Claim at v1.2.0

street-busway14-crosstown-time

f53914c2a27adc340bdb953b2ea90c0cca0b1e4323dbb13dcafa6d93f0fa3cbc

Data

{
  "has_outcome": true,
  "kind": "conflict",
  "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 (2026-11-13), read as midnight. The pre-busway trips (July 2016 to September 2019) were aggregated for the design run before registration. The October 2019 to February 2020 files \"were downloaded but not opened or read\" (data_seen). Referee: the post-period aggregate was written at 12:06:20, after the 12:05:53 commit, and file access times match; \"A last-access time records only the latest read\".",
        "fetched": "2026-11-13T00:00:00",
        "name": "TLC Yellow Taxi Trip Records (monthly parquet)"
      }
    ]
  },
  "outcome": "street-busway14-crosstown-time",
  "prereg_branch": {
    "name": "prereg/street-busway14-crosstown-time",
    "sha": "d865ceb299eae96b255e90dbb9a722ae20f50703"
  },
  "prior": 0.35,
  "registered_at": "2026-11-13T12:05:53-0500",
  "registrant": "Claude (agent, street-space task), 13 November 2026",
  "registration": {
    "alternatives": [
      "Year-to-year drift of the contrast between the 14th Street corridor and the control corridors (seen in the design run: +3.4% from 2017-18 to 2018-19 in the plain year-over-year DiD), which the summer-netted estimator is meant to remove.",
      "The L train weeknight and weekend slowdown (27 April 2019 to 26 April 2020), present in both July-September 2019 and October 2019-February 2020; the primary sample is weekday daytime when the L ran normally.",
      "The M14 Select Bus Service (1 July 2019), present in both netted periods.",
      "Changes in which trips taxis serve (shorter or longer trips within a zone pair), which would move duration and distance together.",
      "Construction or events specific to one corridor in autumn 2019 (for example Hudson Yards in the control-adjacent West Chelsea zone, which is excluded from every group).",
      "October 2019 includes 1-2 October, before the busway, which dilutes the October month by about 6%."
    ],
    "data_seen_before_registration": [
      "All yellow-taxi trips between the listed zones from July 2016 to September 2019, aggregated (design run: data/derived/street_space/design.json, test_design.json, series_pre.csv).",
      "The parquet files for October 2019 to February 2020 were downloaded but not opened or read; no trip after 30 September 2019 has been aggregated or seen.",
      "From the press: DOT's December 2019 summary (side streets +0 to 2 minutes) and INRIX's first-week note, as listed in literature.",
      "ATVC (7ym2-wayt) row counts for 12th to 18th Streets 2017-2021 (counts only, no volumes): only four count locations, all in October 2019, so no volume test of evaporation is possible."
    ],
    "design_file": "data/derived/street_space/test_design.json",
    "discrimination": {
      "D_if_estimate_at_prediction": 8.38,
      "note": "With the design se of 1.69% (log), an estimate at the predicted +5% is 2.9 se from zero (D = 8.4 >= 4); a true effect of zero meets refute_if with probability about 0.83, and a true +5% meets support_if with probability about 0.82. Both placebo seasons (2017, 2018) lie inside +/-1.3%.",
      "placebo_E_pct": {
        "2017": -0.723,
        "2018": 1.254
      },
      "predicted_log": 0.0488,
      "rule": 8,
      "se_log_design": 0.01686
    },
    "draws": 2000,
    "edge": {
      "cause": "busway (removal of through car traffic from 14th Street)",
      "effect": "door-to-door crosstown taxi trip duration"
    },
    "graph_evidence": [
      {
        "concepts": "traffic volume, bus lane, pedestrian street, road closure, travel time, congestion",
        "direction": "none on this link",
        "n_findings": 4,
        "pubmed_ids": [
          14743825,
          17824287,
          24357516
        ],
        "summary": "System Public Graph variable searches (data/derived/street_space/graph_evidence.json, graph_detail.json). No variable for busway, transit priority, road-space reallocation, traffic evaporation or displaced traffic; 'road closure' and 'street closure' match nothing relevant; 'travel time' matches travel to health care. Nearest findings, abstracts read: Peace Bridge traffic fell 50% after 11 September 2001 with respiratory admissions nearly significant (PMID 14743825; the graph stores the 50% as a mean difference of 50); Busan traffic control in 2002 lowered pollutants 1-25%, childhood asthma RR 0.73 (0.49-1.11) (PMID 17824287); bus stops on a bus-only lane in Lima, pedestrian collision OR 2.36 (1.02-5.42) (PMID 24357516). None bears on travel time or displacement."
