INDEPENDENT MINDS. EXCELLENT TELEVISION.THE PREVIEW EDITION / SEPTEMBER 2026
SOURCE NOTEBOOK

TV Shows & Streaming Database — Deep Research Brief, Part 03: TV Data Model, Identifiers, and Metadata Standards

100 preserved extracts. Research planning document; requested work is not evidence that the work was completed.

← Research library
100 extracts
  1. NOTE 3622

    TV Shows & Streaming Database — Deep Research Brief, Part 03: TV Data Model, Identifiers, and Metadata Standards

    Conduct exhaustive research on how TV data is structured, identified, classified, and matched across the industry and open data projects: entities and relationships, episode ordering schemes, specials and multi-part episodes, localized titles, credits and characters, franchises and format adaptations, status lifecycles, worldwide content ratings, genre and mood taxonomies, external identifiers, schema.org markup, entity resolution, and provenance. The goal is a field-level, source-tested data model and canonical ID strategy that the website can implement directly, with documented edge cases and test sets that prove it handles real shows correctly.

  2. NOTE 3623

    1. How to research this part

    Treat named identifiers, standards versions, field names, and example IDs as leads to verify against current documentation.

  3. NOTE 3624

    1. How to research this part

    Minimum coverage: document the TV data model of at least 12 sources (IMDb, TMDB, TheTVDB, TVmaze, Trakt, Wikidata, EIDR, Gracenote, AniDB, AniList, MyAnimeList, MyDramaList); build an edge-case test set of at least 150 shows; collect at least 25 content-rating systems; produce an identifier crosswalk for at least 500 series.

  4. NOTE 3625

    1. How to research this part

    Prefer primary documentation: API references (TMDB v3/v4, TheTVDB v4, TVmaze, Trakt), IMDb non-commercial dataset docs, Wikidata property pages, EIDR data-model and registration guides, schema.org type pages, EBU/SMPTE and MovieLabs standards (Media Entertainment Core, MEC; Common Metadata), and national ratings bodies.

  5. NOTE 3626

    1. How to research this part

    For each claim about a source's model, record the endpoint or dump field that proves it, and the access date.

  6. NOTE 3627

    1. How to research this part

    Do not copy proprietary schemas or data wholesale; describe structure and cite documentation.

  7. NOTE 3628

    2. Core entities and relationships

    Research how each source models these, then recommend a canonical model:

  8. NOTE 3629

    2. Core entities and relationships

    Series (show), Season (and part/volume/chapter/"book"), Episode, Special, Segment (for anthologies and sketch shows), Edit/Version (broadcast cut, extended cut, censored edit, streaming re-edit).

  9. NOTE 3630

    2. Core entities and relationships

    Person, Credit (person × work × role × episode range), Character, CharacterPortrayal (character × person × work × age/era, e.g., young and old versions).

  10. NOTE 3631

    2. Core entities and relationships

    Organization: network, platform, production company, distributor, studio, format owner.

  11. NOTE 3632

    2. Core entities and relationships

    Franchise/Universe, Relationship (spin-off, sequel, prequel, reboot, revival, remake, format adaptation, crossover, backdoor pilot, companion aftershow).

  12. NOTE 3633

    2. Core entities and relationships

    SourceWork (novel, comic, manga, light novel, webtoon, game, podcast, film, format).

  13. NOTE 3634

    2. Core entities and relationships

    Award, Nomination, RatingCertificate, Availability, Asset (image, trailer), ChartEntry, StatusEvent.

  14. NOTE 3635

    2. Core entities and relationships

    Question: which sources model "part" and "volume" as seasons vs. as sub-season groupings? Worked example (verify): Stranger Things 4 Vol. 1/2, Money Heist Part 1–5 (Netflix parts vs. Antena 3 original two seasons), Better Call Saul S6 split, Attack on Titan Final Season Parts 1–3 plus "The Final Chapters" specials.

  15. NOTE 3636

    3. Episode ordering schemes

    Document supported orders per source: TheTVDB (Aired, DVD, Absolute, Alternate, Regional, Alternate DVD; verify list), TMDB (default seasons plus "episode groups" types: original air date, absolute, DVD, digital, story arc, production, TV; verify), TVmaze (single order), Trakt (follows TMDB since 2024 migration; verify), IMDb (season/episode numbers only), AniDB (per-cour entries with specials "S", credits "C", trailers "T", parodies "P", other "O").

