The work of data scientists and analysts isn’t much different than that of historians. Such a scholar collects information on a topic of interest — including people, places and events — and synthesizes those points into a unified narrative.
The story then lends insight to cause and effect, change over time, and, ultimately, what the past can tell us about our present and future.
A similar approach is used when analyzing data. Advanced statistics and clever visualization packages —along with your staple math operations — are used to organize billions of data points into a cohesive story.
A job done well should illuminate the following:
- Trends
- Correlations
- Potential cause and effect relationships (verified by testing, testing, and more testing)
- Predictions
When developing these narratives, it’s important to consider what exactly you’re trying to communicate. Which, in turn, will dictate the kind of story you tell. Each element enables brands to interact with their customers more effectively.
Below are four general types of stories data can tell, originally identified in a list of ten by a senior advisor to Deloitte Analytics.
- Reporting – Stories about the past that use descriptive analytics to detail what happened last week, month, or year. This information tells you how you’ve been doing.
- Explanatory survey – Stories about the present that utilize surveys to gauge what’s happening. Survey analysis can help identify potential cause and effect relationships.
- Predictions – Stories about the future. Using data from the past, statistical models are built to predict what may occur in the future. Generally, there’s some presumption that the future will mirror the past. Predictions are assigned probabilities of how likely they are to happen. Refining your model and the predictions it makes follows a cyclical process: model -> prediction -> testing -> data -> model -> review of your assumptions.
- Eureka! – These stories tend to be the longest and most difficult, but also the most fun. They are long, winding down-the-rabbit-hole investigations into complex problems. Naturally, they also tend to be the most inefficient investigations. It’s important to strike a balance between satisfying your inner nerd and ensuring you stay on track to produce something of value for your clients. Even if nothing particularly compelling arises, you learn a lot from the journey. However, if resolved, these stories can lead to significant returns for your company and that “Eureka!” moment.
Determining which of these four story types your analysis speaks to can go a long way in creating a compelling narrative for your clients.