      }
    ],
    "id": "street-busway14-crosstown-time",
    "kind": "conflict",
    "literature": [
      {
        "citation": "Cairns S, Atkins S, Goodwin P (2002). Disappearing traffic? The story so far. Proceedings of the ICE - Municipal Engineer 151(1):13-22.",
        "doi": "10.1680/muen.2002.151.1.13",
        "finding": "Over 70 case studies of road-space reallocation in eleven countries; predicted traffic problems were often not realised and overall traffic often fell (secondary summaries give a mean reduction near 16-25% on treated roads or areas; not re-read)."
      },
      {
        "citation": "NYC DOT / Sam Schwartz Engineering, 14th Street Transit & Truck Priority Pilot preliminary report (fall 2019), as summarised by the Mayor's Office and Streetsblog, 18 December 2019.",
        "doi": null,
        "finding": "Bus travel times on 14th Street fell 22-47%; car travel times on 12th to 19th Streets rose 0-2 minutes in weekday peaks, 17th Street 3.4 minutes from 3rd to 9th Avenue; most avenues changed by under 1 minute, 3rd Avenue +1.6 minutes. Probe-vehicle runs on single streets, not door-to-door trips."
      },
      {
        "citation": "INRIX blog, first week of the 14th Street busway (October 2019).",
        "doi": null,
        "finding": "No discernible change in performance on neighbouring roads in the first week."
      }
    ],
    "plan": "Data: TLC yellow-taxi trip records (monthly parquet, d37ci6vzurychx.cloudfront.net/trip-data/yellow_tripdata_YYYY-MM.parquet), July 2016 to February 2020, held in ~/opt/nyc-concordance/tlc_yellow_1720/. Scripts: analyses/street_space/common.py (zones, windows), extract.py (aggregation by directed pair, date and hour; filter 0<mi<=50, 60 s to 3 h, 0.5-50 mph), series.py (monthly contrast c_m: treated pairs minus control pairs of pair-demeaned mean log outcome; pairs with at least 20 trips in every month), test.py (the estimator E and its standard error), all committed before this registration. Treated: directed pairs between {Meatpacking/West Village West 158, West Village 249, East Chelsea 68, Greenwich Village North 113} and {East Village 79, Stuy Town 224, Alphabet City 4, Gramercy 107}. Control: the same kind of east-west pairs in Midtown ({48, 50, 100, 230} with {170, 162, 233, 229}) and downtown ({231, 125, 211} with {148, 232, 144, 45}). Sample: weekdays except federal holidays, pickup hour 07 to 18. d_m = c_m - c_(m-12). E = mean d_m over October 2019 to February 2020 minus mean d_m over July to September 2019. se = s*sqrt(1/5+1/3), s = sd of d_m from July 2017 to September 2019; 95% interval E +/- 1.96 se; percent = 100(exp(.)-1). No resampling, so no seed enters the verdict. Verdict by PRECEDENCE rules 1 and 2; conditions are on interval bounds (rule 10). Secondary, not deciding: all-day busway hours (06-21) and overnight (22-05) samples; the plain year-over-year DiD with a day bootstrap (estimate.py, 2,000 draws, seed 20261113); per-pair changes; a monthly event-study figure. The pair set is fixed by the 20-trips-every-month rule over the whole input, so adding the post months may drop a pair and move the placebo values slightly; the outcome reports both. Outcome: log trip duration (s_lt).",
    "prediction": {
      "direction": "+",
      "magnitude": "E between +2% and +10%, central +5%. Reasoning: DOT's probe runs found 0-2 minutes added on parallel streets over 3rd to 9th Avenue (roughly 0-20% of a 10-minute peak run); a door-to-door taxi trip between these zones spends perhaps half its time on the crosstown leg, and some through traffic evaporates or shifts. The kind is 'conflict' because the evaporation literature (Cairns et al.) and DOT's summary lead to an expectation of little or no lengthening, and this registration bets against them.",
      "refute_if": "The upper 95% bound of E is below +5.0% (the predicted central lengthening is excluded).",
      "support_if": "E >= +2.0% AND the lower 95% bound of E is above 0."
