Claim at v1.2.0

street-busway14-fhv-north-south

b7da1cc8bc8b2a7e5b2f623ab0a17945565656686551322f6789a38a63f1a092

Data

{
  "has_outcome": true,
  "kind": "confirmation",
  "lineage": {
    "audit": "docs/lineage_audit.md; data/derived/lineage_audit.json",
    "class": "b",
    "class_meaning": "a design, power or calibration read, pre-period data, or metadata",
    "classes": [
      "b"
    ],
    "early_sources": [
      {
        "class": "b",
        "dataset": "tlc-trip-records",
        "evidence": "Fetch time is a date only, read as midnight. The FHV design used February-September 2019. Referee: \"four of the five post-period files did not yet exist when the claim was registered\".",
        "fetched": "2026-11-13T00:00:00",
        "name": "TLC High Volume FHV Trip Records (monthly parquet)"
      }
    ]
  },
  "outcome": "street-busway14-fhv-north-south",
  "prereg_branch": {
    "name": "prereg/street-busway14-fhv-north-south",
    "sha": "0f85b3ad535a5ca0b118b088f50043bc41fec0d9"
  },
  "prior": 0.45,
  "registered_at": "2026-11-13T13:18:18-0500",
  "registrant": "Claude (agent, street-space task), 13 November 2026",
  "registration": {
    "alternatives": [
      "FHV trips differ from taxi trips in purpose and are routed by apps (Waze-like routing may use side streets differently).",
      "The FHV seasonal swing may differ from the yellow swing; the design measures month-level fleet differences only.",
      "Four north pairs carry most of the claim, so a local event in East Chelsea, Gramercy or Stuy Town (construction, the L train work on 14th Street stations) could produce it.",
      "The yellow swing used as baseline includes years before the January 2019 congestion surcharge."
    ],
    "data_seen_before_registration": [
      "FHV trips between the listed zones, February to September 2019 (design, data/derived/street_space/fhv_design.json).",
      "The FHV files for October 2019 to February 2020 were downloaded but not opened.",
      "Yellow-taxi data to February 2020, including the referee's north/south split (the hypothesis comes from it)."
    ],
    "design_file": "data/derived/street_space/fhv_design.json",
    "discrimination": {
      "D_north_if_estimate_at_prediction": 7.8,
      "fleet_tracking": {
        "north_corr": 0.547,
        "north_r": 0.01298,
        "south_corr": 0.9,
        "south_r": 0.00697
      },
      "north_se_log": 0.02417,
      "note": "D = 7.8 at the predicted +7% (north). At the low end of the range (+3%) D is 1.5, so a small true effect would most likely leave the claim unsettled. The fleet-difference term r is measured on February to September 2019 and cannot show a fleet difference in the winter-summer swing itself.",
      "placebo_yellow_2018_north_pct": -2.73,
      "placebo_yellow_2018_south_pct": 3.447,
      "predicted_log_north": 0.0677,
      "rule": 8,
      "south_se_log": 0.02189
    },
    "edge": {
      "cause": "busway",
      "effect": "door-to-door crosstown trip duration for trips exposed to 14th Street"
    },
    "graph_evidence": [
      {
        "concepts": "none",
        "direction": "none",
        "n_findings": 0,
        "summary": "As for street-busway14-crosstown-time: the graph holds no finding on road-space reallocation or displaced traffic."
      }
    ],
    "id": "street-busway14-fhv-north-south",
    "kind": "confirmation",
    "literature": [
      {
        "citation": "critique/street-busway14-crosstown-time.md (referee, 13 November 2026)",
        "doi": null,
        "finding": "Exploratory split of the yellow-taxi data, chosen after the data: the 4 north pairs +7.4% (+2.8 to +12.1), the southern pairs -1.5% (-5.5 to +2.7). This registration tests that split on data nobody has read."
      },
