Chapter 13 Data Moves Overview

Over the years, some colleagues and I have watched students (and ourselves) do interesting, rich data tasks. From those observations, we have come to believe in “data moves”: common, identifiable actions that alter a dataset in important ways— enhancing it, reorganizing it, nudging it into a form where it can tell its story. We even wrote a paper about data moves. We see that these moves, while simple on the surface, contain great power for those that use them well— and spell trouble for students who don’t understand them. Furthermore, perhaps because they seem simple, and because they are largely unnecessary in a traditional statistics class, they don’t get taught.

Some students (perhaps the ones who later become statistics teachers) just naturally pick them up through exposure to data…but others never do. Maybe, if we pay attention to data moves in our teaching and curriculum design, more students will be successful, learn about data and data analysis, and bring much-needed perspectives to the field.

Even in this short introduction to data science, students use the most important (“core”) data moves; not only that, they can identify the data moves they are using. We have tried to identify data-move moments in the lessons and commentary.

In the next few chapters, we’ll elaborate a bit more on the data moves that appear in this book, how they fit together, and, importantly, how you actually do them in CODAP. The blue section headings below are links to the corresponding chapters.

Filtering

Filtering means looking at a subset of the data. Choosing the right subset can crack a tough analysis right open. Good filters can also give you insight into what to do next.

We have put some comments on selection in this chapter as well, because filtering in CODAP requires selection.

Grouping

Putting data into groups is a fundamental part of data analysis. We often group in order to compare one group to another. But it turns out that grouping has other functions as well, for example, to facilitate other computations.

Calculating

When you don’t have an attribute that’s exactly what you want, sometimes you can make it out of attributes you already have. This often involves calculation (for example, in a units conversion) but not always.

Summarizing

Summarizing, in this context, means characterizing a group. A clear example is computing the mean height of a group of people. That mean height summarizes the group, and helps us compare that group to some other group. Sometimes, people call that value a measure or an aggregate measure of the group.

Joining

When you need to use data from two or more datasets as part of the same analysis, you have to join them together. This is a more advanced move, but some students have ideas they want to pursue that require joining.

And two more

As I suggested in the introduction, data moves are not all there is to data science. You’ll find two more chapters bundled with the data moves listed above, in their own chapters:

Preparing and cleaning data

In this book, students use data we provide, or data they choose from data portals. When you bring in your own data, it’s often “unruly.” That requires a whole new set of skills.

Visualization and communication

This is a giant topic to which we cannot do justice. Still, there are a few things that need to be said.