Free print-and-play · Set 1

PIPELINE:The Data Engineering Card Game

Connect data sources. Deploy engines. Run queries before the clock runs out. Build the better data team in exactly ten rounds.

  • 2players
  • 10rounds
  • 40cards per deck
  • 200unique Set 1 cards
Data Warehouse Source cardBatch Cluster Engine cardDaily Sales Rollup Query card

The idea

Build a pipeline. Score the query.

Each player is a data team racing to score Query Points (QP). Your Sources supply Data, your Engines supply Compute, and your Queries turn both into points. Bottlenecks disrupt the other team; Optimizations keep yours moving.

01

Connect Sources

Play up to one Source each turn. Engage it to produce the Data pips needed to pay card costs.

Resource
02

Deploy Engines

Engines have Compute and Concurrency. A matching Query type gets an Affinity discount to its Complexity.

Capacity
03

Run Queries

Pay the Data cost, invite responses, then compare Complexity with Compute. Successful Queries score QP.

Points
The clock is fixed. After both players finish their tenth turn, subtract Tech Debt from QP. More points wins. Ties go to fewer failed Queries, then more Sources.

At the table

Four steps, every turn

  1. 01

    Refresh

    Ready your Sources and Engines. Advance Long-Running Queries and resolve any that finish.

  2. 02

    Ingest

    Draw one card. If your Backlog is empty, take one Tech Debt instead. The first player skips this step on turn one.

  3. 03

    Build

    Connect a Source, deploy Engines, attach cards, play effects, and run Queries in the order you choose.

  4. 04

    Commit

    Resolve end-of-turn effects and discard down to seven cards.

When a Query runs

Choose an available Engine, pay the Data cost, and open a response window. Opponents can play Incidents; you can answer with Instant cards. Last played resolves first. If Compute is lower than Complexity, the Query fails, scores nothing, and gives you one Tech Debt.

Long-Running work

These Queries occupy an Engine slot while Progress counters tick down on your Refresh steps. They can still be disrupted. Unfinished work at game end fails without adding Tech Debt.

Your resources

Five kinds of Data

Card costs use colored pips. A numbered gray pip can be paid with any type; a Join asks for multiple specific types.

B

Batch

Big, steady scans

S

Stream

Fast, fragile queries

J

Semi-structured

Flexible joins and filters

U

Unstructured

AI and Similarity plays

G

Governance

Protection and control

Work with Alex