Camera Data

In the NYC Open Data Portal, the camera locations are only available in yearly datasets. Because these datasets are so large, I had to filter, group by, and aggregate the data within the portal before uploading them to R in order to find the “implementation date” for each camera (so that we can properly mark crashes as pre-camera or post-camera). In order to get this information, I took each Parking Violations Issued dataset from. FY 2014 - 2026 and performed the following queries: Filter for “Violation Code = 36” (this is the violation code for school zone speed camera tickets), Group By “Street Name” and “Intersecting Street”, and Aggregate By “Issue Date”, where it provides the minimum issue date (in order to find the “first” issued violation from each camera, using that date as the implementation date). The uploaded datasets are the filtered, grouped, and aggregated versions of the Parking Violations Issued datasets for each FY from 2014 - 2026.

Loading & combining the speed violation datasets

camera_dates_2014 <- read.csv("~/Downloads/Parking_Violations_Issued_-_Fiscal_Year_2014_20260407.csv")
camera_dates_2015 <- read.csv("~/Downloads/Parking_Violations_Issued_-_Fiscal_Year_2015_20260407.csv")
camera_dates_2016 <- read.csv("~/Downloads/Parking_Violations_Issued_-_Fiscal_Year_2016_20260407.csv")
camera_dates_2017 <- read.csv("~/Downloads/Parking_Violations_Issued_-_Fiscal_Year_2017_20260407.csv")
camera_dates_2018 <- read.csv("~/Downloads/Parking_Violations_Issued_-_Fiscal_Year_2018_20260407.csv")
camera_dates_2019 <- read.csv("~/Downloads/Parking_Violations_Issued_-_Fiscal_Year_2019_20260407.csv")
camera_dates_2020 <- read.csv("~/Downloads/Parking_Violations_Issued_-_Fiscal_Year_2020_20260407.csv")
camera_dates_2021 <- read.csv("~/Downloads/Parking_Violations_Issued_-_Fiscal_Year_2021_20260407.csv")
camera_dates_2022 <- read.csv("~/Downloads/Parking_Violations_Issued_-_Fiscal_Year_2022_20260407.csv")
camera_dates_2023 <- read.csv("~/Downloads/Parking_Violations_Issued_-_Fiscal_Year_2023_20260407.csv")
camera_dates_2024 <- read.csv("~/Downloads/Parking_Violations_Issued_-_Fiscal_Year_2024_20260407.csv")
camera_dates_2025 <- read.csv("~/Downloads/Parking_Violations_Issued_-_Fiscal_Year_2025_20260407.csv")


camera_dates_2014 <- camera_dates_2014 %>% rename(Street.Name = Street)
camera_dates_2014 <- camera_dates_2014 %>% mutate(cross_street = paste(Street.Name, Intersecting.Street))
camera_dates_2015 <- camera_dates_2015 %>% mutate(cross_street = paste(Street.Name, Intersecting.Street))
camera_dates_2016 <- camera_dates_2016 %>% mutate(cross_street = paste(Street.Name, Intersecting.Street))
camera_dates_2017 <- camera_dates_2017 %>% mutate(cross_street = paste(Street.Name, Intersecting.Street))
camera_dates_2018 <- camera_dates_2018 %>% mutate(cross_street = paste(Street.Name, Intersecting.Street))
camera_dates_2019 <- camera_dates_2019 %>% mutate(cross_street = paste(Street.Name, Intersecting.Street))
camera_dates_2020 <- camera_dates_2020 %>% mutate(cross_street = paste(Street.Name, Intersecting.Street))
camera_dates_2021 <- camera_dates_2021 %>% mutate(cross_street = paste(Street.Name, Intersecting.Street))
camera_dates_2022 <- camera_dates_2022 %>% mutate(cross_street = paste(Street.Name, Intersecting.Street))
camera_dates_2023 <- camera_dates_2023 %>% mutate(cross_street = paste(Street.Name, Intersecting.Street))
camera_dates_2024 <- camera_dates_2024 %>% mutate(cross_street = paste(Street.Name, Intersecting.Street))
camera_dates_2025 <- camera_dates_2025 %>% mutate(cross_street = paste(Street.Name, Intersecting.Street))

# combining all the datasets
camera_dates_all <- bind_rows(camera_dates_2014, camera_dates_2015, camera_dates_2016, camera_dates_2017, camera_dates_2018, camera_dates_2019, camera_dates_2020, camera_dates_2021, camera_dates_2022, camera_dates_2023, camera_dates_2024, camera_dates_2025)

Preliminary location matching

Due to a very large (and messy) dataset, I tried to sort them first without fully cleaning to dwindle down easily identifiable duplicates and then work from there.

sort_letters <- function(x) {
  sapply(lapply(strsplit(x, ""), sort), paste, collapse = "")
}

df_fast1 <- camera_dates_all %>%
  mutate(
    clean = str_remove_all(tolower(paste0(Street.Name, Intersecting.Street)), "[^a-z0-9]"),
    letter_id = sort_letters(clean)
  )
# so this is sorting them by "letter id" (all its letters put into alphanetical order) to see which ones match up right away

df_deduplicated1 <- df_fast1 %>%
  mutate(
    clean_date = ymd_hms(min_issue_date)
  ) %>%
  group_by(letter_id) %>%
  arrange(clean_date, .by_group = TRUE) %>%
  slice(1) %>%
  ungroup()
# finding first instance of each camera

