LoginSign Up
← Back to all sparks
S

survivoR

INFRA · APIS
Velocity0.0

Data from all Seasons of Survivor (US) TV Series in Tidy Format

github.com/doehm/survivor ↗

A Survivor data package that keeps turning a TV show into a relational database

data-packager-packagetelevisionreality-tvrelational-datasports-analytics
◆Current state
survivoR ships Survivor franchise data as R data frames covering the US, Australian, UK and New Zealand versions. Recent releases track broadcast: US48, US49 and US50 arrived across 2.3.6, 2.3.9 and 2.3.12, alongside AU09, AU12 and Australia vs. The World. The 2.3.12 release also reworked castaway_scores into an explicit three-tier structure - standardised residual scores, probabilistic scores bounded on [0,1], and combined scores - and added advantage_timeline as a long-format view of advantage movement.
◆Where it's heading
The work is less about adding seasons than about making the tables join cleanly. 2.3.1 rebuilt challenge_description and challenge_results around a shared challenge_id, added challenge characteristic flags and result notes, and put logical finalist, winner and jury flags on castaways. Since then the pattern repeats at smaller scale: boot_order as its own table, season_name deprecated everywhere except season_summary, castaways cleaned so people booted twice appear once. Derived analytical columns are being separated from raw records rather than mixed into them.
◆Prediction
Expect the next release to add the current season's data on the same broadcast-following cadence, with any structural work continuing to split derived scores out of the raw tables.

◆Recent moves

  1. 2mo ago

    US50 and AU12 land, castaway scores split into three tiers

    Completes US50 and AU12 and adds advantage_timeline as a long-format companion to advantage_movement. The castaway_scores rework is the structural piece: standardised residual, probabilistic and combined scores are now named and separated rather than delivered as one opaque column set.

    View source ↗
  2. 7mo ago

    US49 data completed and season 50 cast added

    A straightforward broadcast-following release: US49 filled in once the season finished, plus cast details for the then-upcoming season 50. No schema movement, which is the norm between the larger restructuring releases.

    View source ↗
  3. 11mo ago

    Survivor Australia vs. The World added

    Adds a crossover series to the franchise coverage. Small in content but consistent with the package's stated scope of covering every version rather than the US show alone.

    View source ↗
  4. 1y ago

    boot_order table added and castaways deduplicated

    Adds boot_order as a dedicated table, overhauls castaway_scores, and removes people booted more than once from repeated rows in castaways. The season_name deprecation outside season_summary is the kind of normalisation that breaks existing scripts but stops the same label being maintained in six places.

    View source ↗
  5. 2y ago

    Complete AU09 data reaches CRAN

    A single-season data drop with no accompanying structural change. Routine for this package, and the shortest release note in the recent history.

    View source ↗
  6. 2y ago

    Challenge tables rebuilt around a shared linking key

    The largest structural release in the visible history: challenge_description and challenge_results were refreshed around a common challenge_id, gaining characteristic flags, challenge types and result notes, while castaways picked up finalist, winner and jury logicals and episodes gained episode_label. This is what made challenge-level analysis joinable, and the smaller normalisations since follow its lead.

    View source ↗