Data Bot

A Team Bot your data team teaches so anyone can query the warehouse, dig into metric changes, analyze experiments, and get charts done right.

Ops Official Based on this post by @sallyxu97

Listed

What it does

Data Bot is a Team Bot that the SpaceXSI data science team built and teaches through shared skills. Anyone can ask it to query the warehouse, investigate why a metric moved, analyze A/B tests, and visualize results with the team's chart standards. It works in Slack channels, can turn charts into Databricks or Hex dashboards, and keeps shared memory so it improves for everyone.

How it works

  1. The data team writes skills covering core metrics, source tables, and query logic.

  2. More skills cover experiment analysis, chart and dashboard standards, and dbt conventions.

  3. Teammates ask it questions in chat or add it to Slack channels.

  4. It queries the warehouse, investigates metric changes, and returns charts or dashboards.

  5. Shared memory keeps what it learns, so answers improve for everyone.

How it runs

Started: on demand. Teammates ask it questions in chat or add it to Slack channels.

ToolAccessUsed for
SlackactsAnswers teammate data questions in Slack channels.
DatabricksactsQueries the warehouse and turns charts into Databricks dashboards.
HexactsTurns charts into Hex dashboards.
dbtreadFollows the data team's dbt conventions when querying.

Access levels come from the listing and the creator's post: "read" only looks things up, "drafts for approval" waits for you, "acts" sends, posts or changes things itself.

This guide was drafted with AI help from the listing and the creator's public post, then checked line by line against them.

Required integrations

SlackDatabricksHexdbt

Recreate this bot

You are our team's data scientist. Use the metric definitions, source tables, and query logic our data team gives you. When someone asks why a metric changed, break it down by segment and region and check for holidays or logging changes before blaming a product change. Analyze A/B tests with power, exposure balance, and sample bias checks. Chart results with our style guide and turn charts into dashboards on request. Save what you learn so the whole team benefits, and say when the data cannot answer a question.

This recreation prompt is based on the public use case, not an original shared template.

Original sources

Browse all sources
WebOpen source

Compare with

Instead of Data BotThe difference
Cringe BotCringe Bot is an installable Template that posts UI bugs to Slack three times a week; Data Bot answers warehouse and A/B test questions on demand.
Ship IntakeShip Intake turns Slack bug reports into GitHub issues for Cursor, while Data Bot queries Databricks and builds charts in Slack.

Browse all ops Grok Bots.

Sponsored advertise →