nrow(df_deduplicated1)
## [1] 4551
street_fixes <- c("CENTRAL P W" = "Central Park West", "GRND CNTRL PWY" = "Grand Central Parkway", "GRAND CE" = "Grand Central", "GR CENTRAL" = "Grand Central", "HRCE HRDNG XPY" = "Horace Harding Expressway","GR AND CONCOURSE" = "Grand Concourse", "GRNDCONCRSE." = "Grand Concourse", "GRNDCONCRSE" = "Grand Concourse", "GRAN D CONCOURSE" = "Grand Concourse", "GRND CONCRSE"  = "Grand Concourse","GRNDCO NCRSE"  = "Grand Concourse","GRN DCONCRSE" = "Grand Concourse", "BRNX RVR"= "Bronx River", "HRLM RVR" = "Harlem River", "CRSS" = "Cross", "CRO SS" = "Cross", "CROS S" = "Cross", "GOULDN" = "Goulden", "NORTHRN" = "Northern", "NORT HERN" = "Northern", "NORTHER N" = "Northern", "F RANCS LEWIS" = "Francis Lewis", "FRANCS LEW IS" = "Francis Lewis", "FRNCS LWS" = "Francis Lewis", "FR NCS LWS" = "Francis Lewis", "FRANCIS LWS" = "Francis Lewis", "FRANCS L"= "Francis Lewis", "BELMNT" = "Belmont", "BELM NT" = "Belmont", "B ELMNT"= "Belmont", "LEXINGTN" = "Lexington", "LEXI NGTON" = "Lexington", "LEXING TON" = "Lexington", "LEXNG TN" = "Lexington", "D YCKMAN" = "Dyckman", "DYC KMAN"= "Dyckman", "BDFD" = "Bedford", "BEDFRD"= "Bedford", "BDFORD"= "Bedford", "BED FORD"= "Bedford", "BE DFORD"= "Bedford", "SPRINGFIE LD" = "Springfield", "S PRINGFIELD" = "Springfield", "SPRNGFLD" = "Springfield", "SPRIN GFIELD" = "Springfield", "SPRING FIELD" = "Springfield", "SPRINGFIEL"= "Springfield", "MANH ATTAN" = "Manhattan", "GUY BREWER" = "Guy R Brewer", "PLEASNT" = "Pleasant", "PL EASANT" = "Pleasant", "MALCLM" = "Malcolm", "MLCOLM" = "Malcolm", "AMST ERDAM" = "Amsterdam", "AMSTRDM" = "Amsterdam", "AMSTERDM" = "Amsterdam", "AMSTRDAM" = "Amsterdam", "AMSTE RDAM" = "Amsterdam", "WOODHAVN" = "Woodhaven", "WO ODHVN" = "Woodhaven", "WOOD HVN" = "Woodhaven", "W OODHAVEN" = "Woodhaven", "WOODHVN" = "Woodhaven", "WOODHV N" = "Woodhaven", "WOODH AVEN"= "Woodhaven", "WOOD HAVEN"= "Woodhaven", "FARMRS" = "Farmers", "FA RMERS"= "Farmers", "YELLOWSTNE" = "Yellowstone", "WASHNGTN" = "Washington", "WASHINGTO N" = "Washington", "WSHNGTN"= "Washington", "WASH INGTON"= "Washington", "WASHI NGTON"= "Washington", "WA SHINGTON"= "Washington", "WAS HINGTON"= "Washington", "WAS HINTON"= "Washington", "FT WASHINGTON" = "Fort Washington", "FT." = "FORT", "RIVERSID E" = "Riverside", "HEMPSTD" = "Hempstead", "HEMP STD" = "Hempstead", "BCH" = "Beach", "B EACH"= "Beach", "PE NROD" = "Penrod", "LA S ALLE" = "Lasalle", "MORNINGSDE" = "Morningside", "SEAV IEW" = "Seaview", "SEAVW"  = "Seaview", "S EAVW" = "Seaview", "S EAVIEW"= "Seaview",  "ROOSEVLT" = "Roosevelt", "WARR N" = "Warren", "WARRN" = "Warren", "FREDERICK DOUGLAS S" = "Frederick Douglass", "FREDRCK DOUGLS" = "Frederick Douglass", "FRDCK DLGASS" = "Frederick Douglass", "FRED ERICK DOUGLAS" = "Frederick Douglass", "FRDCK DG LASS" = "Frederick Douglass", "FREDRCK DGLS" = "Frederick Douglass", "FRE DERICK D"  = "Frederick Douglass", "FREDERICK D" = "Frederick Douglass", "FLAT LANDS" = "Flatlands", "FLATLA NDS" = "Flatlands", "FLATLAN DS" = "Flatlands", "FL ATLANDS" = "Flatlands", "FLATLNDS"= "Flatlands", "FLTLNDS"= "Flatlands", "FLTL NDS" = "Flatlands","FLTLN DS"= "Flatlands", "S. CONDUIT" = "South Conduit", "S CO NDUT"= "South Conduit", "S CONDUT"= "South Conduit", "ROCKA WAY" = "Rockaway", "ROCKAWY" = "Rockaway", "ROCKWY" = "Rockaway", "ROCKAWA" = "Rockaway", "ROCK AWAY" = "Rockaway", "RO CKAWAY" = "Rockaway", "ROCKAWA Y" = "Rockaway", "ROC KAWAY" = "Rockaway", "EDWRD L GRNT" = "Edward L Grant", "WALTN" = "Walton", "PROSP ECT" = "Prospect", "PROSPEC T" = "Prospect", "PRO SPECT" = "Prospect", "PROS PECT" = "Prospect", "PROSPE CT"= "Prospect", "PR OSPECT"= "Prospect", "PRSPCT"= "Prospect", "TIBB ET" = "Tibbet", "BRIELL" = "Brielle", "BRIELL E"= "Brielle", "S VC" = "Service", "WHITESTNE" = "Whitestone", "MA LBA" = "Malba", "BROOK HAVEN" = "Brookhaven", "PAERD EGAT" = "Paerdegat", "FT HAMILTN" = "Fort Hamilton", "FT. HAMILTN" = "Fort Hamilton", "NOR THRN" = "Northern", "HA MILTON" = "Hamilton", "HAMILT ON"= "Hamilton", "METROPOLTN" = "Metropolitan", "MTRPOLTN" = "Metropolitan", "MTROPLTN" = "Metropolitan", "METROPO LITAN" = "Metropolitan",  "METROPLTN"= "Metropolitan", "VA N HORN" = "Van Horn", "NURG E" = "Nurge", "RE NTAR" = "Rentar", "MADISN" = "Madison", "MADI SON"= "Madison", "ADM CLAYTN PWL JR" = "Adam Clayton Powell Jr", "ADAM CLAYTN PWL J R" = "Adam Clayton Powell Jr", "ADAM CLAYTN PWL JR" = "Adam Clayton Powell Jr", "ADAM C POWELL JR" = "Adam Clayton Powell Jr", "WADSORTH" = "Wadsworth", "POWE LL" = "Powell", "BWAY" = "Broadway", "BROA DWAY" = "Broadway", "BROADWA Y" = "Broadway", "BROAD WAY" = "Broadway", "BROADW AY" = "Broadway",  "BRDW AY" = "Broadway", "BRDWAY" = "Broadway", "LIN DEN" = "Linden", "LINDE N" = "Linden", "L INDEN"= "Linden", "LI NDEN"= "Linden", "N ICHOLAS" = "Nicholas", "NICHLS" = "Nicholas", "NICH LS" = "Nicholas", "NI CHOLAS"= "Nicholas",  "NIC HOLAS"= "Nicholas", "NICHLAS"= "Nicholas", "NICH LAS"= "Nicholas", "MELR OSE" = "Melrose", "BRNXWOOD" = "Bronxwood", "BRON XWOOD" = "Bronxwood", "BRNXWD" = "Bronxwood", "BRXWD"= "Bronxwood",  "BRXNWD" = "Bronxwood", "B RONXWOOD"= "Bronxwood", "LAURE LTN" = "Laurelton", "MATI LDA" = "Matilda", "R IVERSDE" = "Riverside", "TREMNT" = "Tremont", "TREMO NT" = "Tremont", "TREM ONT"= "Tremont", "T REMONT"= "Tremont", "CL EARVW" = "Clearview", "CLERVW" = "Clearview", "ALEXNDR" = "Alexander", "ALEX ANDER" = "Alexander", "BRCKNR" = "Bruckner", "B RUCKNR"= "Bruckner",  "BRUCKNR"= "Bruckner",  "BR UCKNER"= "Bruckner", "BRUCKNE"= "Bruckner", "D ONGAN" = "Dongan", "HILLSI DE" = "Hillside", "HIL LSIDE" = "Hillside", "ME RRICK" = "Merrick", "WSTCHESTR" = "Westchester", "WSTCH ESTR" = "Westchester", "WESTSCHSTR"  = "Westchester", "WES TCHSTR" = "Westchester", "WESTCHSTR"= "Westchester", "P ARSNS" = "Parsons", "PARSO NS" = "Parsons", "PAR SONS"= "Parsons", "PA RSONS"= "Parsons", "PARS ONS"= "Parsons", "DA HILL" = "Dahill", "UN DERHILL" = "Underhill", "UNDRHILL" = "Underhill", "UNDE RHILL" = "Underhill", "UND ERHILL"= "Underhill", "UN DEHILL"= "Underhill", "H OMELAWN" = "Homelawn", "BE NSON" = "Benson", "MEADO" = "Meadow", "LITHON IA" = "Lithonia", "BRON X" = "Bronx", "EDGEW OOD" = "Edgewood", "EDGEWD"= "Edgewood", "EDGE WD"= "Edgewood", "MORR ISON" = "Morrison", "PINE GROVE" = "Pinegrove", "COUR TLANDT" = "Courtlandt", "SOUTHRN" = "Southern", "BRIDG ETON" = "Bridgeton", "UNIO N" = "Union", "FRSH MDW" = "Fresh Meadow", "BTH MEMORIAL" = "Booth Memorial", "STRO NG" = "Strong",  "STRON G" = "Strong", "O CEAN" = "Ocean", "OC EAN"= "Ocean", "OCEA N"= "Ocean", "OCE AN"= "Ocean", "FIND LAY" = "Findlay", "SHER IDAN" = "Sheridan", "COLUMBS" = "Columbus", "COLUM BUS" = "Columbus", "REVE RE" = "Revere", "REVER E" = "Revere", "FO RDHAM" = "Fordham", "FORDHM" = "Fordham", "FOR DHM" = "Fordham", "F ORDHAM"= "Fordham", "FORD HAM"= "Fordham", "PER OT" = "Perot", "MERMA ID" = "Mermaid", "M ERMAID"= "Mermaid", "HORACE HRD NG" = "Horace Harding", "HO RACE HRDNG" = "Horace Harding", "HRCE HRDNG" = "Horace Harding", "HORACE HRDING" = "Horace Harding", "HRC E HRDNG"= "Horace Harding", "HO RACE HRDING" = "Horace Harding", "HORA CE HRDG" = "Horace Harding", "HORCE HRDG" = "Horace Harding", "STRATFO RD" = "Stratford", "STRAT FORD" = "Stratford", "NEPTNE" = "Neptune", "NEPTUN E" = "Neptune", "BRGHT N" = "Brighton", "BRGHTN" = "Brighton", "BRIGHT N" = "Brighton", "WILLETSPNT" = "Willets Point", "WILLETS PNT" = "Willets Point", "CORTE LYOU" = "Cortelyou", "KINGS LAND" = "Kingsland", "KING SLAND" = "Kingsland", "BARN ES" = "Barnes", "B ARNES"= "Barnes", "BRUN ER" = "Bruner",  "BUSS ING" = "Bussing", "BU SSING"= "Bussing", "LACO NIA" = "Laconia", "LAC ONIA"= "Laconia", "MARB LE" = "Marble", "EXT ERIOR" = "Exterior", "PAUL DING" = "Paulding", "PAULD NG"= "Paulding", "PAULD ING"= "Paulding", "SCHI EFFELIN" = "Schieffelin", "SCHI EFFELN"= "Schieffelin", "KATO NAH" = "Katonah", "WHITE PLNS" = "White Plains", "WHITE PLAI" = "White Plains", "WHIT E PLAINS" = "White Plains",
