Formula

Add new features to your dataset.

Inputs

• Data: input dataset

Outputs

• Data: dataset with additional features

Formula allows computing new columns by combining the existing ones with a user-defined expression. The resulting column can be categorical, numerical, textual, or datetime. For numeric variables, it sufices to provide a name and an expression.

1. Add new variable.
2. Remove selected variable.
3. New variable name.
4. Expression in Python.
5. If checked, the option will place selected variable in metas.
6. Select a feature.
7. Select a function.
8. Assign values to categorical variables.
9. A list of new variables.
10. Press Send to communicate changes.

Example

Here is a short example using the iris data set from the File widget. We constructed three new variables, one numeric, one categorical, and one textual. For the numeric variable, we computed a square of petal length using the expression petal_length**2. For the categorical variable, we mapped sepal length to three new categories using the expression 0 if sepal_length < 6 else 1 if sepal_length < 7 else 2. We also mapped the newly-created bins to values 0, 1, and 2. Finally, for the text variable, we removed the redundant iris label from class values, retaining only the species name using the expression iris.split("-")[1].

We can observe the changes in a Data Table widget.

Python math language

If you are unfamiliar with Python math language, here's a quick introduction. Expressions can use the following operators:

• +, -, *, /: addition, subtraction, multiplication, division
• //: integer division
• %: remainder after integer division
• **: exponentiation (for square root square by 0.5)
• <, >, <=, >= less than, greater than, less or equal, greater or equal
• == equal
• != not equal
• if-else: value if condition else other-value (see the above example)

See more here.