institution, check out our second webinar in our Women In Data Science series. Abigail Komlenic, Associate Director of Advancement Analytics at Swarthmore College, will share not only how their advancement team shares predictive modeling results, but also how the models were developed and what metrics were used.
To register for Abigail’s webinar, click on the link below.
Assoc. Dir. of Advancement Analytics
Modeling Major Gift Likelihood
Translating Data Into Action
Abigail discusses how her Advancement team translated their predictive model for major gift likelihood into descriptive, easy-to-interpret scores that intrigued colleagues to use them before the project was fully rolled out.