"EDEN WALD" = "Edenwald", "MURD OCK" = "Murdock", "MONT ICELLO" = "Monticello", "SKILL MN" = "Skillman", "SKILLMN" = "Skillman", "SKILLMA N" = "Skillman", "C ANEY" = "Caney", "BENSO N" = "Benson", "B AYRIDGE" = "Bayridge", "BAY RIDGE"= "Bayridge", "RYD R" = "Ryder", "RYDR" = "Ryder", "RY DER"= "Ryder", "NOSTRA ND" = "Nostrand", "NOST RAND", "NSTRN D"= "Nostrand", "NOSTRN D"= "Nostrand", "NSTRND"= "Nostrand", "NO STRAND"= "Nostrand", "BRKLYN" = "Brooklyn", "BR OOKLYN" = "Brooklyn", "BKLYN"= "Brooklyn", "BROOKLY N"= "Brooklyn", "BRO OKLYN"= "Brooklyn", "WO ODSIDE" = "Woodside", "LENWD" = "Glenwood", "GLEN WOOD" = "Glenwood", "GLENW OOD" = "Glenwood",  "GLENWOO D" = "Glenwood", "GLENWD"= "Glenwood",  "GLENWO OD"= "Glenwood", "CLARE NDON" = "Clarendon", "CLAREND ON"= "Clarendon", "CLARNDN"= "Clarendon", "CL ARENDON"= "Clarendon", "CLARN DN"= "Clarendon", "COL ONIAL" = "Colonial", "C OLONIAL"= "Colonial", "JU NIPER" = "Juniper", "ALDERTN" = "Alderton", "FILLM ORE" = "Fillmore", "FILL MORE"= "Fillmore", "CATAMAR AN" = "Catamaran", "HAR WAY" = "Harway", "K INGSBURY" = "Kingsbury", "HAMI LTON" = "Hamilton", "HAMILTN" = "Hamilton", "FLAT BUSH" = "Flatbush", "KINGS TON" = "Kingston", "KI NGSTON"= "Kingston", "KIN GSTON" = "Kingston", "KIN GSTON"= "Kingston", "HENDRIC KSON" = "Hendrickson", "UTR ECHT" = "Utrecht", "CLAREMNT" = "Claremont", "CLAREM ONT" = "Claremont", "CLERMNT" = "Clermont", "CLRMN T"= "Clermont", "WEBSTR" = "Webster", "WEB STR" = "Webster", "WE BSTR"= "Webster", "WE BSTER"= "Webster", "WEB STER"= "Webster", "FRANKLI N" = "Franklin", "FRAN KLIN"= "Franklin", "FRAN KLYN"= "Franklin", "CONEY ILND" = "Coney Island", "CONE Y ILND"= "Coney Island", "RIVINGT N" = "Rivington","RIVINGTN" = "Rivington", "SEN ATOR" = "Senator", "FLUSHIN G" = "Flushing", "THOMSN" = "Thomson", "THOMS N" = "Thomson", "RAL PH" = "Ralph", "RA LPH" = "Ralph", "RALP H"= "Ralph", "LIVERP OOL" = "Liverpool", "NEV INS" = "Nevins", "NE VINS"= "Nevins", "N EVINS"= "Nevins", "P RINCE" = "Prince", "RID GE" = "Ridge", "RIDG E"= "Ridge", "UTI CA" = "Utica", "UTIC A"= "Utica", "CHURC H" = "Church", "CHUR CH" = "Church", "CHU RCH"= "Church", "MAYFAI R" = "Mayfair", "VETERA NS" = "Veterans", "STRIC KLAND" = "Strickland", "HIGHLA WN" = "Highlawn", "RID GE" = "Ridge", "BEVERL EY" = "Beverley", "BEVE RLY" = "Beverley", "BEVER LEY"= "Beverley", "BEV ERLEY"= "Beverley", "BEVE RLEY"= "Beverley", "CR OSSBAY" = "Crossbay", "LEGIO N" = "Legion", "WILMO HR" = "Willmohr", "CA FFREY" = "Caffrey", "BO LTON" = "Bolton", "BOLTO N" = "Bolton", "BOLT ON"= "Bolton", "V ANDALIA" = "Vandalia", "GAT ES" = "Gates", "GA TES"= "Gates",  "GATE S"= "Gates", "VANDERBLT" = "Vanderbilt", "VANDERBIL T" = "Vanderbilt", "VA NDERBILT"= "Vanderbilt", "DEK ALB" = "Dekalb", "DE KALB" = "Dekalb", "DEKAL B"= "Dekalb", "DEKA LB"= "Dekalb", "SCHENE CTADY" = "Schenectady", "SC HENECTADY"= "Schenectady", "SCHENCTDY"= "Schenectady", "SCHE NECTADY"= "Schenectady", "SCHENECTA DY"= "Schenectady", "SCHCT DY"= "Schenectady", "SCHNCTDY"= "Schenectady", "SCHN CTDY"= "Schenectady", "LA CMBE" = "Lacombe", "LA COMBE" = "Lacombe", "LACO MBE" = "Lacombe", "LACMBE" = "Lacombe", "SE WARD" = "Seward", "LA FAYTTE" = "Lafayette", "LAF AYTTE" = "Lafayette", "LAFA YETTE" = "Lafayette", "LAF AYETTE" = "Lafayette", "LAFAYTTE"= "Lafayette", "LAFAY ETTE"= "Lafayette", "L AFAYETTE"= "Lafayette", "CASTLETN" = "Castleton", "CASTL ETON"= "Castleton", "OAKL AND" = "Oakland","CLARKS ON" = "Clarkson", "VANBRUN T" = "Vanbrunt", "VAN BRUNT" = "Vanbrunt", "ANNADAL E" = "Annadale", "ADELA IDE" = "Adelaide", "SOUN DVW" = "Soundview", "SOUNDVW" = "Soundview", "SOUND VIEW"= "Soundview", "CASTLHLL" = "Castlehill", "CAST LHLL" = "Castlehill", "CASTLH" = "Castlehill", "MA CDOUGAL" = "MacDougal", "GIFFOR DS" = "Giffords", "SNEDIK ER" = "Snediker", "LOUSIA NA" = "Louisiana", "LOUSIANA" = "Louisiana", "LOU ISIANA"= "Louisiana", "TRATMA N" = "Tratman", "GR EENE" = "Greene", "RA ILROAD" = "Railroad", "R AILROAD" = "Railroad", "RAILRD" = "Railroad", "HIG HLAND" = "Highland", "HIGHLND"= "Highland", "HOGHL ND"= "Highland", "CO RTLANDT" = "Cortlandt", "W ALLACE" = "Wallace", "AMBO Y" = "Amboy", "A MBOY"= "Amboy", "AMB OY"= "Amboy", "MANHAT TAN" = "Manhattan", "TAYL OR" = "Taylor", "TAY LOR"  = "Taylor",  "T AYLOR"= "Taylor", "TAYLO R"= "Taylor", "SACK T" = "Sackt", "MORR IS" = "Morris", "M ORRIS"= "Morris", "SARATGA" = "Saratoga", "SANF ORD" = "Sanford", "SANFO RD"= "Sanford", "BELA IR" = "Belair", "MARY LAND" = "Maryland", "B RINSMADE" = "Brinsmade", "OV ERBROOK" = "Overbrook", "MARATHO" = "Marathon", "BRY ANT" = "Bryant", "B RYANT"= "Bryant", "NEWLTS" = "New Lots", "NEWLOT S"= "New Lots", "PEN NSYLVANIA" = "Pennsylvania", "PNSLVANIA" = "Pennsylvania", "PNNSLVNIA" = "Pennsylvania", "PENNSYL VANIA"= "Pennsylvania", "PNNSYLVNIA"= "Pennsylvania", "PENNSYLVANNI"= "Pennsylvania", "EUCLI D" = "Euclid", "EUC LID", "FUL TN" = "Fulton", "FUL TN"= "Fulton", "FULT ON"= "Fulton", "FULTN"= "Fulton", "CL ARK" = "Clark", "CL ASSON" = "Classon", "CLAS SON" = "Classon", "CLASSN"= "Classon", "CLASS ON"= "Classon", "CO RNAGA" = "Cornaga", "FAN CHON" = "Fanchon", "ROCHSTR" = "Rochester", "ROC HESTER"= "Rochester", "ATLAN TIC" = "Atlantic", "ATLANTC" = "Atlantic", "ATLA NTC" = "Atlantic", "ATL ANTIC" = "Atlantic", "ATLNTC" = "Atlantic", "AT LNTC" = "Atlantic", "ATLANT IC"= "Atlantic", "A TLANTIC"= "Atlantic","CALHOU N" = "Calhoun", "CAL HOOUN"= "Calhoun", "CAL HOUN"= "Calhoun", "D ITMAS" = "Ditmas", "DITM AS"= "Ditmas", "HILLMA N" = "Hillman", "EDISO N" = "Edison", "E DISON"= "Edison", "VINCE NT" = "Vincent", "ACADEM Y" = "Academy", "PATCH EN" = "Patchen", "PAR K S" = "Park South", "ST UYVESNT" = "Stuyvesant", "S TUYVESANT"= "Stuyvesant", "STUY VESANT"= "Stuyvesant",  "STYVESN T"= "Stuyvesant", "ST UYVESANT" = "Stuyvesant", "STUYVESA NT"= "Stuyvesant", "M ARCY" = "Marcy", "MA RCY"= "Marcy", "MARC Y" = "Marcy", "ALL ERTON" = "Allerton", "VAN SI CLEN" = "Van Siclen", "VAN SCKLN"= "Van Siclen", "PELHM" = "Pelham", "WARIN G" = "Waring", "WA RING"= "Waring", "PI TKIN" = "Pitkin", "KEARN EY" = "Kearney", "KE ARNEY"= "Kearney", "W ILLIAMS" = "Williams", "WILL IAMS"= "Williams", "PRENT ISS" = "Prentiss", "T WNPNS" = "Twin Pines", "TWNPNS" = "Twin Pines", "TH ROOP" = "Throop", "THROO P"= "Throop", "THRO OP"= "Throop", "GEORG IA" = "Georgia", "MILLE R" = "Miller", "TIF FANY" = "Tiffany", "STEINWA Y" = "Steinway", "HUNTS PNT" = "Hunts Point", "HUN TSPNT" = "Hunts Point", "W ORTMAN" = "Wortman", "WO RTMAN"= "Wortman", "STI CKBALL" = "Stickball", "BATHGTE" = "Bathgate", "BA THGATE" = "Bathgate", "BAR RETT" ="Barrett", "HYLA N" = "Hylan", "RESRVOIR"="Reservoir", "BAINBRI DGE" = "Bainbridge", "BAINBRDG"= "Bainbridge", "BN BRIDGE"= "Bainbridge", "LEFFTRS" = "Lefferts", "LEFFER TS" = "Lefferts", "LE FFERTS"= "Lefferts", "WATERBRY" = "Waterbury", "W ATERBURY"= "Waterbury", "WEST CHSTR" = "Westchester", "WESTCHES"= "Westchester", "COMMRCE" = "Commerce", "KNICKRBCKR" = "Knickerbocker", "KNICKR BCKR" = "Knickerbocker", "CENTRL" = "Central", "CENTR AL"= "Central", "CENTR AL"= "Central", "BUTTRIC K" = "Buttrick", "HANCO CK" ="Hancock", "BAYCHSTR" = "Baychester", "BAYC HESTER" = "Baychester", "NEED HM" = "Needham", "NEEDHM" = "Needham", "ALBEMAR IE" =  "Albemarle", "AL BEMARLE" =  "Albemarle", "COOPE R" = "Cooper", "BIO NIA" = "Bionia", "BION IA"= "Bionia", "BL ACKROCK" = "Blackrock", "FARRAG T" = "Farragut", 
