22 skills · category

Data Engineering skills for Claude Code

Skills for moving, cleaning and analysing data: pandas and Polars, dbt models, orchestrated pipelines, data quality checks, analytical SQL, messy CSVs, reproducible notebooks, honest charts, Kafka streams, ethical scraping, and A/B test analysis.

Download all 22 (.zip) Browse the list

Every data engineering skill

Each card opens the full playbook, example prompts and install commands.

No. 215

Airflow DAGs

Airflow 3 DAGs that parse fast and rerun cleanly

Gives Claude the Airflow 3 rules it tends to miss: airflow.sdk imports, removed SLAs and execution_date, and the cron default that makes data_interval_start equal data_interval_end. It writes tasks that read their window from the run, keeps DAG files free of parse-time I/O, makes waits deferrable with explicit timeouts, and ships a stdlib checker that flags Airflow 2 leftovers before they reach the scheduler.

airflow-dags 318 lines · 2 files

No. 216

Analytics SQL

Stated grain, no fan-out, numbers that reconcile

Claude writes the metric definition first, comments the grain of every CTE, and checks row counts after each join so duplicate keys cannot inflate totals. It writes explicit window frames, zero-fills missing days with a date spine, adapts date functions to Postgres, Snowflake, BigQuery or DuckDB, and reconciles the result with a direct total and one hand-checked entity.

analytics-sql 171 lines · 2 files

No. 217

BI Dashboards

One definition per number, one owner per dashboard

Claude writes a dashboard brief and a metric register before any chart, then puts definitions in the tool's semantic layer: Metabase models, Superset dataset metrics, Power BI DAX measures or LookML. Layout runs from headline KPIs to trends, breakdowns and detail, row-level security is tested per role, refresh follows the pipeline, and headline numbers are reconciled against a trusted source.

bi-dashboards 148 lines

No. 218

CSV Cleaning

From mojibake and shifted columns to clean UTF-8

Claude preserves the raw file, runs a bundled profiler that detects encoding, BOM, delimiter, ragged rows, ambiguous dates and European number formats, then repairs structure and parses every column explicitly. Duplicates are removed by a stated rule, bad lines go to a rejects file, and a cleaning log records rows in and out for each step.

csv-cleaning 206 lines · 2 files

No. 219

Data Anonymization

Share data without sharing the people in it

Claude scans every column with a bundled PII detector, classifies direct and quasi-identifiers, and applies a recorded treatment per column: drop, keyed HMAC pseudonyms, generalization, consistent synthetic values or date shifting. It measures k-anonymity before release, builds referentially intact test subsets from production, and re-scans the output so nothing high-risk slips through.

data-anonymization 184 lines · 2 files

No. 220

Data Pipelines

Rerun any day, get the same data

Claude scopes every task to an explicit partition or data interval, replaces appends with partition overwrites, merges or transactional delete-and-insert, and keeps heavy compute out of the scheduler. Retries and timeouts cover transient failures while validation failures alert a person, and idempotency is proven by rerunning a processed partition before any backfill.

data-pipelines 203 lines

No. 221

Data Quality Checks

Catch bad data before the dashboard does

Claude starts from the numbers consumers rely on, profiles history to pick thresholds, and adds schema, key, validity, freshness, volume and reconciliation checks in the tool the project already uses. Each check gets a severity that blocks or warns, bad rows land in quarantine with a reason, and every new check is proven by making it fail on injected faults.

data-quality-checks 219 lines

No. 222

Data Visualization

Charts that make one point, honestly

Claude starts from the takeaway sentence and picks the chart form for that question, then builds it with zero-based bar axes, small multiples instead of dual axes, a highlighted series against gray, direct labels and a title that states the finding. Before export it checks values against the data, grayscale and colorblind readability, and writes alt text.

data-visualization 179 lines

No. 223

dbt Models

Declared grain, tested keys, idempotent increments

Claude organizes dbt projects into source-backed staging views, reusable intermediate models and documented fct and dim marts. Every mart states its grain and enforces it with tests, incremental models get a unique key plus a late-data lookback, and changes are built with dbt build on the modified subgraph, then reconciled against the source system.

dbt-models 257 lines

No. 224

Dimensional Modeling

Declare the grain, conform the dimensions

Claude starts warehouse design with a bus matrix and a one-sentence grain for every fact table, then checks each measure's additivity at that grain. Dimensions get surrogate keys, unknown members and a deliberate SCD type per attribute, special cases use junk, role-playing or bridge dimensions, and the model is reconciled against source totals before anyone builds a dashboard on it.

dimensional-modeling 181 lines

★ No. 225

Experiment Analysis

A/B decisions you can defend

Claude writes the hypothesis, primary metric, guardrails and ship rule before any data arrives, sizes the test from baseline and MDE with a bundled calculator, and checks sample ratio mismatch before reading outcomes. Results come as effects with confidence intervals, with CUPED, delta-method ratios and multiple-comparison corrections where needed, and it calls out peeking and segment fishing.

experiment-analysis 174 lines · 2 files

No. 226

Geospatial Analysis

Right CRS, indexed joins, maps you can trust

Claude inspects every layer's CRS, bounds and validity before doing anything, reprojects to a UTM or equal-area CRS before measuring, and runs point-in-polygon and nearest joins with spatial indexes while reporting unmatched and double-matched features. It repairs invalid geometry, pushes large workloads into PostGIS with ST_DWithin and GiST, and checks results on a map.

geospatial-analysis 179 lines

No. 227

Jupyter Notebooks

Restart, run all, same answer every time

Claude restructures notebooks into a fixed layout with one tagged parameters cell, removes absolute paths and in-notebook installs, pins the environment behind a named kernel, and moves reusable logic into modules with tests. It sets up stripped outputs or jupytext pairing for clean diffs and proves reproducibility with a headless run from a fresh kernel.

