V06 Detect profile drift

Use your current course session. Preparation and course runner · R script ↓

Optional deep dive. Enable only the options named in each block, after preparing its packages or services.

V06a Inspect a data profile

Goal: Separate observation from rules.

Prerequisites: H01; extras.

A profile describes types, missingness and ranges. It does not create quality rules.

training_step(
  "V06a",
  {
    profile_now <- dr_profile_data(policies)
    print(profile_now)
  },
  needs = c("H01"),
  enabled = training_options$extras,
  optional = TRUE
)

Expected result: Three profile rows, each describing six observations.

Task: Is an observed maximum a business limit?

V06 Detect profile drift

Goal: Compare deliveries for change.

Prerequisites: V06a.

Change only the share of missing premiums.

training_step(
  "V06",
  {
    profile_baseline <- dr_profile_snapshot(policies, "delivery-1")
    missing_delivery <- policies
    missing_delivery$premium[1:2] <- NA_real_
    profile_current <- dr_profile_snapshot(missing_delivery, "delivery-2")
    print(dr_profile_compare(profile_baseline, profile_current))
  },
  needs = c("V06a"),
  enabled = training_options$extras,
  optional = TRUE
)

Expected result: Increased missingness in premium.

Task: Is the delivery automatically rejected?