  16. NOTE 3637

    3. Episode ordering schemes

    Worked examples to verify and store as tests:

  17. NOTE 3638

    3. Episode ordering schemes

    Firefly — broadcast order (Fox, 2002) vs. intended/DVD order; pilot "Serenity" aired last.

  18. NOTE 3639

    3. Episode ordering schemes

    Futurama — production seasons vs. broadcast seasons across Fox, Comedy Central, and Hulu revival (2023).

  19. NOTE 3640

    3. Episode ordering schemes

    Doctor Who — 1963 serials, 2005 series, 2023 specials, and 2024 "Season 1" renumbering on Disney+ (verify).

  20. NOTE 3641

    3. Episode ordering schemes

    One Piece — absolute episode numbers (1000+) vs. arc seasons on streaming services.

  21. NOTE 3642

    3. Episode ordering schemes

    Rick and Morty, The Simpsons — production codes.

  22. NOTE 3643

    3. Episode ordering schemes

    Community, Arrested Development S4 remix "Fateful Consequences" (2018) as alternate edit.

  23. NOTE 3644

    3. Episode ordering schemes

    Love, Death & Robots — Netflix's per-user episode ordering experiments (verify).

  24. NOTE 3645

    3. Episode ordering schemes

    Produce: an ordering table with scheme, source, mapping to canonical episode IDs, and conflict notes.

  25. NOTE 3646

    4. Specials, pilots, multi-part, and unaired episodes

    Specials: holiday specials, reunion specials (Friends: The Reunion), aftershows, recaps, clip shows, behind-the-scenes, webisodes, minisodes. Where do sources place them (Season 0 in TMDB/TheTVDB; separate titles in IMDb)?

  26. NOTE 3647

    4. Specials, pilots, multi-part, and unaired episodes

    Multi-part episodes: two-part episodes aired as one on streaming (or the reverse). Test cases (verify): Lost two-hour finales counted as one or two episodes; Sherlock feature-length episodes; broadcast double-episodes split for syndication.

  27. NOTE 3648

    4. Specials, pilots, multi-part, and unaired episodes

    Unaired and unreleased: unaired pilots (Game of Thrones original pilot), pulled episodes (Community S2 "Advanced Dungeons & Dragons" blackface scene removed; verify), completed seasons shelved or pulled for tax write-offs (e.g., HBO Max removals, 2022–2023; verify titles), banned episodes (Pokémon "Dennō Senshi Porygon").

  28. NOTE 3649

    4. Specials, pilots, multi-part, and unaired episodes

    Fields: episodekind (regular, special, pilot, unairedpilot, recap, clip, webisode, minisode, extended, bonus), partsof (for merged), splitinto (for split), releasestatus (aired, released, unaired, withdrawn, lost).

  29. NOTE 3650

    5. Titles and localization

    Title types: original title (native script), romanized title (with system, e.g., Revised Romanization, Hepburn, Pinyin), English title, localized titles per territory and language, working titles, alternate/marketing titles, series vs. season subtitles (Cobra Kai, American Horror Story: Coven).

  30. NOTE 3651

    5. Titles and localization

    Test cases (verify): La Casa de Papel / Money Heist; Shingeki no Kyojin / Attack on Titan; Bron/Broen / The Bridge (and remakes); Kaun Banega Crorepati; Dark (German); Squid Game / Ojingeo Geim and Korean vs. Netflix romanization; Ted Lasso titles in Latin American Spanish vs. Spain Spanish.

  31. NOTE 3652

    5. Titles and localization

    Episode title localization: per-language episode titles, missing translations, and TMDB/TheTVDB translation completeness by language.

  32. NOTE 3653

    5. Titles and localization

    Languages: BCP 47 tags (e.g., pt-BR, es-419, zh-Hant-TW), original language(s), spoken languages, dub/subtitle tracks per availability record.

  33. NOTE 3654

    5. Titles and localization

    Fields: titleid, text, language, script (ISO 15924), territory, titletype, isprimary, romanizationsystem, source, validfrom, validto.

  34. NOTE 3655

    6. People, characters, and credits

    Credit granularity: series-level "starring" vs. per-episode credits vs. "also starring" and guest credits; uncredited roles; voice roles by language (Japanese seiyū vs. English dub cast); crew roles (creator, developer, showrunner, executive producer, writer, story editor, director, composer). Showrunner is often not a formal credit; how do sources record it?

