DSMA Admission case studies
DSMA Admission · Put your knowledge to work in context.
The DSMA Admission exam is the entry assessment for a data science and analytics master's-style program, and it is designed to check whether you have the quantitative and computational foundation to start graduate-level coursework. It suits candidates coming from statistics, computer science, economics, engineering, or a working analytics role who can already read code and reason about data. The DSMA Admission case studies are 5 cases and about 2,289 words.
1Reusing Hospital Records for a Vendor Sepsis Model
A regional hospital's analytics unit holds 1.2 million de-identified encounter records from 2018–2023. The original patient notice said data would be used 'to support treatment, payment, and hospital operations.' A vendor offers a free sepsis-prediction model but requires the full encounter table-admission and discharge dates, 5-digit ZIP, and free-text triage notes-to retrain it.
436 words, about 2 min2Shipping a Mobile-Only Checkout Variant After a Noisy A/B Test
An online retailer ran a 14-day A/B test on its checkout page. Traffic was split 50/50 by user ID.
451 words, about 2 min3A Revenue Dashboard That Double-Counts Refunds
A finance team reports that a SQL dashboard overstates net revenue by about 12% and omits some regions. The query joins customers to orders and then to refunds: SELECT c.region, SUM(o.amount) AS gross FROM customers c LEFT JOIN orders o ON o.customer id = c.customer id LEFT JOIN refunds r ON r.order id = o.order id WHERE r.status < 'cancelled' GROUP BY c.region; orders has one row per order; refunds has zero to three rows per order and a nullable status.
499 words, about 2 min4Nightly Pandas Merge That Drops Stores and Margins
A retail chain's nightly Python job merges 40 store CSV files with a store master table and produces a margin report. Store IDs look like '00123' in the CSVs; pandas reads them as integers, while the master stores them as strings.
455 words, about 2 min5A Churn Model With Post-Event Leakage and a Costly Threshold
A subscription company built a churn model on 300,000 customers with a 2% churn rate. The random-split test AUC is 0.88, but precision in the top decile is 0.18 and recall at the chosen 0.50 threshold is 0.35.
448 words, about 2 min