Overview
The Binning node is a data preparation node in a Weave workflow that distributes a continuous numeric column’s values across a configurable set of bins (between 1 and 20), producing a categorical feature for the column. For each selected column, the node appends a new column named <column>_bins to the output, containing each row’s bin number. Original column values are preserved alongside the new bin column — Binning is additive, not destructive.
Binning nodes are added from the Data Preparation category of the node panel. Configuration is two-step: select columns via the column-selection popup, then enter the bin count in the Choose Number of Bins field. One bin count applies to all selected columns within a single node — for different bin counts per column, chain multiple Binning nodes.
Note: “Weave” is the current name for what earlier interface versions called “Workflow.” This reference uses “Weave.”
When to use it
- Converting a continuous numeric column into a categorical feature for downstream categorical analysis, segmentation, or visualization.
- Preparing data for machine-learning models that require discretized features (decision trees, rule-based systems, certain classifiers).
- Simplifying high-cardinality numeric columns into a small number of analyzable groups.
- Generating bin labels that can serve as categorical Dimensions in Lens Reports — where the continuous column wouldn’t directly fit a categorical Dimension role.
Configuration
Column selection
| Setting | Description |
| Select (button) | Opens the Binning column-selection popup. |
| Select All | Checks every column. Use to bin all columns at once; non-numeric columns may produce errors. |
| Individual column checkboxes | Pick specific columns to bin. Unselected columns pass through unchanged. |
| Apply | Commits the column selection. |
Choose Number of Bins
| Setting | Description |
| Enter Binning value | Single integer between 1 and 20 inclusive. The field validates this with the error “Enter valid input (1-20)!” for out-of-range values. The same bin count applies to all selected columns. |
| Allowed range | 1 to 20. |
Output schema
Binning’s effect on schema is its most distinctive structural fact. The node adds columns rather than transforming existing ones.
| Aspect | Behavior |
| Original columns | Preserved unchanged. All upstream columns flow through to downstream output. |
| Bin columns | One new column per selected column, named <column>_bins. For example, selecting Profit produces a Profit_bins column; selecting Revenue adds Revenue_bins. |
| Bin column data type | Integer. Each row’s value is the bin number — between 1 and the configured bin count. Treated as a categorical feature downstream. |
Key behaviors
Binning produces a categorical feature from a continuous column. The node’s analytical purpose is converting a continuous numeric column (e.g., Profit) into a categorical feature (Profit_bins). The categorical feature has a small fixed set of values — integers from 1 to the configured bin count — that downstream nodes can group, segment, or aggregate by.
Binning is additive, not transformative. Unlike Cleansing (which modifies column values in place), Imputation (which fills nulls in place), Outlier Treatment’s IQR Method (which drops rows), or Z-score Normalization (which transforms column values), Binning adds new columns to the output. Original column values are preserved. This makes Binning safe to apply broadly — downstream nodes have access to both the original continuous column and the new bin column, choosing whichever fits the analytical need.
One bin count per node. The Choose Number of Bins field applies to all selected columns uniformly. To bin Profit into 4 bins and Revenue into 10 bins, use two Binning nodes — one per bin count, each with the appropriate column selection. This is parallel to how Imputation requires multiple nodes for different fill methods and Outlier Treatment requires multiple nodes for different treatment methods.
Original column values are preserved. Downstream Reports, transformations, and visualizations have access to both the original numeric column and the bin column. Use the original for continuous comparisons (averages, sums); use the bin for categorical grouping (count per bin, segmentation analyses).