Labeling
The Labeling Node makes it possible to add one or more additional columns to the data and manually fill those with new data.
Labeling differs from most other nodes in that it does not compute data but uses data set by the user, i.e. you. Every labeling action you trigger adds to or changes the already existing data without completely recomputing the output. The settings control of the labeling node does not display settings that are automatically applied to the node upon recomputation, instead it offers controls for the labeling operations.
Output
Labeling has one output that outputs all the data available at its input, plus as many columns as the you created in this node.
Aggregation
The labeling node is not able to aggregate, as all data at its input is also forwarded to its output.
Settings
- Columns:
- The topmost area of the settings displays the columns created by the labeling node and the currently selected target column. You can add new columns, and change their name and data type. Once a labeling operation has been performed on a column, its name and data type become fixed and cannot be changed any more. Exactly one column is always selected as a current target. All labeling operations will be performed on that column.
- Target Description:
- This is an optional description for a column. If you write something into this text field, the description of the current target column will be added (or replaced) the next time you start a labeling operation. These column descriptions can be exported via the data export.
- Target Marking:
- The target marking determines, which area of a column will be labeled by the next labeling operation. Most of the time the target marking will be the current marking. Note that in many cases Couple Cursors should be set to false on the Common Settings tab. In addition to the current marking, any marking that is set on either the input or the output data of the labeling node will work, making it possible to e.g. label all rows found by a value search. A labeling operation will set all rows of the target column selected by the target marking to the label value.
- Quick Labels:
- The quick labels are a number of labels you can set and then use via a button or via keyboard shortcuts in Keyboard Mode. Quick Labels are particularly helpful, if a column is to be labeled with a limited number of lengthy labels: They can be used as labels with a single keystroke instead of entering them repeatedly.
The labeling node has three modes of operation:
- Standard Mode
- Labeling is only possible with the buttons next to the quick labels. This is the default mode, but only because the keyboard grabbing of the other modes can interfere with the expected behavior of other parts of Blickshift Analytics.
- Quick Label Mode
- Allows labeling with the keyboard shortcuts displayed next to each quick label. The number keys can be used to change the target column or (Ctrl+Number) the target marking. Additionally, pressing Return will pop up a small window that allows entering labels not present in the list of quick labels. Pressing Ctrl+Return will switch to Immediate Labeling Mode.
- Immediate Labeling Mode
- Allows directly entering data into the marked areas. A labeling operation is finished by simply selecting another area. Pressing Esc will return to the previously active mode.