xmetric vs Preset and Apache Superset: Shelf Grammar vs Chart-Type Forms

Composable shelf grammar vs Superset's chart-picker forms

Apache Superset and its managed cloud service Preset use a chart-plugin form workflow in Explore: you select a dataset and visualization type and map columns into form controls such as Time Range, Group by, and Metrics, then click Run Query.

xmetric is built around Tableau's visual and calculation grammar:

1. You do not pick a rigid chart template upfront. In the xmetric app, you drag any dimension or measure onto Rows, Columns, Color, Size, Label, or Detail.
2. Adding a second dimension to Rows or Columns splits the view into small multiples; adding a second measure gives you dual axes or stacked measures without losing your configuration.
3. You can test LOD and table calculations in the formula playground before using them on shelves.

Multi-grain calculations: LOD expressions vs virtual datasets in SQL Lab

In Superset and Preset, the thin semantic layer supports virtual metrics (SQL aggregations like SUM(recovered) / SUM(confirmed)) and virtual calculated columns (non-aggregated SQL expressions like CAST(recovery_rate AS float)), which evaluate at the grain of the chart's GROUP BY query. To compute a customer's first order date or join across grains, you open SQL Lab, write a custom SQL query with CTEs or window functions, and save it as a Virtual Dataset.

In xmetric, you write a single Level of Detail (FIXED, INCLUDE, EXCLUDE) or table calculation formula as a calculated field and use it on shelves.

Customer Cohort Acquisition Month (xmetric LOD syntax):

DATETRUNC('month', { FIXED [User ID] : MIN([Event Timestamp]) })

SQL Lab Virtual Dataset equivalent in Superset/Preset:

SELECT
  user_id,
  event_timestamp,
  revenue,
  DATE_TRUNC('month', MIN(event_timestamp) OVER (PARTITION BY user_id)) AS cohort_month
FROM events;

Category Share of Total Revenue (xmetric Table Calculation):

SUM([Revenue]) / TOTAL(SUM([Revenue]))

SQL Lab equivalent:

SELECT
  category,
  SUM(revenue) AS category_revenue,
  SUM(revenue) / NULLIF(SUM(SUM(revenue)) OVER (), 0) AS pct_of_total
FROM events
GROUP BY category;

Deployment and local file trade-offs

Self-hosting Apache Superset requires deploying a Python web server, Celery workers, caching, a metadata database (PostgreSQL or MySQL), and SQLAlchemy database drivers via Docker Compose, Kubernetes/Helm, or PyPI. Preset Cloud removes that operational work (Starter is free forever for up to 5 users on 1 workspace; Preset Professional is $20/user/month billed annually; $25/user/month billed monthly; Preset Embedded is Viewer licenses from $500/month for 50 licenses for 50 viewer licenses), and requires connecting a database or enabling Allow file uploads to database on a writable database connection to upload CSV, Excel, or Parquet files into a database table.

xmetric runs inside your browser tab for free, with no server or database setup and no account needed until you want to save online.

Spec comparison: xmetric vs Preset and Apache Superset