      {
        "citation": "NYC DOT / Sam Schwartz Engineering preliminary report, December 2019",
        "doi": null,
        "finding": "17th Street, north of 14th, +3.4 minutes from 3rd to 9th Avenue at weekday peaks; other side streets 0-2 minutes."
      }
    ],
    "plan": "analyses/street_space/fhv.py (committed before this registration). Data: TLC fhvhv_tripdata February 2019 to February 2020 (~/opt/nyc-concordance/fhvhv_1920/, from d37ci6vzurychx.cloudfront.net/trip-data/), fields trip_time and trip_miles, shared matched rides dropped, same filter, zones, weekday 07-18 sample and pair rule as the yellow tests. E_fhv(g) = FHV winter-minus-summer swing of the contrast in 2019-20 minus the mean yellow swing of 2016-17 to 2018-19 for the same subgroup; se from the yellow year-over-year spread and the FHV-yellow month difference; 95% interval E +/- 1.96 se. No resampling. Secondary: log trip_miles; the pooled t14 group. Rules 1, 2 and 10.",
    "prediction": {
      "direction": "+",
      "magnitude": "North: E between +3% and +12%, central +7%. South: E near 0 (between -3% and +3%).",
      "refute_if": "North upper 95% bound below +7% (the predicted central lengthening is excluded), OR south lower 95% bound above +3%.",
      "support_if": "North E >= +3% AND north lower 95% bound above 0 AND south upper 95% bound below +5%."
    },
    "prior": 0.45,
    "registered_at": "2026-11-13T13:18:18-0500",
    "registrant": "Claude (agent, street-space task), 13 November 2026",
    "script_commit": "832efc5",
    "scripts": [
      "analyses/street_space/fhv.py"
    ],
    "sha256": "983488eaf0654b12927630f39477c6cc81d021550dc9f8e3ff80b38f049afe8d",
    "statement": "On high-volume for-hire trips (Uber, Lyft and others), weekday daytime trips between East Chelsea and Stuy Town or Gramercy (zones on both sides of Fifth Avenue north of 14th Street) lengthened by roughly 3% to 12% after the busway began, relative to Midtown and downtown crosstown pairs, while trips between the zones south of 14th Street did not."
  },
  "registration_commit": {
    "commit": "0f85b3ad535a5ca0b118b088f50043bc41fec0d9",
    "committed_at": "2026-11-13T13:18:18-05:00"
  },
  "registration_file_sha256": "d5208d0ebff27a3d1f12ef335d04212fd31647e68e32807a0e373dabada9f13f",
  "registration_path": "claims/registered/street-busway14-fhv-north-south.json",
  "registration_sha256": "983488eaf0654b12927630f39477c6cc81d021550dc9f8e3ff80b38f049afe8d",
  "registration_text": "{\n \"id\": \"street-busway14-fhv-north-south\",\n \"kind\": \"confirmation\",\n \"statement\": \"On high-volume for-hire trips (Uber, Lyft and others), weekday daytime trips between East Chelsea and Stuy Town or Gramercy (zones on both sides of Fifth Avenue north of 14th Street) lengthened by roughly 3% to 12% after the busway began, relative to Midtown and downtown crosstown pairs, while trips between the zones south of 14th Street did not.\",\n \"edge\": {\n  \"cause\": \"busway\",\n  \"effect\": \"door-to-door crosstown trip duration for trips exposed to 14th Street\"\n },\n \"graph_evidence\": [\n  {\n   \"concepts\": \"none\",\n   \"n_findings\": 0,\n   \"direction\": \"none\",\n   \"summary\": \"As for street-busway14-crosstown-time: the graph holds no finding on road-space reallocation or displaced traffic.\"\n  }\n ],\n \"literature\": [\n  {\n   \"citation\": \"critique/street-busway14-crosstown-time.md (referee, 13 November 2026)\",\n   \"doi\": null,\n   \"finding\": \"Exploratory split of the yellow-taxi data, chosen after the data: the 4 north pairs +7.4% (+2.8 to +12.1), the southern pairs -1.5% (-5.5 to +2.7). This registration tests that split on data nobody has read.