TOPIC
Data modeling
These are the same questions found in the library, grouped here by the capability each one mainly examines.
138 questions
Questions 51–100 of 138 on this page.
A · Professional problem solving and implementation8
- Offline evaluations keep improving while a specific user segment regresses online. Which data leakage, slice coverage, and deployment differences would you investigate?
- A workspace contains hundreds of items that people find by customer, status, owner, or date. Its navigation follows the company org chart, so users struggle to find things. How would you reorganize it and test findability?
- Users can edit a profile on a phone while offline and may also edit it on the web. The device must sync when connectivity returns without silently overwriting valuable changes. How would you design versioning, conflict detection, merging, and user confirmation?
- An AI assistant retrieves documents from several business systems, where users have different permissions in each. How would you ensure retrieval, caching, summaries, and citations follow source permissions, including after access is revoked?
- A business-rule correction requires recalculating historical records across several services, but some consumers have already sent notifications or performed external actions. How would you design correction events, impact lists, replay boundaries, and side-effect suppression?
- A model begins returning structured JSON, while mobile, server, and analytics consumers upgrade at different rates; new fields may be absent or enum values may change. How would you define the output contract, handle validation failures, and preserve compatibility?
- Overall task success is unchanged after a model update, but outcomes diverge more across users by language, device, and experience. How would you define evaluation segments, manage small-sample risk, and decide whether to continue rollout?
- The team needs real user conversations to reproduce model failures, but logs contain personal and sensitive business content. How would you minimize collection, redact data, authorize access, and set retention while preserving debugging value?
No questions in this section match your target. ·
B · Business understanding and domain modeling41
- Business units use an “active customer” metric, but some define it by payment, others by usage, and others by likelihood to renew. How would you decide which definitions should be standardized and which should remain distinct concepts?
- Recommendation and risk teams both need a “spend in the last 30 days” feature, but one uses calendar days and the other a rolling hour window, and refund handling differs. How would you decide whether the definition can be shared?
- A risk value is usually present in training data, but live serving often substitutes a default and model performance fluctuates. How would you find whether the cause is data collection, timing, missing-value handling, or implementation differences?
- Models, operational rules, and reports share a core feature, but the upstream business definition has changed and some consumers cannot migrate yet. How would you manage old and new definitions, ownership, and migration order?
- The commercial rule for a service changes from a free trial to a refund after conditions are met, while historical orders were signed under the old terms. How would you evolve the model and workflow so both old and new transactions remain explainable?
- A supplier API replaces one “delivered” status with “dispatched” and “received,” but historical data has no basis for distinguishing them. How would you update the domain model, external contract, and related reporting?
- Teams define “done,” “awaiting acceptance,” and “paused” differently, distorting roll-up data. How would you design a shared state model and transition rules?
- A user can belong to several organizations and hold different roles in each. How would you model identity, membership, and permissions without conflating accounts with business roles?
- Customers upgrade, downgrade, or pause subscriptions mid-cycle, while sales and finance disagree on revenue recognition. How would you express the rules and handle historical invoices?
- Business reports mix orders, order lines, and payment attempts, so metrics disagree across pages. How would you establish fact-table grain and validate aggregation semantics?
- The source database has deleted a customer, but a change data capture (CDC) pipeline and downstream summary views still show the old record. How would you trace the deletion event, update dependent views, and verify cleanup?
- Two models call a feature with the same name but receive different values due to defaults and missing-data handling. How would you establish a contract and compatibility checks? Follow-up: If a critical feature refresh misses the real-time decision window, how would you choose among stale values, fallbacks, or refusing the decision?
- An order moves through created, awaiting confirmation, processing, partially fulfilled, canceled, and refunded states; different pages and services interpret “complete” differently. How would you model state, events, and valid transitions so each client does not invent its own meaning?
- A user upgrades mid-cycle and later requests a refund. The contract requires a calculation based on remaining service time, while promotional discounts and taxes also affect settlement. How would you decompose the rules and make the amount calculation traceable?
- A user clicks a campaign link on Monday and purchases on Tuesday, but the purchase event uploads on Friday because the client was offline. How would you define event time, arrival time, attribution windows, and backfill rules so results remain explainable?
- A database field has been replaced, but reports, exports, and external integrations still read it; deleting it directly could cause silent errors. How would you plan compatibility migration, consumer confirmation, and final removal?
- An enterprise customer uses one contract for several departments; employees may work across departments, and administrators may operate on behalf of multiple accounts. How would you distinguish organizations, contract accounts, individual identities, memberships, and authorization scope?
