Claim at v1.1.0
bridge-heat-shootings-ems
5601865b26f27ab9d7c70b1d7f88178513e7a710272256e3b0863d6b496eea07
Data
{
"has_outcome": true,
"kind": "bridge",
"lineage": {
"audit": "docs/lineage_audit.md; data/derived/lineage_audit.json",
"class": "a",
"class_meaning": "exposure or treatment only, not the outcome",
"classes": [
"a",
"b"
],
"early_sources": [
{
"class": "b",
"dataset": "5ucz-vwe8",
"evidence": "NYPD shootings, cached from heat-shootings, build the chain's first link (the composed prediction), not the endpoint. The endpoint, EMS SHOT dispatches by day, was fetched at 16:08:40, after registration (referee).",
"fetched": "2026-11-12T15:21:22-05:00",
"name": "Shootings (2006-Present)"
},
{
"class": "b",
"dataset": "833y-fsy8",
"evidence": "NYPD historic shootings, cached from heat-shootings; input to the composed prediction, not the endpoint (EMS SHOT dispatches, fetched 16:08:40).",
"fetched": "2026-11-12T15:21:36-05:00",
"name": "ARCHIVED_NYPD Shooting Incident Data (Historic)"
},
{
"class": "b",
"dataset": "833y-fsy8",
"evidence": "NYPD historic shootings, cached from heat-shootings; input to the composed prediction, not the endpoint (EMS SHOT dispatches, fetched 16:08:40).",
"fetched": "2026-11-12T15:21:41-05:00",
"name": "ARCHIVED_NYPD Shooting Incident Data (Historic)"
},
{
"class": "a",
"dataset": "GHCN-Daily v3 (NCEI), doi:10.7289/V5D21VHZ",
"evidence": "Daily temperature (GHCN-Daily) is the exposure, not the outcome. It is weather, fetched for earlier heat claims and reused.",
"fetched": "2026-11-12T15:06:44-05:00"
}
]
},
"outcome": "bridge-heat-shootings-ems",
"prior": 0.6,
"registered_at": "2026-11-12T16:04:50-0500",
"registrant": "Claude (bridges agent), for Jeff Vernon",
"registration": {
"alternatives": [
"Fireworks around 4 July are reported as shots; holidays are in the model, and the warm-season sensitivity checks the rest.",
"EMS call-type coding changed over the period; year terms absorb level changes but not changes in seasonal coding.",
"Heat raises outdoor presence and hence the number of people who report shots, independent of shootings."
],
"chain": [
"heat-shootings",
"bridge-shot-ems-link"
],
"composed_prediction": {
"file": "data/derived/bridges/composed_predictions.json",
"hi": 1.1476,
"lo": 1.0494,
"mean_log": 0.09427432105922165,
"median": 1.099,
"scale": "IRR per 5 C of daily maximum temperature, Central Park",
"sd_log": 0.025144994166285625
},
"graph_evidence": [
{
"concepts": [
"temperature",
"injury"
],
"direction": "+",
"n_findings": 4,
"n_studies": 3,
"pubmed_ids": [
37448837
],
"relationship_id": "4FaXcuNJRt|rULjCt1BqH",
"summary": "Girma et al. 2023, NYC ED visits 2005-2011: intentional injury among 20-25 year olds OR 1.54 (1.28 to 1.85) for daily minimum 77 F vs 48 F. Used for direction only; it is an ED measure, not EMS, and not firearm-specific."
}
],
"id": "bridge-heat-shootings-ems",
"kind": "bridge",
"literature": [
{
"citation": "Reeping PM, Hemenway D (2020). The association between weather and the number of daily shootings in Chicago (2012-2016). Injury Epidemiology 7:31.",
"doi": "10.1186/s40621-020-00260-3",
"finding": "Shootings +34.0% (weekdays) and +41.6% (weekends) per 10 C."