  35. NOTE 3656

    6. People, characters, and credits

    Test cases (verify): Doctor Who Doctors as one character with many portrayals; The Crown recasting by era; This Is Us child/teen/adult actors; Bojack Horseman dual role; anime dub cast across Funimation and Crunchyroll releases.

  36. NOTE 3657

    6. People, characters, and credits

    Person identity edge cases: stage names, name changes, transgender performers' current names (show current name; do not display deadnames; research how IMDb updated its policy in 2019–2020; verify), people with the same name (IMDb nm disambiguation), deceased persons, privacy requests for birth dates.

  37. NOTE 3658

    6. People, characters, and credits

    Credit fields: creditid, personid, workid (series/season/episode), rolecategory (cast, voice, guest, crew), job (controlled vocabulary), characterid, creditedas, billingorder, episodecount (derived), firstepisodeid, lastepisodeid, languageofperformance, isuncredited, source, verifiedby.

  38. NOTE 3659

    6. People, characters, and credits

    Controlled job vocabulary: compare TMDB departments/jobs, EIDR contributor roles, MovieLabs roles, and IMDb categories.

  39. NOTE 3660

    7. Franchises, spin-offs, remakes, and format adaptations

    Relationship types and direction: spinoffof, sequelto, prequelto, rebootof, revivalof (same continuity, e.g., Twin Peaks: The Return), remakeof, adaptationofformat, crossoverwith, continuationundernewtitle (Sabrina vs. Chilling Adventures; Cobra Kai from films), sharesuniverse (Arrowverse, Chicago franchise, Yellowstone universe, Taylor Sheridan shows, One Chicago, FBI).

  40. NOTE 3661

    7. Franchises, spin-offs, remakes, and format adaptations

    Format adaptations (unscripted): Big Brother, Survivor (Expedition Robinson), The Masked Singer (from King of Mask Singer), Love Island, Got Talent, The Office (UK → U.S., India, Germany's Stromberg (verify legal relationship)), Skins, Shameless, The Bridge. Record format owner (Banijay, Fremantle, ITV Studios, BBC Studios, Sony) and local version as separate series linked by format.

  41. NOTE 3662

    7. Franchises, spin-offs, remakes, and format adaptations

    Continuations with changed titles and networks: Brooklyn Nine-Nine Fox → NBC; Lucifer Fox → Netflix; The Expanse Syfy → Prime Video; Arrested Development Fox → Netflix; Futurama; Manifest NBC → Netflix; Family Guy. Should these be one series record with networkhistory? Recommend yes, and document how sources differ.

  42. NOTE 3663

    8. Status lifecycle

    Statuses: rumored, indevelopment, pilotordered, seriesordered, preproduction, filming, postproduction, scheduled, airing/returning, onhiatus, renewed, ended (planned finale), cancelled, cancelledafterfilming (unaired), revived, inlimbo.

  43. NOTE 3664

    8. Status lifecycle

    Store status as dated events with source and confidence, not one mutable field. Test cases (verify): 1899 cancelled after S1; Warrior Nun cancelled then revival announced; Westworld; Mindhunter on indefinite hold; Lockwood & Co.; The OA; Lucifer revived by Netflix; Timeless cancelled, revived, ended with film.

  44. NOTE 3665

    8. Status lifecycle

    Map source vocabularies: TMDB (Returning Series, Planned, In Production, Ended, Canceled, Pilot), TVmaze (Running, Ended, To Be Determined, In Development), TheTVDB statuses (verify).

  45. NOTE 3666

    9. Content ratings worldwide

    Collect for at least 25 systems: U.S. TV Parental Guidelines (TV-Y, TV-Y7, TV-Y7-FV, TV-G, TV-PG, TV-14, TV-MA; descriptors D, L, S, V, FV), MPA (for TV movies), BBFC (U, PG, 12, 12A, 15, 18, R18), Ofcom broadcast watershed (not a rating), FSK (0, 6, 12, 16, 18), ACB Australia (G, PG, M, MA15+, R18+), NZ OFLC, CBFC India (U, UA 7+, UA 13+, UA 16+, A; 2024 amendments; verify) and IT Rules OTT categories, Canada (Canadian TV Classification; Québec Régie), France CSA/Arcom (-10, -12, -16, -18), Spain ICAA/CNMC, Italy, Netherlands Kijkwijzer (AL, 6, 9, 12, 14, 16, 18; verify), Nordics (Medietilsynet, Statens medieråd, KAVI), Brazil ClassInd (L, 10, 12, 14, 16, 18), Mexico RTC, Korea KMRB (All, 7, 12, 15, 19), Japan (broadcast self-regulation, EIRIN for film), Singapore IMDA, Philippines MTRCB, Saudi GCAM, UAE, Netflix's own maturity ratings where no local rating exists.

