Step 2: Summarize the data
In this step, you build a DataBrew recipe—a set of transformations that can be applied to this dataset and others like it. When the recipe is complete, you publish it so that it's available for use.
In the game of chess, players can be rated based on how well they perform against
other players. (For more information, see https://en.wikipedia.org/wiki/Chess_rating_system
To summarize the data
-
On the transformation toolbar, choose Filter, By Condition, Greater than or equal to.
-
Set these options as follows:
-
Source column -
white_rating -
Filter condition – Greater than or equal to 1800
To see how the transform works, choose Preview changes. Then choose Apply.
-
-
Repeat the previous step, but this time set Source column to
black_rating. After you apply your changes, the sample data contains only those games where the players on each side (black and white) were Class A or above. -
Summarize the data to determine how many games were won by each side. To do this, on the transformation toolbar, choose Group.
-
For the Group properties, do the following:
-
In the first row, choose
winnerfor Column name. Leave Aggregate set to Group by. -
In the second row, choose
victory_statusfor the Column name. Leave Aggregate set to Group by. -
Choose Add another column.
-
In the third row, choose
winnerfor Column name. Set Aggregate to Count. -
For Group type, choose Group as new table. The preview pane shows you what the result will look like.
-
Choose Finish.
-
-
Choose Publish to save your work, at right on the recipe pane.
-
For Version Description, enter First version of my recipe. Then choose Publish.