}
],
"plan": "Daily counts 2006-01-01 to 2026-08-31 of 76xm-jjuj with initial_call_type = 'SHOT'. Weather and calendar terms built exactly as in analyses/heat/shootings.py (Central Park TMAX/5, day of week, month, year, holiday, rain >= 1 mm). Negative binomial GLM, alpha by ML, HC0 errors. Secondary, reported but not deciding: warm season May-September; 2006-2019 only; STAB dispatches as a contrast outside the chain; SHOT plus T-SHOT.",
"prediction": {
"direction": "+",
"magnitude": "Composed IRR 1.099 (1.049 to 1.148) per 5 C. Below the shootings IRR (1.146) because victim dispatches are only 25 to 50 percent of SHOT dispatches (NYPD recorded 974 shooting incidents in 2023; SHOT dispatches numbered 3,159 that year, a level seen before registration) and the other dispatches are assumed to follow temperature less, between not at all and as much as shootings.",
"refute_if": "|D| >= 1.96 SE. Verdict contested. An observed IRR above the composed range would mean that the non-victim dispatches follow heat more than shootings do; below it, that the chain overstates.",
"support_if": "With D = log(observed IRR) - mean_log(composed) and SE = sqrt(se_observed^2 + sd_log(composed)^2), |D| < 1.96 SE. Verdict confirmed."
},
"prior": 0.6,
"registered_at": "2026-11-12T16:04:50-0500",
"registrant": "Claude (bridges agent), for Jeff Vernon",
"sha256": "bda255d3ebffd6189ea84b6060c0e85675e809fee6bcbdbb17a4a6281ff11099",
"statement": "Composed from heat and shootings (heat-shootings, NYPD) and the share of EMS gunshot dispatches that are shooting victims (bridge-shot-ems-link), daily FDNY EMS dispatches for gunshot wounds rise with the day's maximum temperature by the composed rate ratio."
},
"registration_commit": {
"commit": "e783221b5d7cc4d6d50238ac6a32ba600bfd162b",
"committed_at": "2026-11-12T16:04:50-05:00"
},
"registration_file_sha256": "c880deea51a7d1e73ff43134e790890f5b9fb45df061727bdcfb42b4fe68e976",
"registration_path": "claims/registered/bridge-heat-shootings-ems.json",
"registration_sha256": "bda255d3ebffd6189ea84b6060c0e85675e809fee6bcbdbb17a4a6281ff11099",
"registration_text": "{\n \"id\": \"bridge-heat-shootings-ems\",\n \"kind\": \"bridge\",\n \"statement\": \"Composed from heat and shootings (heat-shootings, NYPD) and the share of EMS gunshot dispatches that are shooting victims (bridge-shot-ems-link), daily FDNY EMS dispatches for gunshot wounds rise with the day's maximum temperature by the composed rate ratio.\",\n \"graph_evidence\": [\n {\n \"relationship_id\": \"4FaXcuNJRt|rULjCt1BqH\",\n \"concepts\": [\n \"temperature\",\n \"injury\"\n ],\n \"n_findings\": 4,\n \"n_studies\": 3,\n \"direction\": \"+\",\n \"summary\": \"Girma et al. 2023, NYC ED visits 2005-2011: intentional injury among 20-25 year olds OR 1.54 (1.28 to 1.85) for daily minimum 77 F vs 48 F. Used for direction only; it is an ED measure, not EMS, and not firearm-specific.\",\n \"pubmed_ids\": [\n 37448837\n ]\n }\n ],\n \"literature\": [\n {\n \"citation\": \"Reeping PM, Hemenway D (2020). The association between weather and the number of daily shootings in Chicago (2012-2016). Injury Epidemiology 7:31.\",\n \"doi\": \"10.1186/s40621-020-00260-3\",\n \"finding\": \"Shootings +34.0% (weekdays) and +41.6% (weekends) per 10 C.