"FA RRAGUT"= "Farragut", "FARRA GUT"= "Farragut", "FARR AGUT"= "Farragut", "CRAN FORD" = "Cranford", "HAVNWD" = "Havenwood",  "HAVN WD" = "Havenwood", "BO RMAN" = "Borman", "SMIT H"="Smith", "CUMB RLND" = "Cumberland", "CUMBRLND"= "Cumberland", "CORNELI A" = "Cornelia", "HYA TT" = "Hyatt", "BAL TIC" = "Baltic", "MARKHM" = "Markham", "CRS BX" = "Cross Bronx", "RODM AN" = "Rodman", "J IMMY" = "Jimmy", "HAI GHT" = "Haight", "MCBA INE" = "McBaine", "CE BRA" = "Cebra",  "TEN BROECK" = "Tenbroeck", "WICK HAM" = "Wickham", "BE AR" = "Bear", "BEA R", "NETHER LAND" = "Netherland", "NETHE RLAND"= "Netherland", "BOU NDARY" = "Boundary", "BERG LUND" = "Berglund", "A LAMEDA" = "Alameda", "RAMB LEWOOD" = "Ramblewood", "RAI LRD" = "Railroad", "DURNT" = "Durant", "ARLE NE" = "Arlene", "JER OME" = "Jerome", "JER ME"= "Jerome", "JE ROME"= "Jerome", "JERO ME"= "Jerome", "TH ERESA" = "Theresa", "BER RIMAN" = "Berriman", "HARVES T" = "Harvest", "BAR RETTO" = "Barretto", "BA RRETTO"= "Barretto", "BARET TO"= "Barretto", "FAIR LAWN" = "Fair Lawn", "AG UILAR" = "Aguilar", "N OBLE" = "Noble", "FOST ER" = "Foster", "BAYSWA TER" = "Bayswater", "CRANBE RRY" = "cranberry", "STER LING" = "Sterling", "STE RLING"= "Sterling", "ST ERLING"= "Sterling", "PL AZA" = "Plaza", "HANN IBAL" = "Hannibal", "ISLND" = "Island", "I SLND" = "Island", "SOMR S" = "Somers", "SOME RS"= "Somers", "CHAU NCY" = "Chauncey", "CHAUN CEY"= "Chauncey", "GRAVSND" = "Gravesend", "GRAVE SEND"= "Gravesend", "GRAV ESEND"= "Gravesend", "PARKSID E" = "Parkside", "SCHRDRS"  = "Schroeders", "LOCST" = "Locust", "CHRIS TOPHER"= "Christopher", "CH ARLES" = "Charles", "CL ERMONT" = "Clermont", "CAT TARAUGUS"  = "Cattaraugus", "CRO TONA" = "Crotona", "C ROTONA"= "Crotona", "SEA MAN"  = "Seaman", "SNYDR"  = "Snyder", "RICHMND" = "Richmond", "RICH MOND"= "Richmond", "RI CHMND"= "Richmond", "LARED O"  = "Laredo", "LEVRETT"  = "Leverett", "SH EFFIELD"  = "Sheffield", "SH EPHERD"  = "Shepherd", "NE ILL" = "Neill", "HAGA MAN"  = "Hagaman", "RUTL AND" = "Rutland", "RUTL ND"= "Rutland", "HI NSDALE"  = "Hinsdale", "SCHURZ"  = "Schurz", "MAYF LOWER"  = "Mayflower", "M AYFLOWER"= "Mayflower", "C ROES" = "Croes", "S HARPE"  = "Sharpe", "NE W"  = "New", "FAI LE" = "Faile", "SEN ECA"  = "Seneca", "SE NECA"= "Seneca", "AMERIC AS"  = "Americas", "EASTER N"  = "Eastern", "EASTRN" = "Eastern", "SAG ONA"  = "Sagona", "WALLAC E" = "Wallace", "RODNE Y" = "Rodney", "MAC KAY" = "Mackay", "CARL TON" = "Carlton","HAN NAH" = "Hannah", "FA RADAY"= "Faraday", "ELMWO OD"  = "Elmwood", "ELMWD"= "Elmwood", "ELMW OOD"= "Elmwood", "PA RADE" = "Parade", "SHEL DN"  = "Sheldon", "SHELDN"= "Sheldon", "PA TTERSON"  = "Patterson", "MAS ON"  = "Mason", "FERN DALE"  = "Ferndale", "FE RNDALE"= "Ferndale", "MYR TLE"  = "Myrtle", "MRTL E" = "Myrtle", "W ATSON"  = "Watson", "WATS"= "Watson", "GRAN D"  = "Grand", "GRA ND"= "Grand", "PA RK" = "Park", "P `ARK" = "Park", "ARNL D"  = "Arnold", "FORS T"  = "Forest", "FOR EST"= "Forest", "F OREST"= "Forest", "SHAK ESPEARE"  = "Shakespeare", "ZER EGA"  = "Zerega", "VAN NAM E" = "Van Name", "MAL TA" = "Malta", "WAVERL Y"  = "Waverly", "WA VERLY"= "Waverly", "WAVER LY"= "Waverly", "ST ANLEY" = "Stanley", "STANLY" ="Stanley", "STANL Y" = "Stanley", "SEMINOL E"  = "Seminole", "S EIGEL"  = "Seigel", "M IRIAM" = "Miriam", "GARRE TSON"  = "Garretson", "RENWIC K" = "Renwick", "BRE WER"  = "Brewer", "JEFF ERSN" = "jefferson", "HEFF RSN"= "jefferson", "JE FFERSON"= "jefferson", "JEFFRSN"= "jefferson", "JEFFER SON"= "jefferson", "J EFFERSON"= "jefferson", "BURB ANK"  = "Burbank", "BL ISS" = "Bliss", "AL BERTA"  = "Alberta", "HAR BOUR"  = "Harbour", "FAIR VI EW"  = "Fairview", "FA IRVIEW"= "Fairview", "MOT TE"  = "Motte", "BAIS LEY" = "Baisley", "VAN BUREN"  = "Van Buren", "SH ORE"  = "Shore", "SHOR E"= "Shore", "FE ATHERBED"  = "Featherbed", "BR ANDT"  = "Brandt", "FREEBO RN" = "Freeborn", "ARBUTU S"  = "Arbutus", "TOMPKIN S"  = "Tompkins", "TOMP KINS"= "Tompkins", "TOMPK INS"= "Tompkins", "CLINTN"  = "Clinton", "CLIN TN"  = "Clinton", "CLIN TON"= "Clinton", "CL INTON"= "Clinton", "BAYCHESTR"  = "Baychester", "DON IZTTI"  = "Donizetti", "ALD RICH"  = "Aldrich", "PO NTIAC"  = "Pontiac", "CA NTERBURY"  = "Canterbury", "SCHENC K"  = "Schenck", "S CHENCK"= "Schenck", "SCH ENCK"= "Schenck", "OLM STD"  = "Olmstead", "CHATTERTN"  = "Chatterton", "KIS WICK"  = "Kiswick", "WA TERBURY"  = "Waterbury", "BLA CKROCK"  = "Blackrock", "VERNN"  = "Vernon", "VERNO N"= "Vernon", "W ILLMOHR"="Willmohr", "WILLIAMSBRDG"  = "Williamsbridge", "WILLMSBRDG"= "Williamsbridge", "WILLIAMSBRDG"= "Williamsbridge", "RHNLNDR"  = "Rhinelander", "FENI MORE"  = "Fenimore", "FENI MRE"= "Fenimore", "FEN IMORE"= "Fenimore", "FE NIMORE"= "Fenimore", "FENIMRE"= "Fenimore", "R IDGE"  = "Ridge", "PL NS"  = "Plains", "PLNS"= "Plains", "CLEV ELAND"  = "Cleveland", "CEDARV IEW"  = "Cedarview", "EASTCH ESTER"  = "Eastchester", "EAST CHESTER"= "Eastchester", "EASTCHSTR"= "Eastchester", "BLAK E"  = "Blake", "BLA KE", "MONTR EAL"  = "Montreal", "MONT REAL"= "Montreal", "H ART" = "Hart", "GRA CE"  = "Grace", "FO RBELL"  = "Forbell", "ME NAHAN"  = "Menahan", "M ENAHAN"= "Menahan", "MA CE" = "Mace", "VAR ET"  = "Varet", "JOH NSON"  = "Johnson", "JOHNSO N"= "Johnson", "JO HNSON"= "Johnson", "WILLOU GHBY" = "Willoughby", "WILL OUGHBY"= "Willoughby", "MAU JR" = "Maujer", "MAUJE R"= "Maujer", "BLOND L"  = "Blondell", "RUDYAR D" = "Rudyard", "WOODRW" =  "Woodrow", "WO ODROW"=  "Woodrow", "ELLSWR TH" =  "Ellsworth", "ELLSWRTH"=  "Ellsworth", "ELL SWORTH"=  "Ellsworth", "BALLRD" =  "Ballard", "LY DIG" =  "Lydig", "SHE EPSHEAD" =  "Sheepshead", "VAN NST"  = "Van Nest", "DELA FIELD"  = "Delafield", "BEDEL L"  = "Bedell", "BE DELL"= "Bedell", "BULL ARD"  = "Bullard", "WES TWOOD"  = "Westwood", "A DEE" =  "Adee", "THMAS S BOYLND" =  "Thomas Boyland", "HA LL"= "Hall", "DAL Y" =  "Daly", "BOSTN" =  "Boston", "BOST N"=  "Boston", "BO STON"=  "Boston", "HAMD EN" =  "Hamden", "O LD" =  "Old", "TIEM ANN" =  "Tiemann", "TIEMAN N"=  "Tiemann", "T IEMANN"=  "Tiemann", "AL BANY" =  "Albany", "TU RNBULL"  = "Turnbull", "DU RYEA"  = "Duryea", "DE AN" =   "Dean", "EDIT H"  = "Edith", "GO ETHALS"  = "Goethals", "W ESTBROOK" =  "Westbrook", "M AGENTA" =  "Magenta", "PLHM PKWY" = "Pelham Parkway", "LENO X"=  "Lenox", "WINEGA R"  = "Winegar", "WHIT MAN"  = "Whitman", "AINS WORTH"  = "Ainsworth", "PARK INSON"  = "Parkinson", "MATTH EWS"  = "Matthews", "BU FFALO" =  "Buffalo", "BERGE N" = "Bergen", "BERG EN"= "Bergen", "ALPIN E" =  "Alpine", "T HEBES" =  "Thebes", "ENTR" =  "Enter", "BOLL ER" =  "Boller", "JANS EN"  = "Jansen", "BR ADY" = "brady", "WAL THAM" =   "Waltham", "WINT HROP" =  "Winthrop", "WIN THROP" =  "Winthrop", "HICK S" =  "Hicks", "MCDONLD" =  "McDonald", "MCD ONLD"=  "McDonald", "LIN WOOD" =  "Linwood", "LINW OOD" =  "Linwood", "L ONGFELLOW"=  "Longfellow", "LO NGFELLOW"=  "Longfellow",  "GL EASON" =  "Gleason", "MCKI NLEY" = "McKinley", "MCKIN LEY" = "McKinley", "P INTO" =  "Pinto", "PR ESCOTT" =  "Prescott", "SPENC R" =   "Spencer", "PARK WAY" =   "Parkway", "LINCO LN"=  "Lincoln", "LI NCOLN"=   "Lincoln", "L INCOLN"=   "Lincoln", "LINC OLN"=   "Lincoln", "W ARNING" = "Warning", "GR EENWOOD"=   "Greenwood", "GREE NWOOD"=   "Greenwood", "DEL AWARE" =  "Delaware", "RETFO RD"=  "Retford",