    },
    "prior": 0.35,
    "registered_at": "2026-11-13T12:05:53-0500",
    "registrant": "Claude (agent, street-space task), 13 November 2026",
    "script_commit": "8b0917f",
    "scripts": [
      "analyses/street_space/common.py",
      "analyses/street_space/extract.py",
      "analyses/street_space/series.py",
      "analyses/street_space/test.py",
      "analyses/street_space/estimate.py"
    ],
    "seed": 20261113,
    "sha256": "7ce1f43da35120ec69043b550b2af31513667efa399735e6aaeb37cfb600b130",
    "statement": "After the 14th Street busway began on 3 October 2019, weekday daytime yellow-taxi trips between zones on the west and east sides of the 14th Street corridor lengthened by about 5% relative to comparable crosstown zone pairs in Midtown and downtown, net of the same comparison in the preceding summer."
  },
  "registration_commit": {
    "commit": "d865ceb299eae96b255e90dbb9a722ae20f50703",
    "committed_at": "2026-11-13T12:05:53-05:00"
  },
  "registration_file_sha256": "805bd298133944dc22a5824f68906a9b14e0db8a15263a2b04da7008c02ad0b4",
  "registration_path": "claims/registered/street-busway14-crosstown-time.json",
  "registration_sha256": "7ce1f43da35120ec69043b550b2af31513667efa399735e6aaeb37cfb600b130",
  "registration_text": "{\n \"id\": \"street-busway14-crosstown-time\",\n \"kind\": \"conflict\",\n \"statement\": \"After the 14th Street busway began on 3 October 2019, weekday daytime yellow-taxi trips between zones on the west and east sides of the 14th Street corridor lengthened by about 5% relative to comparable crosstown zone pairs in Midtown and downtown, net of the same comparison in the preceding summer.\",\n \"edge\": {\n  \"cause\": \"busway (removal of through car traffic from 14th Street)\",\n  \"effect\": \"door-to-door crosstown taxi trip duration\"\n },\n \"graph_evidence\": [\n  {\n   \"concepts\": \"traffic volume, bus lane, pedestrian street, road closure, travel time, congestion\",\n   \"n_findings\": 4,\n   \"direction\": \"none on this link\",\n   \"summary\": \"System Public Graph variable searches (data/derived/street_space/graph_evidence.json, graph_detail.json). No variable for busway, transit priority, road-space reallocation, traffic evaporation or displaced traffic; 'road closure' and 'street closure' match nothing relevant; 'travel time' matches travel to health care. Nearest findings, abstracts read: Peace Bridge traffic fell 50% after 11 September 2001 with respiratory admissions nearly significant (PMID 14743825; the graph stores the 50% as a mean difference of 50); Busan traffic control in 2002 lowered pollutants 1-25%, childhood asthma RR 0.73 (0.49-1.11) (PMID 17824287); bus stops on a bus-only lane in Lima, pedestrian collision OR 2.36 (1.02-5.42) (PMID 24357516). None bears on travel time or displacement.\",\n   \"pubmed_ids\": [\n    14743825,\n    17824287,\n    24357516\n   ]\n  }\n ],\n \"literature\": [\n  {\n   \"citation\": \"Cairns S, Atkins S, Goodwin P (2002). Disappearing traffic? The story so far. Proceedings of the ICE - Municipal Engineer 151(1):13-22.\",\n   \"doi\": \"10.1680/muen.2002.151.1.13\",\n   \"finding\": \"Over 70 case studies of road-space reallocation in eleven countries; predicted traffic problems were often not realised and overall traffic often fell (secondary summaries give a mean reduction near 16-25% on treated roads or areas; not re-read).