The Explore query builder in PushMetrics with the data panel on the left, the chart in the middle, and the visualization settings on the right

An Explore is a saved chart built on a Metric Block. You pick the measures and dimensions you want, add filters or a time range, choose a chart type, and PushMetrics compiles your selections into a single SQL query, runs it, and renders the result. No SQL required, though you can always see exactly what ran.

Because an Explore is anchored on a Metric Block, the numbers stay consistent: the same measure means the same thing in every chart.

Creating an Explore

1
On the Content page, click + New → New Explore.
2
Pick the Metric Block you want to explore. This anchors the chart. (You can also Create new a Metric Block right from here.)
The Pick a metrics view to explore screen, listing available Metric Blocks such as car_sales, chinook_artist_revenue, and store_orders, with a Create new button

The query builder

The Explore screen has three areas:

  • Data panel (left) holds the measures and dimensions available on the Metric Block, plus the shelves where you build the query.
  • Result (middle) shows the chart, with tabs for Raw results (a table) and SQL (the generated query).
  • Visualization (right) is the chart type picker and per-chart settings.

Shelves

Drag or click fields onto these shelves to shape the query:

Shelf What it does
Rows Dimensions to break the data down by (row groups)
Columns Dimensions that pivot across columns
Segment by A dimension that splits the result into small multiples or color groups
Measures (Add measure) The numbers to plot, picked from the block's measures
Filters Narrow the rows (see operators below)
Sort Order the result by any dimension or measure, ascending or descending
Limit Cap the number of rows returned (default 1000)

Filters

Add a filter, pick a field, and choose an operator. Available operators:

=   !=   >   <   >=   <=   IN   NOT IN   LIKE   NOT LIKE   IS NULL   IS NOT NULL

For IN and NOT IN, enter a comma-separated list (a, b, c). IS NULL and IS NOT NULL take no value.

Time ranges and grains

If the Metric Block has a timeseries column, you get time controls for free:

  • Time range is a relative range like Today, Yesterday, Last 7 days, Last 30 days, Last 90 days, This week, This month, This year, or Last 1 year. You can also pick an absolute date range yourself.
  • Time grain buckets a date dimension by Year, Quarter, Month, Week, Day, Hour, or raw (no bucketing).
  • Time function controls how the grouping works. Truncate uses DATE_TRUNC, which keeps the date type and pins to the start of the period (best for time series). Extract uses EXTRACT, which returns the number, for example 4 for April (best for seasonality).

Running the query

1
Click the measures and dimensions you want to plot. Add filters or a time range if needed.
2
Click Run.
3
On the right, under Visualization Type, pick your chart. Check Raw results for the underlying table and the SQL tab for the generated query.
4
Name it at the top and click Save. It now appears in the Content list under Explore.
An Explore showing the data shelves on the left, a bar chart of total revenue by car model in the middle, and the visualization type set to Bar Chart on the right

Visualization types

Pick the chart that fits your data. Each type has its own settings (which field is the X axis, what to group by, stacking, etc.) in the panel on the right.

Type
Best for
KPI Card
A single headline number, with optional sparkline and comparison.
Bar Chart
Comparing categories. Vertical bars with optional grouping or stacking.
Line Chart
Trends over time. Time series with optional grouping.
Area Chart
Cumulative trends. Stacked area over time.
Donut Chart
Proportions and share of total. Pie or donut.
Bar List
Ranked top-N lists, shown as horizontal bars.
Spark Chart
A tiny inline trendline with no axes.
Table
A pivot table with row and column dimensions, measure cells, and optional inline heatmap bars.
Custom
A custom JSX or React component for cases the built-in charts don't cover.

Builder vs YAML

Like Metric Blocks, an Explore has a Builder view (the visual shelves) and a YAML view (the same query and visualization as text). Toggle between them at the top. The YAML is useful for copying a chart definition, reviewing it, or version-controlling it.

Exporting

From a saved Explore you can Export the result directly. The formats are:

  • PNG for a chart image (set width, height, and background)
  • PDF for a document (with page size, orientation, and margin)
  • CSV for the raw data
  • XLSX for an Excel workbook

To deliver a chart on a schedule instead of a one-off export, attach the Explore to a report. See Charts in Reports & Slack.

Next steps