Use this panel to build pivot tables and charts from your project data with a simple drag‑and‑drop interface (powered by PivotTable.js).

Quick start#

1. Choose a data source: Links, Factors (Work in progress!), or Sources.
2. Choose the stage: After Analysis Pipeline (matches your filters elsewhere) or Before Analysis Pipeline (raw data).
3. Pick a display: Table, Heatmap, or a Plotly chart.

Arrange fields (drag and drop)#

• Drag field chips from the left list into the drop zones:
• Rows: categories listed down the left.
• Columns: categories across the top.
• Values / Measure: the numeric field used for calculations (when needed).
• Reorder by dragging within a zone; remove by dragging out or clicking ×.

Choose the calculation ("Summarize by")#

• Use the aggregator dropdown to select how numbers are calculated:
• Count (default) – how many rows fall into each cell.
• Sum of / Average of / Min / Max – pick a numeric field (e.g. ai_confidence, text length, counts).
• Unique Count – how many distinct values of a field occur.
• Example: to average AI confidence by region and sentiment 1) Put custom_-Region in Rows and sentiment in Columns 2) Set aggregator to Average of and choose ai_confidence as the field

Filter or exclude values#

• Every field chip has a filter menu. Click it to include/exclude values.
• Use the search box, then tick/untick items. "Select All" toggles everything.
• Filters apply immediately and are remembered in the URL so you can share the view.

Sort and tidy#

• Click any row/column header to sort by label or by totals.
• Subtotals and grand totals are included automatically in table views.
• Remove empty columns quickly by excluding the blank value in that field's filter.

Heatmaps and charts#

• Switch the renderer to Heatmap, Row Heatmap, or Col Heatmap for quick intensity views (colors match the app palette).
• Choose Plotly renderers for interactive Bar, Line, Scatter, Stacked Bar, Area, or Multiple Pie charts.

Export and sharing#

• Use the toolbar to Copy to Clipboard or Download XLSX of the current table.
• The current configuration is saved in the URL automatically; bookmark or share it with collaborators.

Notes on the datasets#

• Links: every causal link plus metadata; includes AI fields (e.g. confidence) and reserved columns like original_cause, original_effect.
• Factors: unique factors with frequency, source count, citations, and original_label (ALL underlying original labels for the displayed factor, concatenated with line breaks, derived from the current stage's links like original_cause/effect).
• Sources: document metadata and flattened custom fields (custom_*).

💡Tip: For results that match other panels, use After Analysis Pipeline.