Build dashboards

A dashboard starts with a data source. Widgets query it; dashboard filters coordinate the view.

Data Sources

A data source is a thin wrapper around a data store table:

Field	Description
Name	A human-readable label shown in the widget builder
Table	The underlying data store table the data source reads from
Detected schema	Fields with their types and which ones are aggregatable
Time dimension	The date field used for dashboard date-range filtering (set during creation, changeable later)
Sync config	Optional — if the data source is populated by a recurring sync workflow, the schedule and last-sync status are surfaced here

Aggregating Multiple Tables

A single data source can aggregate multiple tables across organizations. This is useful when one dashboard needs a combined view, such as call volume across several customer organizations. QuickFlo exposes a virtual _customer dimension for grouping and filtering; each row uses the alias assigned to its source table.

Aggregate sources work for both:

• Cross-org aggregation — your own table plus tables shared from other organizations (see Sharing Tables Across Organizations)
• Same-org multi-table aggregation — your own table combined with other tables in your own org

Shared Data Sources

A data source can also point at just a table shared into your organization. It appears in the picker with a Shared badge and remains read-only. If the owner revokes access, new queries stop and affected widgets show a clear error.

Widgets

Each widget runs an analytics query and renders a chart. Available widget types:

Widget	Best for
Bar	Comparing measures across categories
Line / Area	Trends over time
Pie / Doughnut	Share-of-total breakdowns (use sparingly — bars are usually clearer)
Stat Card	A single headline number (with optional comparison delta)
Funnel	Multi-stage conversion flow
Data Table	Raw rows or grouped detail with sortable columns
Stats Table	A grid of measures × categories — like a multi-column stat card
Pivot Table	Two-dimensional grid with row/column dimensions, ratio mode, and heatmap shading

Query Configuration

Every widget has the same query shape under the hood:

• Measures — what to aggregate (e.g. count, sum(amount), avg(durationSeconds))
• Dimensions — what to group by (e.g. region, direction, agentName)
• Time dimensions — a date field plus a granularity (15min, 30min, hour, day, week, month, year)
• Filters — narrow the dataset before aggregation
• Joins — optionally join in another data source by a key field
• Order — sort the result set by a measure or dimension
• Limit — cap the number of rows returned
• Time range — the date window the query covers (relative last 7 days, or an absolute start/end)

Pivot Tables

The pivot table widget supports:

• Row and column dimensions — including inline date bucketing via field suffixes (createdAt:hourOfDay, createdAt:dayOfWeek, createdAt:month, etc.) so you can pivot “calls by agent × hour of day” without precomputing the bucket
• Ratio mode — switch the cell display from raw measure values to ratios (with percent formatting) for “what fraction of the total” views
• Heatmap shading — color-scale cells from low to high so outliers pop visually

Filters

Dashboards have two layers of filtering:

Dashboard-Level Filters (Auto-Wire)

Click + Filter in the dashboard toolbar to add a global filter that drives the whole dashboard — a region dropdown, a customer selector, an environment switcher. Each filter has:

• A label shown in the toolbar
• Bindings — one or more (data source, dimension) pairs that this filter targets

The bindings array is the magic: a single filter can wire itself to multiple data sources at once by mapping the same logical concept (region) to whichever field name each data source uses for it. Add a “Region” filter, bind it to region on data source A and geography.region_code on data source B, and selecting “EMEA” updates every widget connected to either source. No widget changes needed.

You can save default filter selections so a dashboard always loads with a sensible starting state.

Widget-Level Filters

Each widget can also have its own filters, applied on top of the dashboard’s. These come in two flavours:

• Simple mode — structured rows with field, operator, and value (status equals active, amount greater than 100). Operators are equals, notEquals, contains, notContains, gt (>), gte (>=), lt (<), lte (<=), set (is set), and notSet (is not set). The equals and notEquals operators support multi-value chip selection with searchable options pulled from the underlying data, so you can pick a list of accepted values without typing them by hand.
• Formula mode — a code editor for expressions that combine AND, OR, NOT, comparisons, arithmetic, and supported functions. Use brackets for field names with spaces, such as [Retry Count]. Member access is reserved for data-store references such as $ds.table.key. Use formula mode when simple rows do not express the rule — for example (status == 'active' OR retryCount > 3) AND createdAt > '2024-01-01'.