# Build tables and analysis

> Drag variables to rows and columns; the cross-tabulation computes instantly. Layer on measures, weights, significance tests, and expressions to refine it — and use formula cells for bespoke calculations.

*Source:* https://recense.ai/docs/tables-and-analysis

## Building a table

1. **Switch to the Canvas view.** — The canvas is the main analysis workspace.
2. **Create a new table from the toolbar or by dragging a variable onto the canvas.** — Either method starts a new cross-tabulation.
3. **Drag variables to the Rows and Columns drop zones.** — Variables appear as dimensions of the table.
4. **The cross-tabulation computes instantly.** — Results appear as you drop — all computation happens in your browser.

## Measures

Use the Measures pill to choose what each cell shows. You can show multiple statistics at once — for example, counts and column percentages in different table bands.

- **Count** — Weighted or unweighted frequency.
- **Column %** — Percentage within each column.
- **Row %** — Percentage within each row.
- **Index** — Over/under-representation vs. the total.

## Weighting

- Select a weight variable from the Weight pill.
- All measures recalculate using the weighted base.
- The unweighted base row shows the actual sample size for context.
- Toggle between weighted and unweighted views.

## Filters

- Apply a filter variable to restrict the base population.
- Filters affect all measures in the table.
- Combine filter conditions using expressions.

## Significance testing

When enabled via the Significance pill:

- Column letter annotations appear on cells that differ significantly from other columns.
- Tests available: chi-square, z-test, t-test.
- Bonferroni correction available for multiple comparisons.
- Significant values highlight in blue (the one accent colour in Recense).

> **ℹ How to read significance letters**  
> Each column is assigned a letter (A, B, C...). A cell showing "AB" means that value is significantly higher than the values in columns A and B.

## Display options

Control how a variable's categories appear in the table:

- All categories — every category gets its own row or column.
- Top 2 Box / Bottom 2 Box — combine the top or bottom scale points into a single summary.
- NET — aggregate all items into one summary row.
- By dimension — show one entry per part of a multi-part variable (e.g. each item in a grid question).

## Expressions

For derived dimensions beyond what drag-and-drop provides:

- Operators: OR, AND, NOT, parentheses.
- Helpers: TOP_BOX(), BOTTOM_BOX(), NET(), MEAN(), COUNT().
- Variable references: by name or group.
- Implicit expansion: referencing a group expands all its member variables.

*Expression example:*

```text
NET(Q5_1 OR Q5_2 OR Q5_3)
```

## Sorting

- Sort rows by any column's values (ascending or descending).
- Sorting respects the current measure — sort by percentage, count, or index.

## Hiding categories

- Hide individual row or column categories without removing them.
- Hidden categories still contribute to totals and bases.

## Calculated rows and columns

You can add new rows or columns that are calculated from existing ones — useful for showing differences, ratios, or other comparisons that aren't in the raw data.

- Difference: show the gap between two categories (e.g. Male minus Female).
- Index: show the ratio between categories (e.g. how one segment compares to the total).
- Calculated items appear alongside the original data in the table.

## Formula cells

- Double-click an empty cell to enter a formula.
- Formulas can reference cells in the same table or across tables by name.
- Cell-reference mode lets you click to select target cells.
- Formulas recompute automatically when source tables change.

## Canvas workspace

- Arrange multiple tables on an infinite canvas.
- Add markdown notes and anchor them to specific cells.
- Group tables visually for organisation.
- Tag tables for filtering and search.
- Use presentation mode for a clean walkthrough view.

## Next steps

- **[Chart and share results](/docs/charts-and-sharing)** — Visualise your tables and publish a shareable canvas.
- **[Methodology](/docs/methodology)** — Which significance test runs when, and how weighting affects bases.
- **[Agent recipes](/docs/agent-recipes)** — Brief the agent to build, refine, and explain your tables.