A point-and-click explorer over the coded dataset. Three tables: Companies (one row each), Products (one row per named product), and Evidence (one row per citation). Drag fields to pivot, filter, sort, and chart. Start from a saved view or build your own.
Click Configure to open the side panel. From there, drag any field into Group By, Split By, Order By, or Where to pivot and filter — and use the plugin picker at the top of the panel to swap between Datagrid and chart types (bar, line, scatter, heatmap, treemap).
Reading the confidence columns
The Companies table carries evidence-quality columns next to the values, because in a cited dataset those are part of the data rather than metadata to be stripped at presentation time:
| Column | What it tells you |
|---|---|
enriched |
Whether the pipeline has researched this company yet. false rows are roster seeds. |
category_confidence |
How confident the coder was. low triggers a second verification pass. |
pricing_model_confidence |
Same, for the field most at risk of fabrication. |
unique_ip_confidence |
Same, for the field most likely to attract marketing paraphrase. |
gap_count / gap_fields |
How many fields could not be established, and which ones. |
evidence_count |
How many verbatim citations back this company's record. |
pricing_published |
true only where an actual figure is published. Distinct from a pricing model being known. |
has_unique_ip |
true only where an IP claim survived the evidence requirement. |
Filtering on enriched == true before charting is usually what you want — otherwise unresearched roster rows dilute every distribution.