LoginSign Up
← Back to all sparks
E

excluder

ANALYTICS
Velocity0.0

Checks for Exclusion Criteria in Online Data. Contribute to ropensci/excluder development by creating an account on GitHub.

github.com/ropensci/excluder ↗

A Qualtrics data-cleaning package that has been in maintenance mode since its CRAN acceptance.

survey-datadata-cleaningr-packagequaltricsmaintenance
◆Current state
excluder marks, checks, and excludes online-survey rows that fail quality criteria — duplicate responses, suspicious IP or geolocation, screen resolution, completion duration, preview rows. The mark_*/check_*/exclude_* verb trio and the column-renaming helpers are the whole public surface. Recent releases are dependency chasing and test robustness rather than new exclusion criteria.
◆Where it's heading
The package is stable and its maintenance load comes from things it does not control: the {iptools} package leaving CRAN, {tidyselect} deprecating the .data pronoun, IP-geolocation tests breaking when the underlying address data shifts. Much of that work is about staying installable, not about better exclusions. Note that several of these entries were backfilled into the feed within the same two-minute window and are not in version order.
◆Prediction
The next release will most likely be another dependency or CRAN-check response rather than a new exclusion criterion, following the pattern of the last three.

◆Recent moves

  1. 1y ago

    Internet-dependent tests and examples made conditional

    Almost entirely test and documentation hygiene driven by IP-geolocation data shifting outside the US. The one user-visible piece is a warning when columns are already renamed.

    View source ↗
  2. 2y ago

    qualtrics_fetch2 dataset and tidyselect deprecation fixes

    A second example dataset with labels as column names makes the rename path testable by users, alongside the .data-pronoun migration that {tidyselect} forced. Typical of the dependency-driven maintenance this package now runs on.

    View source ↗
  3. 3y ago

    CRAN acceptance and graceful IP-lookup failure

    CRAN acceptance and a JOSS paper land together, and the IP functions learn to fail gracefully when the geolocation source is unreachable — the fix that made the package safe to depend on in automated pipelines.

    View source ↗
  4. 3y ago

    dplyr 1.0.8 across()/is.na() compatibility fix

    An upstream dplyr change broke a predicate; the fix swaps across() for if_all(). No behavioural difference for users on a working version.

    View source ↗
  5. 3y ago

    use_labels(), rename_columns(), and orientation-agnostic resolution

    The most substantive feature release in the window: question labels can become column names, and res_min applies a resolution floor to both dimensions instead of assuming every screen shares an orientation. Both address real failure modes in mixed-device survey data.

    View source ↗
  6. 3y ago

    Swaps iptools for ipaddress

    A dependency substitution forced by {iptools} leaving CRAN, with no change to how exclusions behave. Exactly the kind of upkeep that dominates this package's recent history.

    View source ↗