- A product feature is configured by customer contract, adjusted by experiment group, and constrained by regional rules; analysts need to explain which combination shaped a user's experience. How would you model configuration sources, scope, effective time, and decision provenance?
- A dashboard's total conversion rate cannot explain which customer segment or process step caused a change; teams want to drill down by channel, plan, and region, but dimensions have different grains. How would you define fact grain and valid aggregation rules?
- A cross-team objective spans several projects, each with routine tasks, risks, and incidents; retrospectives struggle to trace which decisions and work drove an outcome. How would you model relationships among objectives, projects, work items, events, and evidence?
- Business users can edit rules ahead of time, schedule them after approval, and still have older versions referenced by some users and asynchronous jobs. How would you model draft, review, publication, effective time, and rollback?
- Order revenue is shared by a platform, provider, and channel partner. Weeks later, part of the order is refunded; one party has already settled and lacks sufficient balance. How would you express allocation rules, refund responsibility, recovery, and reconciliation state?
- Business reporting needs the duration of every state transition, while an operations page needs only the object's current state; a wide table overwrote history during updates. How would you separate event facts, snapshots, and current-state models?
- CRM, billing, and product logs use customer IDs, contract IDs, and anonymous device identifiers; analysts need one customer's lifecycle. How would you model identity mappings, confidence, temporal validity, and correction?
- Several teams maintain customer profiles. When data is wrong, no one knows which system is authoritative, and systems also write values back to one another. How would you define field ownership, source of truth, synchronization direction, and dispute handling?
- A device reports a cumulative counter, but the counter resets after device replacement or data correction; the data platform may also backfill historical events. How would you define deltas, reset markers, negative differences, and recomputation boundaries?
- A global product counts daily activity by calendar day, but product analytics, the warehouse, and model serving use user time, UTC, and server time respectively; results often differ near month-end. How would you define boundaries and versioned semantics?
- Billing changes from calendar-month pricing to service-day pricing, with the new rule applying only to new contracts; old contracts cross the release date, and support still needs to explain past invoices. How would you version rules, contract applicability, and invoice snapshots?
- Business categories may be added, merged, or renamed. Reports must reproduce the definition used at the time and compare results under the current definition. How would you preserve source categories, mapping versions, and report definitions?
- An event producer wants to change a field from a string to a structured object, but some consumers deploy slowly and others drop unknown fields when writing records back. How would you stage dual writes, version negotiation, migration, and retirement of the old format?
- A customer-ownership field in historical events is wrong and must be corrected in bulk to recompute downstream metrics; the same events may also trigger notifications and partner callbacks. How would you separate data recomputation from side-effect replay?
- An accountant uses one identity to manage several legal entities, each with separate contracts, invoices, and approvers; some group reports allow cross-entity summaries. How would you model identity, membership, resource ownership, and cross-entity access?
- Several users file cases about one incident. Some can share a resolution, but each has different contract deadlines, communications, and compensation eligibility. How would you model the parent issue, linked cases, individual commitments, and closure criteria?
- Plans sold through different channels share a display name but have different features, quotas, and renewal terms; users may switch channels during an upgrade. How would you distinguish the catalog display, sellable offer, contract version, and effective entitlements?
- Customers can apply points to part of an amount and then use a cash coupon; points may be funded by the platform or a partner and some may need restoration after cancellation. How would you separate payment, discount, point liability, and refund rules?
- An enterprise pays per active seat, but some employees share accounts across shifts while others have accounts they never use; the contract does not define “active.” How would you determine the billable entity, activity definition, and dispute evidence?
- Orders contain nested line items, shipping, and tax, while refunds may apply to only some items; joining order headers directly to lines doubles totals in existing reports. How would you model grain and aggregation semantics?
- Sales performance rolls up through a regional hierarchy. Midyear, a region is split; historical reports must remain unchanged, while future reports use the new hierarchy. How would you represent organization versions, effective time, and attribution?
- Some transaction details must be retained for years under policy, while product analytics wants to use them for user segmentation and experiments. How would you model purpose authorization, retention, and access views?
- A user's transaction enters a behavioral feature window, then is reversed days later due to refund or fraud review; downstream models and reports have already consumed the feature. How would you model reversal, window recomputation, and impact propagation?
- New users have little behavioral history, and the feature pipeline fills missing values with a global average, making “not observed” indistinguishable from “average behavior.” How would you represent cold-start status and evidence boundaries?
No questions in this section match your target. ·