\"\n }\n ],\n \"chain\": [\n \"heat-shootings\",\n \"bridge-shot-ems-link\"\n ],\n \"composed_prediction\": {\n \"scale\": \"IRR per 5 C of daily maximum temperature, Central Park\",\n \"median\": 1.099,\n \"lo\": 1.0494,\n \"hi\": 1.1476,\n \"mean_log\": 0.09427432105922165,\n \"sd_log\": 0.025144994166285625,\n \"file\": \"data/derived/bridges/composed_predictions.json\"\n },\n \"prediction\": {\n \"direction\": \"+\",\n \"magnitude\": \"Composed IRR 1.099 (1.049 to 1.148) per 5 C. Below the shootings IRR (1.146) because victim dispatches are only 25 to 50 percent of SHOT dispatches (NYPD recorded 974 shooting incidents in 2023; SHOT dispatches numbered 3,159 that year, a level seen before registration) and the other dispatches are assumed to follow temperature less, between not at all and as much as shootings.\",\n \"support_if\": \"With D = log(observed IRR) - mean_log(composed) and SE = sqrt(se_observed^2 + sd_log(composed)^2), |D| < 1.96 SE. Verdict confirmed.\",\n \"refute_if\": \"|D| >= 1.96 SE. Verdict contested. An observed IRR above the composed range would mean that the non-victim dispatches follow heat more than shootings do; below it, that the chain overstates.\"\n },\n \"plan\": \"Daily counts 2006-01-01 to 2026-08-31 of 76xm-jjuj with initial_call_type = 'SHOT'. Weather and calendar terms built exactly as in analyses/heat/shootings.py (Central Park TMAX/5, day of week, month, year, holiday, rain >= 1 mm). Negative binomial GLM, alpha by ML, HC0 errors. Secondary, reported but not deciding: warm season May-September; 2006-2019 only; STAB dispatches as a contrast outside the chain; SHOT plus T-SHOT.\",\n \"alternatives\": [\n \"Fireworks around 4 July are reported as shots; holidays are in the model, and the warm-season sensitivity checks the rest.\",\n \"EMS call-type coding changed over the period; year terms absorb level changes but not changes in seasonal coding.\",\n \"Heat raises outdoor presence and hence the number of people who report shots, independent of shootings.\"\n ],\n \"prior\": 0.6,\n \"registrant\": \"Claude (bridges agent), for Jeff Vernon\",\n \"registered_at\": \"2026-11-12T16:04:50-0500\",\n \"sha256\": \"bda255d3ebffd6189ea84b6060c0e85675e809fee6bcbdbb17a4a6281ff11099\"\n}",
"replication": {
"built_at": "2026-11-13T22:57:33-0500",
"checked_at": "2026-11-13T22:57:33-0500",
"decides": "endpoint IRR compatible with the composed prediction 1.099 (1.049 to 1.148), |z| < 1.96",
"deciding_number": "city_effect: IRR per 5 C of Central Park TMAX (lag 0) for daily FDNY EMS SHOT dispatches, 2006 to August 2026, negative binomial GLM as in heat-shootings, robust HC0 95% CI; and its compatibility z with the composed prediction",
"inputs_copied": 6,
"manifest": "replication/bridge-heat-shootings-ems/manifest.json",
"manifest_sha256": "de3b8eb3e59a4fcf1dc942a7489175424b8dfcfb1d06a3134fd077d1ea75dccb",
"package": "replication/bridge-heat-shootings-ems/",
"recorded": {
"compat_z": 0.4149437146358164,
"estimate": 1.1105984346713746,
"hi": 1.1211671770768419,
"lo": 1.100129319081887,
"n": 7548,
"stab_contrast": 1.0759499384976088
},
"status": "MATCH",
"status_from": "replication/bridge-heat-shootings-ems/verify_result.json",
"tables_not_rebuilt": 0,
"tolerance": {
"compat_z": 0.01,
"estimate": 0.0001,
"hi": 0.0001,
"lo": 0.0001,
"n": 0,
"stab_contrast": 0.0001
}
},
"statement": "Composed from heat and shootings (heat-shootings, NYPD) and the share of EMS gunshot dispatches that are shooting victims (bridge-shot-ems-link), daily FDNY EMS dispatches for gunshot wounds rise with the day's maximum temperature by the composed rate ratio."
}History in this collection
- addedv1.0.0Nov 14, 2026this version