\"\n  },\n  {\n   \"citation\": \"NYC DOT / Sam Schwartz Engineering preliminary report, December 2019\",\n   \"doi\": null,\n   \"finding\": \"17th Street, north of 14th, +3.4 minutes from 3rd to 9th Avenue at weekday peaks; other side streets 0-2 minutes.\"\n  }\n ],\n \"prediction\": {\n  \"direction\": \"+\",\n  \"magnitude\": \"North: E between +3% and +12%, central +7%. South: E near 0 (between -3% and +3%).\",\n  \"support_if\": \"North E >= +3% AND north lower 95% bound above 0 AND south upper 95% bound below +5%.\",\n  \"refute_if\": \"North upper 95% bound below +7% (the predicted central lengthening is excluded), OR south lower 95% bound above +3%.\"\n },\n \"discrimination\": {\n  \"rule\": 8,\n  \"north_se_log\": 0.02417,\n  \"south_se_log\": 0.02189,\n  \"predicted_log_north\": 0.0677,\n  \"D_north_if_estimate_at_prediction\": 7.8,\n  \"placebo_yellow_2018_north_pct\": -2.73,\n  \"placebo_yellow_2018_south_pct\": 3.447,\n  \"fleet_tracking\": {\n   \"north_corr\": 0.547,\n   \"south_corr\": 0.9,\n   \"north_r\": 0.01298,\n   \"south_r\": 0.00697\n  },\n  \"note\": \"D = 7.8 at the predicted +7% (north). At the low end of the range (+3%) D is 1.5, so a small true effect would most likely leave the claim unsettled. The fleet-difference term r is measured on February to September 2019 and cannot show a fleet difference in the winter-summer swing itself.\"\n },\n \"plan\": \"analyses/street_space/fhv.py (committed before this registration). Data: TLC fhvhv_tripdata February 2019 to February 2020 (~/opt/nyc-concordance/fhvhv_1920/, from d37ci6vzurychx.cloudfront.net/trip-data/), fields trip_time and trip_miles, shared matched rides dropped, same filter, zones, weekday 07-18 sample and pair rule as the yellow tests. E_fhv(g) = FHV winter-minus-summer swing of the contrast in 2019-20 minus the mean yellow swing of 2016-17 to 2018-19 for the same subgroup; se from the yellow year-over-year spread and the FHV-yellow month difference; 95% interval E +/- 1.96 se. No resampling. Secondary: log trip_miles; the pooled t14 group. Rules 1, 2 and 10.\",\n \"alternatives\": [\n  \"FHV trips differ from taxi trips in purpose and are routed by apps (Waze-like routing may use side streets differently).\",\n  \"The FHV seasonal swing may differ from the yellow swing; the design measures month-level fleet differences only.\",\n  \"Four north pairs carry most of the claim, so a local event in East Chelsea, Gramercy or Stuy Town (construction, the L train work on 14th Street stations) could produce it.\",\n  \"The yellow swing used as baseline includes years before the January 2019 congestion surcharge.\"\n ],\n \"data_seen_before_registration\": [\n  \"FHV trips between the listed zones, February to September 2019 (design, data/derived/street_space/fhv_design.json).\",\n  \"The FHV files for October 2019 to February 2020 were downloaded but not opened.\",\n  \"Yellow-taxi data to February 2020, including the referee's north/south split (the hypothesis comes from it).\"\n ],\n \"design_file\": \"data/derived/street_space/fhv_design.json\",\n \"scripts\": [\n  \"analyses/street_space/fhv.py\"\n ],\n \"prior\": 0.45,\n \"registrant\": \"Claude (agent, street-space task), 13 November 2026\",\n \"script_commit\": \"832efc5\",\n \"registered_at\": \"2026-11-13T13:18:18-0500\",\n \"sha256\": \"983488eaf0654b12927630f39477c6cc81d021550dc9f8e3ff80b38f049afe8d\"\n}",
  "statement": "On high-volume for-hire trips (Uber, Lyft and others), weekday daytime trips between East Chelsea and Stuy Town or Gramercy (zones on both sides of Fifth Avenue north of 14th Street) lengthened by roughly 3% to 12% after the busway began, relative to Midtown and downtown crosstown pairs, while trips between the zones south of 14th Street did not."
}

History in this collection

  1. addedv1.0.0Nov 14, 2026this version