"C URRY"=   "Curry", "THORN YCROFT"=   "Thornycroft", "NE AL"=   "Neal", "NEA L", "HOWA RD"=   "Howard", "HOWRD"=   "Howard", "HENDR X"="Hendrix", "HENDRX"= "Hendrix", "H ENDRIX"= "Hendrix", "HNDR X"="Hendrix", "OA KDALE"=   "Oakdale", "ALT OONA"=   "Altoona", "P OLIT"=   "Polit", "POMPE Y"=   "Pompey", "HA WTHORNE"=   "Hawthorne", "FI GUREA"=   "Figurea", "SA XON"=   "Saxon", "SUY DAM" =  "Suydam", "CASSI DY"=   "Cassidy", "NAGL E"=   "Nagle", "TR AVIS"=   "Travis", "DECAT UR"=   "Decatur", "DE CATUR"=   "Decatur", "DECA TUR"=   "Decatur", "S KIDMORE"=   "Skidmore", "VAN Z"=   "Van Zandt", "NIAGAR A"=   "Niagara", "H UTCHINSON"=   "Hutchinson", "SAN ILAC"=   "Sanilac", "C ASTLe"= "Castle", "MI CKLE"=   "Mickle", "VA N CLEEF"=   "Van Cleef", "CRYS TAL"=   "Crystal", "ES SEX"=   "Essex", "E SSEX"=   "Essex", "ESS EX"=   "Essex", "WILLIAM SON"=   "Williamson", "GRAH AM"=   "Graham", "SI NCLAIR"=   "Sinclair", "RIE GEMANN"=   "Riegemann", "ROGER S"=   "Rogers", "ROG ERS"= "Rogers", "RO GERS"= "Rogers", "COLGA TE"=   "Colgate", "PI NE"="Pine", "X STATE HOSP"= "Creedmoor State Hospital", "E-X STATE HOSP."= "Creedmoor State Hospital", "STATE HOSP."= "Creedmoor State Hospital", "ENT"= "entrance", "ENTRNCE"= "entrance", "N ELL IIOTT"=   "Nellie Elliott", "SCHO LES"=   "Scholes", "REMS EN"=   "Remsen", "HALSE Y"=   "Halsey", "HA LSEY"= "Halsey", "STE LL"=   "Stell", "QUIN CY"=   "Quincy", "QUI NCY"= "Quincy", "QU INCY"= "Quincy", "NE WPOERT"=   "Newport", "N EWPRT"= "Newport", "NEWPRT"= "Newport", "CLAV Er"=   "Claver", "HOLLN D"=   "Holland", "EDGEG ROVE"=   "Edgegrove", "SH ERADEN"=   "Sheraden", "SO UTH"=   "South", "TI LDEN"= "Tilden", "TIL DEN"= "Tilden", "TILD EN"= "Tilden", "TILDE N"= "Tilden", "VANDER VOORT"=   "Vandervoort", "WOO DWARD"=   "Woodward", "DE AN"=   "Dean", "HA RT"=   "Hart", "DEW EY"=   "Dewey", "GRAN DVIEW"=   "Grandview", "T WINE"=   "Twine", "HIN SDALE"=   "Hinsdale", "HIMHR OD"="Himrod", "SHIRL EY"=   "Shirley", "I RWIN"=   "Irwin", "HASPE L"=   "Haspel", "E SPLANADE"=   "Esplanade", "SUTT ER"=   "Sutter", "SU TTER" = "Sutter", "S UTTER" = "Sutter", "SUTTE R" = "Sutter", "WILL IAMS"=   "Williams", "WILLIAM S"=   "Williams", "DU MONT"= "Dumont", "DUMO NT"= "Dumont", "DUMON T"="Dumont", "DEVO N"= "Devon", "D EVON"= "Devon", "WHT NY"= "Whitney", "W HITNEY"="Whitney", "TYN DALL"=   "Tyndall", "WOODH ULL"= "Woodhull", "WOODHU LL"= "Woodhull", "W OODHULL"= "Woodhull", "RIVERD ALE"=   "Riverdale", "KING S"=   "Kings", "KIN GS"=   "Kings", "KI NGS"=   "Kings",  "MORR IS" =  "Morris", "MARI ON"=  "Marion",  "MAIN E"=   "Maine", "WATE RS"=   "Waters", "FOUN TAIN"=   "Fountain", "LU RTING"=   "Lurting", "L URTING"=   "Lurting", "SEYM OUR"=   "Seymour", "A NTHONY"=   "Anthony", "ZULE TTE"=   "Zulette",  "UNIV RSTY"= "University", "UNIVRSTY"= "University", "U NIVERSITY"= "University", "AUTU MN"=   "Autumn", "GERR ITSN"=   "Gerritson", "GERRITSN"=   "Gerritson",  "MAR INE"=   "Marine", "PR ALL"=   "Prall", "WILSO N"=   "Wilson", "WIL SON"=   "Wilson", "YO UNG"=   "Young", "AS TOR"=   "Astor", "HARKNE SS"=   "Harkness", "PUTN AM"=  "Putnam", "PULAS KI" =  "Pulaski", "NORWA Y"=   "Norway",  "T HORNHILL"=   "Thornhill",  "LORTE L"=   "Lortel", "SW AN"=   "Swan", "SUTT ON"=   "Sutton", "MOSHL U"=   "Mosholu", "BR IGGS"=   "Briggs",  "MORRIS N"=   "Morrison",  "WI LLIS" =  "Willis",  "J ARVIS"=   "Jarvis",  "FI NGERBOARD"=   "Fingerboard", "FING ERBOARD"= "Fingerboard",  "FINGE RBOARD" = "Fingerboard","FINGE RBD"= "Fingerboard", "FINGERBD"= "Fingerboard", "FINGERB D"= "Fingerboard", "ORIE NTAL"=   "Oriental", "WES TBURY"=   "Westbury", "BEDE L"=   "Bedell", "BEDE LL"=   "Bedell", "BRE NTWOOD"=   "Brentwood", "B RENTWOOD"=   "Brentwood",  "BU LWER"=   "Bulwer","ELLWO OD"="Ellwood",  "WIL LOWBROOK"=   "Willowbrook",  "ME LBOURNE"=   "Melbourne",  "BA LFOUR"=   "Balfour",  "WELLB ROOK" =  "Wellbrook",  "QUEE NS"=   "Queens", "QU EENS"=   "Queens",  "WEBBR OOK"=   "Webbrook", "MEL BA"=   "Melba", "POYE R"=   "Poyer", "BARB EY"=   "Barbey", "BU RKE"=   "Burke",  "GA BRIELLE"=  "Gabrielle",  "DRIG GS"=  "Driggs",  "VAY LER"=  "Vayler",  "PILL ING"=  "Pilling", "BOYLND"=  "Boyland",  "BLE ECKER"=  "Bleecker", "SULLI VAN"=  "Sullivan", "SUL LIVAN"=  "Sullivan", "RYMND"=  "Raymond", "RAYM OND"=  "Raymond",  "WINDE RMERE"=  "Windermere", "MANO R"=  "Manor", "MA NOR"=  "Manor", "CAULD WELL"=  "Cauldwell", "RADCL IFF"=  "Radcliff",  "RADC LIFF"="Radcliff",  "CUNN INGHAM"=  "Cunningham", "WINC HSTR"=  "Winchester", "COLE RDGE"=  "Coleridge", "COLERDGE"=  "Coleridge",  "BATCHELDE R"=  "Batchelder",  "WOO DMINE"=  "Woodmine", "FALM OUTH"=  "Falmouth", "PRIN CETON"=  "Princeton", "PRI NCETON"=  "Princeton", "LONGW OOD" = "Longwood",  "CHE SEA" = "Chelsea", "SZIC KY"=  "Szickly",  "CULO TTA"=  "Culotta",  "BACH E"=  "Bache", "ST AGG"=  "Stagg", "LOR RAINE"=  "Lorraine", "H OBART"=  "Hobart", "O’B RIEN"=  "O’Brien", "BERESF ORD"= "Beresford",  "NE GUNDO"=  "Negundo", "QUEENS DALE"= "Queensdale",  "KING DOM"=  "Kingdom",  "N BURGER"=  "North Burger",  "B LONDELL"=  "Blondell", "B URNSIDE"=  "Burnside", "BURNSI"=  "Burnside",  "DUN LOP"=  "Dunlop",  "PETE R"=  "Peter",   "HENNSSY"= "Hennessey",  "AMBE R"=  "Amber", "AMB ER"=  "Amber",  "SLEI GHT"=  "Sleight",  "HER ING"=  "Hering",  "LANG HAM" =  "Langham",  "H ASTINGS"=  "Hastings",   "POIL LON"=  "Poillon",  "UNIO N"=  "Union",  "LUTE N"=  "Luten", "VERM ONT"=  "Vermont", "VER MONT"=  "Vermont",  "ALLER TON"=  "Allerton", "BU RNETT"=  "Burnett", "BURN ETT" =  "Burnett", "LAM BRT"=  "Lambert",  "H GMAN"=  "Hegeman", "HGMAN"=  "Hegeman", "GUYO N"=  "Guyon", "BAYA RD"=  "Bayard",  "DUZE R"=  "Duzer", "SUF FOLK"=  "Suffolk", "BAY VIEW"=  "Bayview",   "BEEKMA N"=  "Beekman",  "MONR OE"=  "Monroe", "M ONROE"=  "Monroe", "BREVRT"=  "Brevert", "FO STER"=  "Foster",  "M ARKET"=  "Market", "MARKE T"=  "Market",  "NE WKIRK"=  "Newkirk", "BOYNT ON"=  "Boynton", "STILLW ELL"=  "Stillwell", "STIL LWELL"=  "Stillwell", "BAN KER"=  "Banker",  "BU SH"=  "Bush", "VI SITATION"=  "Visitation", "CORNE LIA"=  "Cornelia", "CENTR E"=  "Centre",  "DICKNSN"=  "Dickenson", "DICK SN"=  "Dickenson",  "F RIEL"=  "Friel", "JOH NS"=  "Johns", "COMMER CE"=  "Commerce",  "PACI FIC"=  "Pacific",  "C RUGER"="Cruger", "R ECTOR"=  "Rector",  "WARW ICK"=  "Warwick", "WANWCK"=  "Warwick",  "JA CKSON"=  "Jackson", "WILS