Hex and Mode Alternative for Zero-Setup Ad-Hoc Data Exploration

When a cloud notebook is overkill for ad-hoc exploration

Data teams love Hex and Mode (acquired by ThoughtSpot and packaged as Analyst Studio) for collaborative SQL + Python notebooks tied to Snowflake, BigQuery, or Databricks. But for fast ad-hoc exploration on a file in front of you, notebook workspaces introduce friction:

1. Setup before your first chart. Creating a project, uploading a file into workspace storage, waiting for a cloud kernel to start, and writing a SELECT cell just to inspect column distributions takes minutes before you see a single bar chart.
2. Chart cells are secondary to code cells. While both tools include chart cells, pivoting across multiple dimensions, adding small multiples, changing mark types (bar, line, area, scatter, heatmap), and nesting Level of Detail aggregations is much slower than Tableau's shelf-and-mark grammar.
3. Per-editor seat pricing adds up. Hex charges $36 per Editor / month on Professional and $75 per Editor / month on Team, and Mode (ThoughtSpot Analyst Studio) requires a paid ThoughtSpot plan (see the vendor's pricing page) or Enterprise contract beyond its free trial.
4. Cloud upload by default. Uploading a local .parquet or .csv export into a hosted notebook sends your data to remote cloud compute.

How to run ad-hoc visual exploration in xmetric

xmetric gives analysts and engineers local file exploration inside a browser tab with no setup:

1. Load local files or BI workbooks. Import a .csv, .xlsx, or .parquet file, or open a .twb or .twbx workbook, in the xmetric app (/app/), with no cloud kernel to start. To look inside .pbix and .pbip files, use the free workbook inspector (/inspect/).
2. Pivot and facet with shelves. Drag dimensions to Columns and Rows, split series by Color or Detail, and build small multiples or cross-tabs from the same fields.
3. Write LODs and table calculations without DAG spaghetti. Instead of chaining 4 SQL cells in a notebook to compute a cohort denominator and join it back, write one { FIXED ... } or WINDOW_* expression. To test one alone and see the equivalent SQL, use the formula playground (/play/).

Replacing a 3-cell notebook chain with one calculation

In a notebook tool, comparing each customer's monthly spend against their own historical average or cohort baseline usually requires a CTE or chained SQL cells followed by a chart cell. In xmetric, you can write it as a single calculated field.

Customer spend vs their personal lifetime average (FIXED LOD)

xmetric calculated field (Order vs Customer Avg):

AVG([Order Amount]) - AVG({ FIXED [Customer ID] : AVG([Order Amount]) })

Equivalent SQL query:

SELECT
  customer_id,
  order_id,
  order_amount,
  order_amount - AVG(order_amount) OVER (PARTITION BY customer_id) AS diff_from_customer_avg
FROM orders;

Moving 4-week average alongside raw weekly numbers (Table Calculation)

xmetric calculated field:

WINDOW_AVG(SUM([Events]), -3, 0)

Equivalent SQL query:

SELECT
  DATE_TRUNC('week', event_time) AS event_week,
  SUM(events) AS weekly_events,
  AVG(SUM(events)) OVER (
    ORDER BY DATE_TRUNC('week', event_time)
    ROWS BETWEEN 3 PRECEDING AND CURRENT ROW
  ) AS rolling_4wk_avg
FROM telemetry
GROUP BY 1;

Current scope and limits

• Local files today: The xmetric app imports .csv, .xlsx, and .parquet files and opens .twb and .twbx workbooks (not every dashboard will look identical to Tableau), with shelves, Show Me charts, calculated fields, LODs, and table calcs.
• Warehouse connections: Direct live connections to Postgres and BigQuery are coming.
• Notebook code execution: xmetric is a visual BI tool (shelves and marks, LODs, table calcs), not a multi-cell Python runtime for training ML models.
• Agent integration: Every action in the app is a command. An MCP server, ChatGPT app, and Claude connector are in development.

Spec comparison: xmetric vs Hex, Mode, and Preset