  46. NOTE 3667

    9. Content ratings worldwide

    Fields: ratingsystemid, jurisdiction, value, descriptors[], ratedwork (series/season/episode), ratedby (official body, platform self-rating, broadcaster), certificatenumber, date, source.

  47. NOTE 3668

    9. Content ratings worldwide

    Question: which platforms apply ratings at episode level (Netflix displays episode-specific advisories in some markets; verify)?

  48. NOTE 3669

    10. Genre, mood, and theme taxonomies

    Compare: TMDB TV genres (e.g., Action & Adventure, Sci-Fi & Fantasy, Kids, News, Reality, Soap, Talk, War & Politics), TVmaze genres, IMDb genres, TheTVDB genres, Gracenote genre and "Video Descriptors" (mood, theme, setting; verify naming), Netflix internal "altgenres" (reported by The Atlantic, 2014) as a model not a source, EBU ContentCS, Wikidata genre items, AniList tags (with spoiler flags and rank), MyDramaList tags, TV Tropes (as reference; licensing CC BY-NC-SA; verify).

  49. NOTE 3670

    10. Genre, mood, and theme taxonomies

    Design facets: format (sitcom, procedural, anthology, docuseries, competition, talk, soap, telenovela, microdrama), genre, subgenre, tone (dark, cozy, satirical), themes (found family, heist), setting (time/place), audience (kids, teens, family), pacing, content advisories (not mood tags), and spoiler-sensitive tags.

  50. NOTE 3671

    10. Genre, mood, and theme taxonomies

    Produce: a proposed taxonomy of at least 400 tags with definitions, parent-child relations, mappings to TMDB/TVmaze/AniList genres, and spoiler flags.

  51. NOTE 3672

    11. Identifiers and crosswalk

    For each identifier: format, scope (series/season/episode/person), assignment authority, public availability, license, stability, and merge/redirect behavior.

  52. NOTE 3673

    11. Identifiers and crosswalk

    IMDb: tt + 7–8 digits for titles and episodes; nm for names; episodes linked via title.episode.tsv (parentTconst, seasonNumber, episodeNumber). Record redirect handling on merges.

  53. NOTE 3674

    11. Identifiers and crosswalk

    TMDB: numeric TV series ID; seasons and episodes addressed by seriesid + seasonnumber + episodenumber, plus internal episode IDs; externalids endpoint returns IMDb, TVDB, Wikidata, etc. (verify list).

  54. NOTE 3675

    11. Identifiers and crosswalk

    TheTVDB: numeric series IDs (v4 also uses slugs); episode IDs; supports remote IDs.

  55. NOTE 3676

    11. Identifiers and crosswalk

    TVmaze: numeric show and episode IDs; lookup by IMDb/TVDB.

  56. NOTE 3677

    11. Identifiers and crosswalk

    Trakt: trakt ID, slug, and linked ids (imdb, tmdb, tvdb).

  57. NOTE 3678

    11. Identifiers and crosswalk

    Wikidata: QIDs; properties P345 (IMDb ID), P4983 (TMDB TV series ID), P4835 (TheTVDB series ID), P8600 (TVmaze series ID) (verify all property numbers); instance of Q5398426 (television series) (verify).

  58. NOTE 3679

    11. Identifiers and crosswalk

    EIDR: DOI-format 10.5240/XXXX-XXXX-XXXX-XXXX-XXXX-C; hierarchy Series → Season → Episode → Edit → Manifestation; alternate IDs.

  59. NOTE 3680

    11. Identifiers and crosswalk

    Gracenote TMS IDs: prefixes SH (series), EP (episode), MV (movie), SP (sports) with 8-digit root and 4-digit suffix (verify).

  60. NOTE 3681

    11. Identifiers and crosswalk

    ISAN: 96-bit with version segment (V-ISAN); root/episode/version structure (verify).