ON"=  "Wilson",  "RCHMNDHILL"=  "Richmond Hill", "WCHMDH ILL"=  "Richmond Hill", "VE DDER"=  "Vedder",  "T EAKWOOD"=  "Teakwood",  "TR EADWELL"=  "Treadwell",  "TYSE N"=  "Tysen", "T YSEN"=  "Tysen",  "NER EID"=  "Nereid", "NERE ID"=  "Nereid",  "CR ESTON"=  "Creston",  "G RISWOLD"=  "Griswold",  "MIDDLTN"=  "Middleton", "GE NESEE"=  "Genesee", "M EMPHIS"=  "Memphis", "N CO NDUIT"=  "North Conduit",   "RET FORD"=  "Retford",    "SW EETGUM"=  "Sweetgum",   "PHI LIP"=  "Philip",  "MORNING STAR"=  "Morningstar", "MORN INGSTAR"=  "Morningstar", "MA XIE"=  "Maxie", "CAR LIN"=  "Carlin", "C RESTON"=  "Creston",  "CROM WELL"=  "Cromwell", "G LOVER"=  "Glover", "REGEN T"=  "Regent", "SIC KLES"=  "Sickles",  "Y ORK"=  "York",  "ELLER Y"=  "Ellery",  "NOR MAN"=  "Norman", "HUNT INGTON"=  "Huntington",   "CA STOR"=  "Castor",  "RIVINGT N"=  "Rivington", "TH OMAS"=  "Thomas",  "SCO TT"=  "Scott",  "HAR T"=  "Hart",   "CARLYL E"=  "Carlyle",  "CYP RESS"=  "Cypress", "CYPR ESS"=  "Cypress", "JULIA NA"=  "Juliana",  "TRKLEMNS"=  "Trockelmans", "TRKL EMNS"=  "Trockelmans",  "SACKM AN"=  "Sackman",  "LA NDING"=  "Landing",  "KE W GARDENS"=  "Kew Gardens",  "RA DFORD"=  "Radford", "H ENRY"=  "Henry",  "THRUW AY"=  "Thruway",   "ARDS LEY"=  "Ardsley",  "WINA NT"=  "Winant",   "DEME YER"=  "Demeyer", "SWINT ON"=  "Swinton",  "DE RUYTER"=  "Deruyter", "LE ONARD"=  "Leonard", "LEONRD"=  "Leonard", "LEO NARD"=  "Leonard", "P RESIDENT"=  "President", "PRSDENT"=  "President", "PRESI DNT"=  "President",  "PRESIDE NT"=  "President",  "MONTGO"=  "Montgomery", "MNT GOMRY"=  "Montgomery", "MON TGOMERY"=  "Montgomery", "MONTGO MERY"=  "Montgomery",  "TA RGEE"=  "Targee",  "DENK ER"=  "Denker",  "TO WNSEND"=  "Townsend",   "POP LAR"=  "Poplar", "EYLAN DT"=  "Eylandt", "ASHF ORD"=  "Ashford",  "G IVAN"=  "Givan",   "MEDIN A"=  "Medina",  "OAK T REE"=  "Oak tree", 
"OXFO RD"=  "Oxford", "PA ULS"=  "Pauls",  "SUYD AM"=  "Suydam", "DIX ON"=  "Dixon",   "T HERIOT"=  "Theriot",  "EL VERTON"= "Elverton",  "EVERG REEN"="Evergreen",  "LE E"=  "Lee", "SIMO NSON"=  "Simonson",  "MO NSEY"="Monsey", "TR OY"=  "Troy",  "L EWIS"=  "Lewis", "ST ANWIX"=  "Stanwix", "KE NNETH"=  "Kenneth",  "T OMLINSON"=  "Tomlinson",  "VOORHIE S"=  "Voorhies", "VOORHI ES"=  "Voorhies", "JOURN EAY"=  "Journey",  "MONTA GUE"=  "Montague", "LORTE L"=  "Lortel", "WAL L"=  "Wall",  "TUCKRTN"=  "Tuckerton", "W HITE"=  "White", "WHI TE"=  "White",  "LI BERTY"=  "Liberty", "F RANICS"=  "Francis", "H OLLYWOOD"=  "Hollywood",  "F TELEY"=  "Fteley",  "BE NNETT"=  "bennett",  "BE LKNAP"=  "belknap",  "M ORRISON"=  "morrison",  "EMPIR E"=  "empire",  "BO ND"=  "bond", "B ROWVALE"=  "browvale",  "BOLLE R"=  "boller", "NEWYORK"=  "new york",  "UNIO N"=  "union", "VI CTORY"=  "victory", "THROG NECK"=  "throgs neck",  "COL LEGE"=  "college", "FOR ST" = "Forest", "RUGB Y"=  "rugby", "COVE RT"=  "covert",  "HOOK CRK"=  "hook creek", "RI DGEDALE"=  "ridgedale", "EDD Y"=  "eddy", "ED DY"=  "eddy", "CLY MER"=  "clymer", "CR EST"=  "crest",  "CO NDUIT"=  "conduit", "CON DUIT"= "conduit", "ME LVIN"=  "melvin",  "FRA NCIS"=  "francis", "CHISHO LM"=  "chisholm",  "WES T"=  "west",  "H OUSTON"=  "houston", "HOUSTO N"=  "houston",  "AUSTI N"=  "austin", "BLMNG DALE"=  "bloomingdale", "BLMNGDALE"="bloomingdale",  "HELIO S"=  "helios",  "EME RALD"=  "emerald", "DE RE IMER"=  "dereimer",  "LOR IMER"=  "lorimer", "LORIM ER"=  "lorimer", "IRE LAND"=  "ireland", "OVIN GTON"=  "ovington",  "GIRA RD"=  "girard",  "LOG AN"=  "logan", "W.KIN GSBRDG"=  "west kingsbridge", "W.KINGBRDG"= "west kingsbridge",  "K ENSINGTON"=  "kensington", "HUMBOLDT"=  "humboldt", "S TEINWAY"=  "steinway", "TYSE NS"=  "tysens",  "WOLVER INE"=  "wolverine",  "BOER UM"=  "boerum", "M ORGAN"=  "morgan", "LURTI NG"=  "lurting",  "BE RRY"="berry",  "M AIN"=  "main", "WIL LETS"=  "willets",  "BRIG HAM"=  "brigham", "VEN ICE"=  "venice", "SCHOHRI E"=  "schoharie", "LOC KWOOD"=  "lockwood", "ST.P ETRS"=  "st. peters", "ST.PETRS"=  "st. peters", "WOOD S"=  "woods", "EL LICOTT"=  "ellicott",   "COMME RCE"=  "commerce",   "COMM ONWEALTH" = "commonwealth",   "CARROL L"=  "carroll", "STE VENSON"=  "stevenson", "STEVN SN"=  "stevenson", "STEVNSN"=  "stevenson",  "FT INDE"=  "fort independence",  "C HESTNUT"=  "chestnut",  "ELR IDGE"=  "elridge",  "COLDE N"=  "colden", "FDR DR"=  "FDR drive",  "ROS E"=  "rose",  "AR CHER"=  "archer",  "CL OVE"=  "clove", "CLOV E"=  "clove", "D WIGHT"=  "dwight", "SUMMERF IELD"=  "summerfield",  "RE ADE"=  "reade",  "SOU THWEST"=  "southwest",  "HENDR ICKS"=  "hendricks",  "EUNI CE"=  "eunice",  "E RICSON"=  "ericson",  "STCKH LM"=  "stockholm", "STCKHLM"=  "stockholm",  "JE NNINGS"=  "jennings", "OSW EGO"=  "oswego",  "PECK S LIP"=  "peck slip",  "KE LLY"=  "kelly", "COYL E"=  "coyle",  "S COTT"=  "scott", "CARLISL E" = "carlisle",   "WE XFORD"= "wexford", "W EXFORD"= "wexford",  "P ILGRIM"=  "pilgrim", "HE LENA"=  "helena",  "HE NLEY"=  "henley",  "WEE D"="weed",  "QUENT IN"=  "quentin", "Q UENTIN"=  "quentin",  "DR UMGOOLE"=  "drumgoole",  "FIE LDSTON"=  "fieldstone",  "J ULES"=  "jules",  "TRO Y"=  "troy", "KILDA RE"=  "kildare",  "FAI RFIELD"=  "fairfield",  "MONITR"=  "monitor",  "U NIONPORT"=  "unionport",  "E LM"=  "elm",  "F ORD"=  "ford",  "FIR TH"=  "firth",  "G UNTHER"=  "gunther", "GUN HILL"="gunhill", "GUNH ILL"=  "gunhill", "DE ISIUS"=  "deisius",  "TELEP ORT"=  "teleport",  "STEW ART"=  "stewart", "HE WITT"=  "hewitt", "AINSL IE"=  "ainslie", "GU ERLAIN"=  "guerlain", "SHADO W"=  "shadow", "DOUGH TY"=  "doughty", "HE RKIMER"=  "herkimer", "ELLIN GTON"=  "ellington", "TREN T"= "trent", "M ILTON"=  "Milton", "Y ORK"=  "York", "THAME S"=  "Thames", "SEXT ON"=  "Sexton", "TR UXTON"=  "truxton",  "GUYBRWR" = "Guy R Brewer",  "GUYBRW R" = "Guy R Brewer", "NO LL"=  "noll", "KNAP P"=  "knapp", "SPIR IT"=  "spirit", "BR OOKVILLE"=  "brookville", "B ROOK"="Brook", "B ROWN"=  "brown", "BROW N"=  "brown", "BR ENTON"=  "brenton",  "K INGSBRDG"=  "kingsbridge", "ROSE WOOD"=  "rosewood",  "PEMBRKE"=  "pembroke",  "HEM LOCK"=  "hemlock", "SCHER MERHORN"=  "schermerhorn",  "O’D DONNELL"=  "o’ddonnell", "MIL FORD"=  "milford", "BOUL DER"=  "boulder", "WES TMINSTER"=  "westminster", 
"FROS T"=  "frost", "CON CORD"=  "concord", "CI CRLE"=  "circle", "WEL DON"=  "weldon", "COLDS PRING"=  "coldspring", "J OHN"=  "john", "H ICKS"=  "hicks", "ORE GON"=  "oregon", "POWE LL"=  "powell", "HOM E"=  "home",  "IRVIN G"=  "irving", "S HILOH"=  "shiloh", "TH OMPSON"=  "thompson", "ROSD ALE" = "rosedale", "A LDEN" = "alden", "H UNTS" = "hunts", "S LIP"=  "slip", "A VENUE"=  "avenue", "A VE."=  "avenue",  "AV"=  "avenue", "AV."=  "avenue", "AVE."=  "avenue", "A V"= "avenue", "PLAC"=  "place", "W AY"=  "way", "S T"=  "street", "ST ."=  "street", "ST"=  "street",  "S T"=  "street",  "S TREET"=  "street", "EX PRESSWAY"=  "expressway", "EX PWY"="expressway", "EXPY"="expressway", "EXPWY"="expressway", "DRIV EWAY"=  "driveway", "TPK"=  "turnpike", "TPKE."="turnpike", "LN."=  "lane", "TE RR."=  "terrace", "HWY"=  "highway", "BLV D"=  "boulevard", "BL VD"="boulevard", "BV"="boulevard", "BV."="boulevard", "BLVD."="boulevard", "BLVD"="boulevard", "PK W."="parkway",  "PK"="parkway","P K"="parkway","DR"= "drive", "PL" = "place")