\"\n  },\n  {\n   \"citation\": \"NYC DOT / Sam Schwartz Engineering, 14th Street Transit & Truck Priority Pilot preliminary report (fall 2019), as summarised by the Mayor's Office and Streetsblog, 18 December 2019.\",\n   \"doi\": null,\n   \"finding\": \"Bus travel times on 14th Street fell 22-47%; car travel times on 12th to 19th Streets rose 0-2 minutes in weekday peaks, 17th Street 3.4 minutes from 3rd to 9th Avenue; most avenues changed by under 1 minute, 3rd Avenue +1.6 minutes. Probe-vehicle runs on single streets, not door-to-door trips.\"\n  },\n  {\n   \"citation\": \"INRIX blog, first week of the 14th Street busway (October 2019).\",\n   \"doi\": null,\n   \"finding\": \"No discernible change in performance on neighbouring roads in the first week.\"\n  }\n ],\n \"prediction\": {\n  \"direction\": \"+\",\n  \"magnitude\": \"E between +2% and +10%, central +5%. Reasoning: DOT's probe runs found 0-2 minutes added on parallel streets over 3rd to 9th Avenue (roughly 0-20% of a 10-minute peak run); a door-to-door taxi trip between these zones spends perhaps half its time on the crosstown leg, and some through traffic evaporates or shifts. The kind is 'conflict' because the evaporation literature (Cairns et al.) and DOT's summary lead to an expectation of little or no lengthening, and this registration bets against them.\",\n  \"support_if\": \"E >= +2.0% AND the lower 95% bound of E is above 0.\",\n  \"refute_if\": \"The upper 95% bound of E is below +5.0% (the predicted central lengthening is excluded).\"\n },\n \"discrimination\": {\n  \"rule\": 8,\n  \"se_log_design\": 0.01686,\n  \"predicted_log\": 0.0488,\n  \"D_if_estimate_at_prediction\": 8.38,\n  \"placebo_E_pct\": {\n   \"2017\": -0.723,\n   \"2018\": 1.254\n  },\n  \"note\": \"With the design se of 1.69% (log), an estimate at the predicted +5% is 2.9 se from zero (D = 8.4 >= 4); a true effect of zero meets refute_if with probability about 0.83, and a true +5% meets support_if with probability about 0.82. Both placebo seasons (2017, 2018) lie inside +/-1.3%.\"\n },\n \"plan\": \"Data: TLC yellow-taxi trip records (monthly parquet, d37ci6vzurychx.cloudfront.net/trip-data/yellow_tripdata_YYYY-MM.parquet), July 2016 to February 2020, held in ~/opt/nyc-concordance/tlc_yellow_1720/. Scripts: analyses/street_space/common.py (zones, windows), extract.py (aggregation by directed pair, date and hour; filter 0<mi<=50, 60 s to 3 h, 0.5-50 mph), series.py (monthly contrast c_m: treated pairs minus control pairs of pair-demeaned mean log outcome; pairs with at least 20 trips in every month), test.py (the estimator E and its standard error), all committed before this registration. Treated: directed pairs between {Meatpacking/West Village West 158, West Village 249, East Chelsea 68, Greenwich Village North 113} and {East Village 79, Stuy Town 224, Alphabet City 4, Gramercy 107}. Control: the same kind of east-west pairs in Midtown ({48, 50, 100, 230} with {170, 162, 233, 229}) and downtown ({231, 125, 211} with {148, 232, 144, 45}). Sample: weekdays except federal holidays, pickup hour 07 to 18. d_m = c_m - c_(m-12). E = mean d_m over October 2019 to February 2020 minus mean d_m over July to September 2019. se = s*sqrt(1/5+1/3), s = sd of d_m from July 2017 to September 2019; 95% interval E +/- 1.96 se; percent = 100(exp(.)