  61. NOTE 3682

    11. Identifiers and crosswalk

    Anime: AniDB aid/eid, AniList ID, MyAnimeList ID (per season/cour), Kitsu ID; mapping projects such as anime-offline-database, Fribb anime-lists, and ScudLee's anime-lists (AniDB→TVDB) (verify repos and licenses).

  62. NOTE 3683

    11. Identifiers and crosswalk

    Asian drama: MyDramaList IDs; Viki container IDs.

  63. NOTE 3684

    11. Identifiers and crosswalk

    Platform-native: Netflix title IDs, Amazon ASINs (by territory), Apple umc.cmc. IDs, Disney+ entity IDs, Roku content IDs, Google Knowledge Graph MIDs (/m/ and /g/).

  64. NOTE 3685

    11. Identifiers and crosswalk

    Produce: crosswalk table (canonicalid, idtype, value, scope, sourceofmapping, confidence, firstseen, lastverified, status: active/redirected/deprecated) for at least 500 series with season/episode depth for 100.

  65. NOTE 3686

    12. schema.org and structured data

    Types: TVSeries, TVSeason, TVEpisode, CreativeWorkSeason, Person, PerformanceRole, Organization/BroadcastService, BroadcastEvent, WatchAction (potentialAction with target and actionAccessibilityRequirement for availability/offers), AggregateRating, VideoObject, ImageObject.

  66. NOTE 3687

    12. schema.org and structured data

    Key properties: partOfSeries, partOfSeason, episodeNumber, seasonNumber, numberOfSeasons, numberOfEpisodes, actor (with PerformanceRole and characterName), director, productionCompany, contentRating, inLanguage, countryOfOrigin, sameAs (link to Wikidata/IMDb), identifier (PropertyValue), startDate/endDate.

  67. NOTE 3688

    12. schema.org and structured data

    Question: which properties does Google currently use for TV features, and what did Google deprecate (verify changes to "Media Actions" partner program)?

  68. NOTE 3689

    12. schema.org and structured data

    Worked example to produce: a JSON-LD block for one episode of The Bear (verify details) including series, season, credits, identifiers, and a WatchAction.

  69. NOTE 3690

    13. Entity resolution

    Signals: exact external-ID matches; normalized title (Unicode NFKC, casefold, punctuation, article stripping, romanization); original title; first-air year ±1; country and network; episode count; creator/lead cast overlap; season/episode date sequences (compare air-date vectors to match orderings).

  70. NOTE 3691

    13. Entity resolution

    Methods to research: deterministic rules plus probabilistic scoring (Fellegi–Sunter), blocking (by year and normalized title tokens), embedding similarity for titles across languages, Splink or Dedupe libraries, human review queues for scores in an uncertain band, and graph-based transitive closure with conflict detection.

  71. NOTE 3692

    13. Entity resolution

    Hard cases for the test set (verify each): same-title different shows (The Office UK/U.S.; Shameless; House of Cards 1990 vs. 2013; Battlestar Galactica 1978/2004; Doctor Who eras); remakes with same title and year proximity; anime seasons as separate entries vs. one series; K-dramas with multiple English titles; episodes with swapped numbers between sources; shows whose IDs were merged or split upstream.

  72. NOTE 3693

    13. Entity resolution

    Produce: gold-standard test set of at least 150 series and 2,000 episode pairs with correct mappings, plus precision/recall targets (e.g., ≥99.5% precision on auto-merge) and an evaluation script plan.

  73. NOTE 3694

    14. Edit history and provenance

    Research how IMDb (contribution review), TMDB (edit history per field, locks), TheTVDB (history and moderation), Wikidata (revisions, references with P248 stated in, P813 retrieved), and MusicBrainz (edit voting) track changes.

  74. NOTE 3695

    14. Edit history and provenance

    Fields per assertion: value, sourceid, sourceurl, retrievedat, assertedby (user/import/editor), method (import, manual, AIsuggested), confidence, reviewstatus, supersededby, licenseofsource.

  75. NOTE 3696

    14. Edit history and provenance

    Recommend field-level provenance for contested facts (air dates, credits, status) and record-level for low-risk fields.