# getting rid of all the directional abbreviations (EB, NB, SB, WB), anything after the hyphen, getting rid of white space at the beginning & end                  
cameras_clean1 <- df_deduplicated1 %>%
  mutate(
    cross_street = toupper(as.character(cross_street)),
    cross_street = str_remove_all(cross_street, "^\\b(EB|NB|SB|WB)\\b\\s+"),
    cross_street = str_remove(cross_street, "-.*$") %>% str_trim()
  )

# changing the street_fixes list to a dataframe in order to apply it to the camera dataset
street_fixes1 <- data.frame(
  Wrong = names(street_fixes),
  Right = as.character(street_fixes),
  stringsAsFactors = FALSE
)

# filtering out blank values (it kept interfering when trying to apply it to the main dataset)
street_fixes1 <- street_fixes1 %>%
  filter(!is.na(Wrong) & Wrong != "")

# creating a regex to apply to camera_clean1$cross_street
for(i in 1:nrow(street_fixes1)) {
  pattern_with_boundaries1 <- paste0("\\b", street_fixes1$Wrong[i], "\\b")
  
  cameras_clean1$cross_street <- str_replace_all(
    string = cameras_clean1$cross_street, 
    pattern = pattern_with_boundaries1, 
    replacement = street_fixes1$Right[i]
  )
}

# more cleaning happening, capitalizing everything, changing all of the "WEST" + "EACH" to "W" + "E", getting rid of lingering special characters, + adding "&" in between the two streets, trimming the ends of white space
cameras_unique1 <- cameras_clean1 %>%
  mutate(
    cross_street = toupper(as.character(cross_street)),
    cross_street = str_replace_all(cross_street, "\\bWEST\\b", "W"),
    cross_street = str_replace_all(cross_street, "\\bEAST\\b", "E"),
    cross_street = str_remove_all(cross_street, "[@\\.\\'/,]"),
    cross_street = str_replace(cross_street,  "(\\bAVENUE\\b|\\bSTREET\\b|\\bROAD\\b|\\bBOULEVARD\\b|\\bPLACE\\b|\\bPARKWAY\\b|\\bEXPRESSWAY\\b|\\bHIGHWAY\\b)", "\\1 &"),
    cross_street = str_remove(cross_street, " &\\s*$"),
    cross_street = str_squish(cross_street)
  )

# getting rid of remaining white space, getting rid of ordinal labels, creating another letter key that splits by "&" and is identifiable by either order (1 St & Main St or Main St & 1 St)
cameras_unique1 <- cameras_unique1 %>%
  mutate(
    no_space1 = str_replace_all(cross_street, " ", ""),
    no_space1 = str_replace_all(no_space1, "(\\d)(ST(?!REET)|ND|RD|TH)", "\\1")
  ) %>%
  rowwise() %>%
  mutate(match_key_nospace1 = paste(sort(str_split(no_space1, "&")[[1]]), collapse = "&")) %>%
  ungroup()

# doing a final consolidation in case there are any matches that weren't found before (there are)
cameras_final1 <- cameras_unique1 %>%
  filter(match_key_nospace1 != "" & !is.na(match_key_nospace1)) %>%
  group_by(match_key_nospace1) %>%
  summarise(
    activation_date1 = min(clean_date, na.rm = TRUE),
    display_name1 = first(cross_street),
    camera_count = n(),
    .groups = 'drop'
  )

Crash Data

Now, I am uploading my dataset from “Motor Vehicle Collisions - Crashes”. Because this dataset was so large, I filtered out any rows that did not have a cross street (as we are only looking at intersections) and any rows that did have a side street. Through the main street and the cross street, I will match each crash to a camera and mark them as “before camera installation” (0) or “after camera installation” (1). Lots of cleaning is happening here to standardize it in the same way the camera data is standardized.

crash_data <- fread("~/Downloads/Motor_Vehicle_Collisions_-_Crashes_20260408.csv")

# lots of street cleaning here, capitalizing everything, changing "WEST" and "EAST" to "W" and "E", removing ordinal labels, creating a match key (identifiable in either street order) in order to match it to the camera data
crashes_final1 <- crash_data %>%
  filter(`ON STREET NAME` != "" & `CROSS STREET NAME` != "") %>%
  mutate(
    ON_CLEAN = toupper(`ON STREET NAME`) %>% 
      str_replace_all("\\bWEST\\b", "W") %>% 
      str_replace_all("\\bEAST\\b", "E"),
    CROSS_CLEAN = toupper(`CROSS STREET NAME`) %>% 
      str_replace_all("\\bWEST\\b", "W") %>% 
      str_replace_all("\\bEAST\\b", "E"),
    on_id = str_replace_all(ON_CLEAN, " ", ""),
    cross_id = str_replace_all(CROSS_CLEAN, " ", ""),
    on_id = str_replace_all(on_id, "(\\d)(ST(?!REET)|ND|RD|TH)", "\\1"),
    cross_id = str_replace_all(cross_id, "(\\d)(ST(?!REET)|ND|RD|TH)", "\\1")
  ) %>%
  rowwise() %>%
  mutate(match_key_nospace1 = paste(sort(c(on_id, cross_id)), collapse = "&")) %>%
  ungroup()

