Choose Your Module
Each module is a self-contained story told through data challenges — SQL, pipeline design, and more. Solve them in order to uncover hidden findings.
Data Case Files
Vol. 1 — Follow the Money
Someone at Meridian Corp is moving money they shouldn't be. Phantom transactions in the ledger, tampered archives, suspicious wire transfers, inflated prices — the evidence is buried in the numbers. You're the detective. Work the cases, follow the money, and piece together a conspiracy that someone went to great lengths to hide.
Station Logs
Kepler-9 Incident Report
Orbital Station Kepler-9 is failing. The station AI (ARIA) has flagged cascading anomalies across sensor arrays, power grids, and telemetry feeds — but the raw data is noisy and possibly corrupted. As the senior data engineer on shift, you need to triage the sensor readings, build pipelines to summarize the readings, and cut through the noise to find what's really going wrong before it's too late.
The Lost Brewery
A Medieval Monastery Mystery
A centuries-old monastery has brewed legendary ale since 1180. Quality has been declining for decades. The abbey's handwritten brewing ledgers were scanned and run through OCR (Optical Character Recognition) to convert them to digital text — but the software mangled the old handwriting badly. You're a data historian hired by the Abbot to clean the garbled records, trace the quality decline, find who changed the recipe — and uncover why.
The Gallery
Art Heist Forensics
Last night, 'The Last Garden' by Elara Voss — worth $40M — vanished from the Thornfield Gallery during a private gala. Security footage is corrupted, but the metadata survived. You're a forensic data analyst. Reconstruct the evening from badge scans, camera logs, and staff access records.
Double Entry
The Flat Table Disaster
Questborne, a small game studio, shipped their online multiplayer game "Realm of Echoes" with one giant table for everything — players, items, guilds, gold, all mashed together. Now the in-game economy is melting down: duplicate items, phantom gold, guild leaders who change retroactively. You're the contractor called in to untangle it.
Terra Incognita
The Lost Trade Routes
A medieval historian has digitized coordinates from fragmentary Silk Road maps recovered from a monastery archive. The data is messy — duplicate waypoints, missing coordinates, conflicting distances between ancient cities. Somewhere in the noise is a trade route that hasn't been traveled in eight centuries, and a lost oasis that cartographers have debated for decades. The Haversine formula and careful spatial analysis are your compass.
Echo Chamber
The Ouroboros Incident
A dead social platform called Ouroboros had its recommendation algorithm go rogue — injecting synthetic posts, creating feedback loops, and burying the evidence in layers of self-referencing data. The platform is offline now, but the database survived. Trace the recursive threads, detect the cycles, and uncover how a system designed to connect people ended up consuming itself.
Model Behavior
The Alignment Incident
Day 1 at Tessera AI. You're supposed to be onboarding — instead, you walk into a live incident. The company's flagship AI model is getting worse with every release — evaluation scores are plummeting — and nobody noticed until customers started complaining. Somewhere in the inference logs — the records of every prediction the model makes — duplicate records are masking the decline, and the automated safeguards that should have caught it never existed. Find what went wrong, prove it in the data, and build the monitoring pipeline that prevents it from ever happening again.
Diamond Data
Baseball Sabermetrics
Sabermetrics is the use of statistical analysis to evaluate baseball players beyond traditional scouting. Small-market teams can't outspend the giants — but they can outsmart them. The Ridgeline Aces haven't made the playoffs in a decade, and new GM Sarah Chen is betting on analytics to turn things around. As the team's first data analyst, you'll crunch batting stats, pitching metrics, and player contracts to find hidden value and build a winning roster on a budget.
The Vinyl Vault
System Architecture
Groove Records has survived forty years selling vinyl from a dusty Brooklyn storefront — Maxine Torres opened it with a crate of records and a conviction that people still want to hold music in their hands. Now, with a warehouse of rare pressings and a cult following, she wants to take it online. You're the architect she's hired to build it right. The catalog is a mess of duplicates, customers want real-time stock alerts, EU regulators are watching, and Maxine's dream of live auctions for ultra-rare pressings won't survive a system that buckles under load.
Cloudbreak
Analytics Architecture
Pelican Point sits at the end of a dirt road on Western Australia's Indian Ocean coast — a legendary left-hand reef break where three-meter swells peel for 200 meters over shallow limestone. Kai Nakamura has surfed it since he was twelve and logged conditions from the cliff above his shop for thirty years: swell height, period, wind direction, tide, water temperature, all in pencil. Now the World Surf League wants to run a competition here. They have an eleven-day holding window and need to pick the best three days — get it wrong, and the world's best surfers paddle into mush on live television. Kai's notebooks hold the patterns, but patterns aren't predictions. His nephew Tane — a data engineer at a surf forecasting startup — has a plan: wire together buoy telemetry, satellite imagery, and decades of handwritten observations into something a model can learn from and a broadcast team can trust. You're spending the summer with Tane, turning pencil sketches into a production data platform — from a raw landing zone to real-time dashboards to wave-quality predictions, one architectural layer at a time.
The Breach
NovaPay Incident Response
NovaPay processes forty million payment records a day. At 3AM, the intrusion detection system fires — someone is inside the network. Raw server logs are the only evidence, and the attacker is still moving. Parse the access logs to reconstruct the initial recon scan. Decode base64-encoded exfiltration payloads hiding in plain HTTP traffic. Unmask spoofed user-agent strings that reveal an insider. Every regex pattern you write is a step closer to the damage report regulators need by morning.