Data Science on the Google Cloud Platform

Chapter 3. Creating Compelling Dashboards

In Chapter 2, we ingested On-time Performance Data from the US Bureau of Transportation Statistics (BTS) so as to be able to model the arrival delay given various attributes of an airline flight—the purpose of the analysis is to cancel a meeting if the likelihood of the flight arriving within 15 minutes of the scheduled arrival time is less than 70%.

Before we delve into building statistical and machine learning models, it is important to explore the dataset and gain an intuitive understanding of the data—this is called exploratory data analysis, and it’s covered in more detail in Chapter 5. You should always carry out exploratory data analysis for any dataset that will be used as the basis for decision making. In this chapter, though, I talk about a different aspect of depicting data—of depicting data to end users and decision makers so that they can understand the recommendation that you are making. The audience of these visual representations, called dashboards, that we talk about in this chapter are not other data scientists, but are instead the end users. Keep the audience in mind as we go through this chapter, especially if you come from a data science background—the purpose of a dashboard is to explain an existing model, not to develop it. A dashboard is an end-user report that is interactive, tailored to end users, and continually refreshed with new data. See Table 3-1.

Table 3-1. A dashboard is different from exploratory data ...