  76. NOTE 3697

    15. Full field-level schema (produce and validate)

    series: id, slug, primarytitleid, originallanguage[], origincountries[], format, kind (scripted, unscripted, animated, anime, docuseries, talk, soap, microdrama, web), firstairdate, lastairdate, statuscurrent (derived), episoderuntimeminutes (typical, range), networkhistory[] (orgid, territory, from, to, role: original/continuation), productioncompanies[], createdby[], sourceworkids[], franchiseids[], genres[], tags[], ratings[], externalids[], defaultorderingscheme, islimited, isanthology, totalseasons (derived), totalepisodes (derived), synopsisids[] (original writing, per language), createdat, updatedat.

  77. NOTE 3698

    15. Full field-level schema (produce and validate)

    season: id, seriesid, orderingscheme, number, kind (season, part, volume, series, cour, specialgroup), title, airdatefirst, airdatelast, episodecount, releasemodel (weekly, binge, split, batch), externalids[].

  78. NOTE 3699

    15. Full field-level schema (produce and validate)

    episode: id, seriesid, primaryseasonid, numberinseason, absolutenumber, productioncode, titleids[], episodekind, runtimeminutes, runtimewithads, releasestatus, airdates[] (territory, network/platform, datetime with time zone, ispremiere), directors/writers via credits, ratingcertificates[], partsof[], splitinto[], externalids[].

  79. NOTE 3700

    15. Full field-level schema (produce and validate)

    episodeordermap: episodeid, scheme, seasonnumber, episodenumber, source.

  80. NOTE 3701

    15. Full field-level schema (produce and validate)

    person, character, characterportrayal, credit, organization, relationship, title, statusevent, ratingcertificate, externalid, provenance — define each with keys, required fields, cardinality, and indexes.

  81. NOTE 3702

    16. Timeline 2019 – 2026

    API/model changes: TheTVDB v4 API (2020–2021), TMDB episode groups expansion, Trakt's move to TMDB metadata (verify date), IMDb dataset field changes, TVmaze license changes, Wikidata property creations, EIDR hierarchy updates, MovieLabs MEC releases.

  82. NOTE 3703

    16. Timeline 2019 – 2026

    Ratings system changes: CBFC UA sub-categories (2024), Kijkwijzer updates, India IT Rules (2021), Korean streaming self-rating (2023; verify).

  83. NOTE 3704

    16. Timeline 2019 – 2026

    Notable renumbering events: Doctor Who (2024), Disney+ Star Wars/Marvel season labeling (verify), anime re-seasoning on Crunchyroll.

  84. NOTE 3705

    17. Deliverables for Part 03

    Full field-level schema (at least 15 entities, 250 fields) with types, constraints, and source mappings.

  85. NOTE 3706

    17. Deliverables for Part 03

    Source data-model comparison across at least 12 sources.

  86. NOTE 3707

    17. Deliverables for Part 03

    Episode-ordering reference with at least 40 documented shows.

  87. NOTE 3708

    17. Deliverables for Part 03

    Edge-case test set: at least 150 series and 2,000 episode pairs with gold mappings.

  88. NOTE 3709

    17. Deliverables for Part 03

    Identifier reference for at least 18 ID systems and a crosswalk of at least 500 series.

  89. NOTE 3710

    17. Deliverables for Part 03

    Content-rating reference: at least 25 systems with values, descriptors, and authority.

  90. NOTE 3711

    17. Deliverables for Part 03

    Genre/tone/theme taxonomy: at least 400 tags with mappings.

  91. NOTE 3712

    17. Deliverables for Part 03

    Relationship-type vocabulary with at least 60 worked examples.

  92. NOTE 3713

    17. Deliverables for Part 03

    schema.org JSON-LD templates for series, season, episode, person.

  93. NOTE 3714

    17. Deliverables for Part 03

    Glossary of at least 80 terms; source list with documentation URLs and access dates.

  94. NOTE 3715

    Research standards

    State the research date and cite the documentation version or endpoint for each source-model claim.

  95. NOTE 3716

    Research standards

    Verify every identifier format and Wikidata property number against live documentation.

  96. NOTE 3717

    Research standards

    Mark example IDs and episode facts as test inputs to verify, not facts.

  97. NOTE 3718

    Research standards

    Record the license of every mapping dataset (e.g., CC0, CC BY-SA, ODbL) before recommending reuse.

  98. NOTE 3719

    Research standards

    Handle person data with privacy in mind; never recommend displaying deadnames or unverified birth dates.

  99. NOTE 3720

    Research standards

    Separate official ratings from platform self-ratings and user-contributed advisories.

  100. NOTE 3721

    Research standards

    Note where a lawyer should review database-right or license compatibility when combining sources.