-1). No resampling, so no seed enters the verdict. Verdict by PRECEDENCE rules 1 and 2; conditions are on interval bounds (rule 10). Secondary, not deciding: all-day busway hours (06-21) and overnight (22-05) samples; the plain year-over-year DiD with a day bootstrap (estimate.py, 2,000 draws, seed 20261113); per-pair changes; a monthly event-study figure. The pair set is fixed by the 20-trips-every-month rule over the whole input, so adding the post months may drop a pair and move the placebo values slightly; the outcome reports both. Outcome: log trip duration (s_lt).\",\n \"alternatives\": [\n  \"Year-to-year drift of the contrast between the 14th Street corridor and the control corridors (seen in the design run: +3.4% from 2017-18 to 2018-19 in the plain year-over-year DiD), which the summer-netted estimator is meant to remove.\",\n  \"The L train weeknight and weekend slowdown (27 April 2019 to 26 April 2020), present in both July-September 2019 and October 2019-February 2020; the primary sample is weekday daytime when the L ran normally.\",\n  \"The M14 Select Bus Service (1 July 2019), present in both netted periods.\",\n  \"Changes in which trips taxis serve (shorter or longer trips within a zone pair), which would move duration and distance together.\",\n  \"Construction or events specific to one corridor in autumn 2019 (for example Hudson Yards in the control-adjacent West Chelsea zone, which is excluded from every group).\",\n  \"October 2019 includes 1-2 October, before the busway, which dilutes the October month by about 6%.\"\n ],\n \"data_seen_before_registration\": [\n  \"All yellow-taxi trips between the listed zones from July 2016 to September 2019, aggregated (design run: data/derived/street_space/design.json, test_design.json, series_pre.csv).\",\n  \"The parquet files for October 2019 to February 2020 were downloaded but not opened or read; no trip after 30 September 2019 has been aggregated or seen.\",\n  \"From the press: DOT's December 2019 summary (side streets +0 to 2 minutes) and INRIX's first-week note, as listed in literature.\",\n  \"ATVC (7ym2-wayt) row counts for 12th to 18th Streets 2017-2021 (counts only, no volumes): only four count locations, all in October 2019, so no volume test of evaporation is possible.\"\n ],\n \"design_file\": \"data/derived/street_space/test_design.json\",\n \"scripts\": [\n  \"analyses/street_space/common.py\",\n  \"analyses/street_space/extract.py\",\n  \"analyses/street_space/series.py\",\n  \"analyses/street_space/test.py\",\n  \"analyses/street_space/estimate.py\"\n ],\n \"seed\": 20261113,\n \"draws\": 2000,\n \"prior\": 0.35,\n \"registrant\": \"Claude (agent, street-space task), 13 November 2026\",\n \"script_commit\": \"8b0917f\",\n \"registered_at\": \"2026-11-13T12:05:53-0500\",\n \"sha256\": \"7ce1f43da35120ec69043b550b2af31513667efa399735e6aaeb37cfb600b130\"\n}",
  "statement": "After the 14th Street busway began on 3 October 2019, weekday daytime yellow-taxi trips between zones on the west and east sides of the 14th Street corridor lengthened by about 5% relative to comparable crosstown zone pairs in Midtown and downtown, net of the same comparison in the preceding summer."
}

History in this collection

  1. addedv1.0.0Nov 14, 2026this version