This is what a data job
actually looks like.

Practice the full analyst workflow: SQL, Python, and Storytelling. On realistic business problems. The way work actually happens.

3 modules per scenario·SQL → Python → Storytelling·Real work, not puzzles

You know SQL.
Here's what's next.

The jump from technical skills to real business impact is a skill in itself. Smartgoose is where you practice it.

Traditional platforms

Problem #247

-- Write a query to return the Nth highest salary

SELECT

salary

FROM

employees

ORDER BY

salary DESC

LIMIT

1 OFFSET N-1;

✓ Accepted·Runtime: 42ms

Smartgoose scenario

SC
Sarah ChenVP Product · Today 9:14 AM

Hey team, premium churn is up 8.3% month-over-month. Marketing thinks it's pricing. Engineering thinks it's a bug. I need data. Pull whatever you need from the warehouse and give me a clear recommendation by EOD.

What do you do next? Which tables do you query? What's your hypothesis?

Skills work better together

SQL, Python, and Storytelling aren't separate subjects. In real work, they're one continuous flow. That's how Smartgoose teaches them.

Real work is open-ended

A business question doesn't come with a schema hint or an expected output. Smartgoose gives you the ambiguity that makes the work meaningful.

The full workflow, finally in one place

Business problem → data pull → analysis → recommendation. That's the real job. Now there's a place to practice all of it.

How it works

One scenario.
Three modules.
One recommendation.

Every scenario flows through the exact same workflow you'll use on the job. From ambiguous question to boardroom-ready answer.

Step 01 · Business Problem

A real business question lands in your inbox

Every scenario starts with a realistic message from a simulated VP: ambiguous, urgent, and exactly what real work looks like. No hints. No schema diagrams. Just a question and a deadline.

# analytics-team
SC
Sarah ChenVP Product · Mon 9:14 AM

Team, premium churn is up 8.3% MoMand I'm getting questions from the board. Marketing thinks it's pricing. Engineering thinks it's a bug. I need data. Give me a clear recommendation by EOD.

Step 02 · SQL Module

Pull only the data that matters

Write CTEs, window functions, and cohort queries against a simulated 6-table corporate schema. The data has intentional messiness: nulls from a migration, weird joins, edge cases that bite you on the job.

churn_analysis.sql
1
2
3
4
5
6
7
8
9
10
-- Churn cohort analysis · 6-table schema
WITH churn_cohorts AS (
SELECT
u.user_id,
DATE_TRUNC('month', s.started_at) AS cohort_month,
LAG(s.status, 1) OVER (
PARTITION BY u.user_id ORDER BY s.created_at)
AS prev_status,
s.tier, s.churned_at
-- subscription_tier has nulls from Mar 2024 migration
)
Step 03 · Python Module

Clean, transform, and visualize

Load your SQL output into Pandas. Handle the messy data your query returned. Build the charts that tell the story. Then decide which one actually answers the VP's question.

Retention by cohort

6-month view
100%75%50%25%M0M1M2M3M4M5AvgFeb
Feb cohort
Avg cohort
Step 04 · Storytelling

Lead with the recommendation

Structure your findings using the Pyramid Principle. Lead with % deltas, not methodology. Give a VP something they can act on immediately. That's the real deliverable.

Executive Summary

Churn is concentrated in the Feb cohort

23% churn vs 8% average (2.9× the baseline)

Root cause

Trial-to-paid friction at day 14, not pricing sensitivity

Recommendation

Reduce paywall friction for free-to-paid upgrade flow

Impact

40% reduction in Feb cohort churn pattern in Q2

See what a scenario looks like

Purpose-built data. Real-world messiness. A clear deliverable at the end.

Smartgoose: Scenario #001 · Subscriber Analytics
90–120 min

Scenario #001

Subscriber Analytics

Modules

01Business ProblemScenario #001

From your inbox

SC
Sarah ChenVP Product · Monday 9:14 AM

Team, premium churn is up 8.3% month-over-month and I'm getting questions from the board. Marketing thinks it's pricing sensitivity, Engineering thinks it's a bug. I need data. Pull whatever you need and give me a clear recommendation by EOD. What's actually driving this?

Your task

Investigate the churn spike. Identify the root cause. Write a structured recommendation with data to back it up, by EOD.

What makes Smartgoose different

01

End-to-end corporate workflow

Start with a stakeholder email. End with a boardroom-ready recommendation. Every step in between is yours to navigate.

02

Real business problem framing

Every scenario opens with an ambiguous question from a simulated VP. The kind that lands in your inbox on a Monday morning with no instructions attached.

03

Purpose-built messy data

Nulls, schema migrations, weird joins, and edge cases designed to mirror the data quality issues you actually encounter in production environments.

04

SQL + Python + Storytelling: Connected

Three modules, one scenario. Your SQL output feeds your Python analysis. Your Python analysis feeds your story. Nothing exists in isolation.

05

The recommendation is the deliverable

Not a quiz score. A structured business recommendation with % deltas and a clear next action. That's the real output of data work.

06

Deadline pressure, simulated

Every scenario has context: who asked, why it matters, and when they need it. Because context is half the work in a real data role.

You have the skills.
Now use them like a pro.

Built for data professionals who are technically strong and ready to make a bigger business impact. One real scenario at a time.

Data Analyst

You write solid SQL and you're ready to connect your analysis to real business outcomes.

SQL proficient and ready to drive business impact.

Product Analyst

You understand the product deeply and you're ready to back your instincts with data.

Cross-functional, stakeholder-facing, metric-driven.

BI Analyst

You build dashboards people love and you're ready to move into more strategic analytical work.

Dashboard builder ready for more strategic work.

Train for Monday morning,
not just the interview.

Early access when Smartgoose launches.