jupyter-notebooks 188 lines

No. 228

Kafka Streaming

Keyed order, durable writes, idempotent consumers

Claude maps ordering needs to message keys, sizes partitions for peak load, and configures replication with min.insync.replicas and idempotent producers. Consumers commit only after processing, dedupe in the sink, and route poison messages to a dead-letter topic. It explains where exactly-once actually applies and checks schema changes against registry compatibility.

kafka-streaming 200 lines

No. 229

Lakehouse with Apache Iceberg

One catalog, pruned scans, maintenance on a schedule

Helps Claude decide whether an Iceberg lakehouse is worth it, then build tables that every engine reaches through one catalog. It picks partition transforms and sort orders that keep files near target size, chooses copy-on-write or merge-on-read per workload, and schedules compaction, snapshot expiry and orphan cleanup in a safe order, with confirmation before anything that removes time travel.

lakehouse-iceberg 265 lines · 2 files

No. 230

Pandas Analysis

Every row accounted for, every number reproducible

Claude loads data with explicit dtypes, profiles nulls, keys and ranges before touching anything, and logs row counts through every filter and join. Merges use validate and indicator so duplicate keys cannot inflate totals, aggregations use named groupby, and the analysis ends with assertions, a cross-check against an independent number and a findings note with caveats.

pandas-analysis 202 lines

No. 231

Polars DataFrames

Lazy plans, pushed-down filters, no row loops

Claude writes Polars as lazy plans built from expressions, checks with explain that filters and column selection reach the scan, and collects once, streaming when data outgrows memory. When migrating pandas code it uses a mapping reference, keeps the pandas output as an oracle, and compares frames until null, NaN and ordering differences are resolved.

polars-dataframes 193 lines · 2 files

No. 232

PySpark Jobs

Fewer shuffles, no stragglers, right-sized files

Claude writes PySpark with explicit schemas and built-in functions instead of Python row UDFs, checks the physical plan before running, and reads Spark UI stage metrics to find spill, skew and extra shuffles. It picks broadcast or salted joins from evidence, caches only reused data, and writes idempotent partition overwrites or MERGE into Delta Lake or Iceberg with sane file sizes.

pyspark-jobs 187 lines

★ No. 233

Statistical Testing

The right test, an effect size, honest corrections

Claude picks the hypothesis test from a decision table keyed on outcome type, number of groups and pairing, checks the assumptions that actually matter such as independence of units, and reports effect sizes with confidence intervals ahead of p-values. A bundled script applies Bonferroni, Holm, Benjamini-Hochberg and BY corrections side by side, and results are written as plain-language statements with limits.

statistical-testing 174 lines · 2 files

No. 234

Streamlit Apps

Data apps that do not re-query on every click

Claude designs Streamlit apps around the rerun model: data loaders cached with TTLs, shared clients in cache_resource, multi-input filters inside forms, and live panels in fragments so one widget does not refresh the world. State lives in session_state with stable keys and callbacks, secrets come from st.secrets, and an AppTest suite covers the main path before deploy.

streamlit-apps 210 lines

No. 235

Time-Series Forecasting

Beat seasonal naive on a real backtest first

Claude defines the horizon and decision before modeling, cleans gaps and breaks explicitly, and always fits naive and seasonal naive baselines alongside ETS, ARIMA, Prophet or gradient-boosted candidates. Every model is scored on rolling-origin backtests over the real horizon with a leakage checklist, scale-free metrics and interval coverage, then summarized in a forecast report.

time-series-forecasting 182 lines

No. 236

Web Scraping

Polite scrapers that notice when sites change

Claude checks robots.txt with a bundled script and reviews terms before writing any code, then looks for an API, sitemap, or embedded JSON before parsing HTML, reaching for Playwright only when content is client-rendered. Fetchers honor crawl delays and Retry-After, cache raw responses, resume from checkpoints, and fail the run when field null rates show the layout changed.

web-scraping 183 lines · 2 files

Choosing Data Engineering skills

Install the skills that match requests you make every week; each costs about 100 tokens of standing context until it loads. If you are not sure where to begin, start with Experiment Analysis and Statistical Testing. For background, read The Best Claude Code Skills in 2026, by Use Case, How to Install Claude Code Skills (5 Ways) and Personal vs Project Skills.

By technology: Python skills SQL skills PostgreSQL skills

Install several

Install several at once

The installer accepts any number of skill names. Add --project to install into the current repository.

curl -fsSL https://claudeskills.page/install.sh | sh -s -- airflow-dags analytics-sql bi-dashboards csv-cleaning data-anonymization data-pipelines data-quality-checks data-visualization dbt-models dimensional-modeling experiment-analysis geospatial-analysis jupyter-notebooks kafka-streaming lakehouse-iceberg pandas-analysis polars-dataframes pyspark-jobs statistical-testing streamlit-apps time-series-forecasting web-scraping

Data Engineering skills FAQ

How do I install every Data Engineering skill at once?

Download the bundle at https://claudeskills.page/downloads/category/data.zip and unzip it into ~/.claude/skills/ (or .claude/skills/ inside a repository). To pick a few, pass several names to the installer: curl -fsSL https://claudeskills.page/install.sh | sh -s -- airflow-dags analytics-sql bi-dashboards.

Will installing 22 skills slow Claude Code down?

No. Claude Code keeps only each skill's short description in context — roughly 100 tokens per skill — and loads the full instructions only when a request matches. Install the whole category if it fits your work; unused skills cost almost nothing.

Which Data Engineering skill should I start with?

Start with Experiment Analysis, Statistical Testing. They cover the most common requests in this category; add the rest as your work calls for them.