# joining crash data & camera data
final_analysis_df1 <- crashes_final1 %>%
  inner_join(cameras_unique1, by = "match_key_nospace1")
kable(head(final_analysis_df1))
CRASH DATE BOROUGH ZIP CODE ON STREET NAME CROSS STREET NAME OFF STREET NAME NUMBER OF PERSONS INJURED NUMBER OF PERSONS KILLED ON_CLEAN CROSS_CLEAN on_id cross_id match_key_nospace1 Street.Name Intersecting.Street min_issue_date cross_street clean letter_id clean_date no_space1
12/14/2021 MEEKER AVENUE LORIMER STREET NA 3 0 MEEKER AVENUE LORIMER STREET MEEKERAVENUE LORIMERSTREET LORIMERSTREET&MEEKERAVENUE NB MEEKER AVE @ LORI MER ST 2022 Sep 30 12:00:00 AM MEEKER AVENUE & LORI MER STREET nbmeekeravelorimerst abeeeeeiklmmnorrrstv 2022-09-30 MEEKERAVENUE&LORIMERSTREET
12/14/2021 MEEKER AVENUE LORIMER STREET NA 3 0 MEEKER AVENUE LORIMER STREET MEEKERAVENUE LORIMERSTREET LORIMERSTREET&MEEKERAVENUE WB MEEKER AVE @ LORI MER ST 2021 May 27 12:00:00 AM MEEKER AVENUE & LORI MER STREET wbmeekeravelorimerst abeeeeeiklmmorrrstvw 2021-05-27 MEEKERAVENUE&LORIMERSTREET
12/16/2021 KINGSLAND AVENUE MEEKER AVENUE NA 1 0 KINGSLAND AVENUE MEEKER AVENUE KINGSLANDAVENUE MEEKERAVENUE KINGSLANDAVENUE&MEEKERAVENUE NB MEEKER AVE @ KING SLAND AVE 2021 Nov 30 12:00:00 AM MEEKER AVENUE & KINGSLAND AVENUE nbmeekeravekingslandave aaabdeeeeegikklmnnnrsvv 2021-11-30 MEEKERAVENUE&KINGSLANDAVENUE
04/24/2022 BROOKLYN 11221 THROOP AVENUE DE KALB AVENUE NA 1 0 THROOP AVENUE DE KALB AVENUE THROOPAVENUE DEKALBAVENUE DEKALBAVENUE&THROOPAVENUE NB THROOP AVE @ DEKA LB AVE 2019 Dec 02 12:00:00 AM THROOP AVENUE & DEKALB AVENUE nbthroopavedekalbave aaabbdeeehklnooprtvv 2019-12-02 THROOPAVENUE&DEKALBAVENUE
04/24/2022 BROOKLYN 11221 THROOP AVENUE DE KALB AVENUE NA 1 0 THROOP AVENUE DE KALB AVENUE THROOPAVENUE DEKALBAVENUE DEKALBAVENUE&THROOPAVENUE WB DEKALB AVE @ THRO OP AVE 2021 Nov 30 12:00:00 AM DEKALB AVENUE & THROOP AVENUE wbdekalbavethroopave aaabbdeeehklooprtvvw 2021-11-30 DEKALBAVENUE&THROOPAVENUE
04/24/2022 QUEENS 11101 THOMSON AVENUE SKILLMAN AVENUE NA 1 0 THOMSON AVENUE SKILLMAN AVENUE THOMSONAVENUE SKILLMANAVENUE SKILLMANAVENUE&THOMSONAVENUE EB THOMSN AVE. SKILL MN AV- 32 PL 2015 Sep 25 12:00:00 AM THOMSON AVENUE & SKILLMAN AVENUE ebthomsnaveskillmnav32pl 23aabeehiklllmmnnopsstvv 2015-09-25 THOMSONAVENUE&SKILLMANAVENUE

163,491 crashes have successfully been matched to a speed camera location.

cameras_with_crashes1 <- final_analysis_df1 %>% 
  distinct(match_key_nospace1) %>% 
  nrow()

There are 2,533 distinct camera locations.

Analysis Prep

This is where I convert crash dates & camera activation dates into one standardized date format and list them as “before” (0) or “after” (1) camera activation. I am also making standardized columns for total injuries, total deaths, and total casualties (injuries + deaths) for each crash.

Finally, I have created a year_window dataset that filters for crashes that have occurred within one year before camera activation and one year after camera activation so that we have balanced before and after groups when analyzing.

analysis_df <- final_analysis_df1 %>%
  mutate(
    crash_date_clean = mdy(`CRASH DATE`),
    activation_date_clean = ymd(clean_date)
  ) %>%
  filter(!is.na(crash_date_clean) & !is.na(activation_date_clean)) %>%
  mutate(
    treatment = if_else(crash_date_clean >= activation_date_clean, 1, 0)
  )

analysis_df <- analysis_df %>%
  mutate(
    total_injuries = `NUMBER OF PERSONS INJURED`,
    total_deaths = `NUMBER OF PERSONS KILLED`,
    total_casualties = `NUMBER OF PERSONS INJURED` + `NUMBER OF PERSONS KILLED`
  )

year_window <- analysis_df %>%
  mutate(days_diff = as.numeric(crash_date_clean - activation_date_clean)) %>%
  filter(days_diff >= -365 & days_diff <= 365) %>%
  mutate(period = if_else(days_diff < 0, "Before", "After"))

Analysis

Descriptive Stats & Initial Crash Analysis

table <- t(table(analysis_df$treatment))
colnames(table) <- c("before", "after")

kable(table)
before after
118367 45954 
summary_stats <- year_window %>%
  group_by(period) %>%
  summarise(
    `Total Crashes` = n(),
    `Total Injuries` = sum(total_injuries, na.rm = TRUE),
    `Total Deaths` = sum(total_deaths, na.rm = TRUE),
    `Total Casualties` = sum(total_casualties, na.rm = TRUE)
  ) %>%
  pivot_longer(cols = -period, names_to = "Metric", values_to = "Value") %>%
  pivot_wider(names_from = period, values_from = Value) %>%
  select(Metric, Before, After) %>%
  mutate(
    `% Change` = paste0(round(((After - Before) / Before) * 100, 1), "%")
  )

kable(summary_stats, caption = "12-Month Before/After Comparison")
12-Month Before/After Comparison
Metric Before After % Change
Total Crashes 12697 10431 -17.8%
Total Injuries 4699 3912 -16.7%
Total Deaths 30 17 -43.3%
Total Casualties 4729 3929 -16.9% 
crash_counts <- year_window %>%
  group_by(period) %>%
  summarise(total_crashes = n())
print(crash_counts)
## # A tibble: 2 × 2
##   period total_crashes
##   <chr>          <int>
## 1 After          10431
## 2 Before         12697

This is a table of the total crashes that occurred before cameras were installed and after in a one-year window.

Analysis: Crashes Overall

Percentage drop in crash volume

crash_before <- crash_counts$total_crashes[crash_counts$period == "Before"]
crash_after <- crash_counts$total_crashes[crash_counts$period == "After"]
crash_pct_drop <- ((crash_after - crash_before) / crash_before) * 100
print(crash_pct_drop)
## [1] -17.84674

There was an observed 17.6% drop in crashes after cameras were installed.

Chi-square test

crash_count_test <- table(year_window$period)
crash_chisq <- chisq.test(crash_count_test)

print(crash_chisq)
## 
##  Chi-squared test for given probabilities
## 
## data:  crash_count_test
## X-squared = 222.01, df = 1, p-value < 2.2e-16

The difference between the number of crashes before and after camera activation is significant!

Total crashes before & after camera activation (one-year window)

crash_plot <- year_window %>%
  group_by(period) %>%
  summarise(total_crashes = n()) %>%
  mutate(period = factor(period, levels = c("Before", "After")))

ggplot(crash_plot, aes(x = period, y = total_crashes, fill = period)) +
  geom_bar(stat = "identity", width = 0.6) +
  geom_text(aes(label = scales::comma(total_crashes)), vjust = -0.5, size = 5) +
  scale_y_continuous(expand = expansion(mult = c(0, 0.1))) +
  scale_fill_manual(values = c("Before" = "#d95f02", "After" = "#1b9e77")) +
  labs(title = "Total Crash Volume: 12-Month Comparison",
       subtitle = "One-year window around camera activation date",
       x = "Period",
       y = "Number of Crashes") +
  theme_minimal() +
  theme(legend.position = "none")

Analysis: Injuries

Percentage drop in injuries

# INJURIES ALONE
raw_injuries <- year_window %>%
  group_by(period) %>%
  summarize(total_injuries = sum(`NUMBER OF PERSONS INJURED`, na.rm = TRUE))

injuries_before <- raw_injuries$total_injuries[raw_injuries$period == "Before"]
injuries_after <- raw_injuries$total_injuries[raw_injuries$period == "After"]
inj_pct_drop <- ((injuries_after - injuries_before) / injuries_before) * 100
print(inj_pct_drop)
## [1] -16.74824

There was an observed 16.5% drop in injuries after cameras were installed.

T-test

inj_t_test <- t.test(total_injuries ~ treatment, data = year_window)
print(inj_t_test)
## 
##  Welch Two Sample t-test
## 
## data:  total_injuries by treatment
## t = -0.51409, df = 22444, p-value = 0.6072
## alternative hypothesis: true difference in means between group 0 and group 1 is not equal to 0
## 95 percent confidence interval:
##  -0.02381561  0.01391855
## sample estimates:
## mean in group 0 mean in group 1 
##       0.3700874       0.3750360

The average differences in injuries before & after camera activation is insignificant (the number of injuries/crash went up, even though crashes went down).

Interrupted time series (ITS)

#aggregating into months
inj_its <- analysis_df %>%
  mutate(month_rel = floor((as.numeric(crash_date_clean - activation_date_clean) / 30.44))) %>%
  group_by(month_rel) %>%
  summarize(injuries = sum(`NUMBER OF PERSONS INJURED`, na.rm = TRUE)) %>%
  mutate(
    camera_active = if_else(month_rel >= 0, 1, 0), # 0 before, 1 after
    time_after = pmax(0, month_rel)                # trend change counter
  )

#segmented poisson regression
# month_rel = overall trend | camera_active = the 'jump' | time_after = the 'slope'
inj_its_model <- glm(injuries ~ month_rel + camera_active + time_after, 
                 family = poisson, 
                 data = inj_its)

summary(inj_its_model)
## 
## Call:
## glm(formula = injuries ~ month_rel + camera_active + time_after, 
##     family = poisson, data = inj_its)
## 
## Coefficients:
##                 Estimate Std. Error z value Pr(>|z|)    
## (Intercept)    6.2843208  0.0086021  730.55   <2e-16 ***
## month_rel      0.0117401  0.0001240   94.67   <2e-16 ***
## camera_active -0.2419005  0.0137391  -17.61   <2e-16 ***
## time_after    -0.0298500  0.0002292 -130.24   <2e-16 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for poisson family taken to be 1)
## 
##     Null deviance: 28579.0  on 296  degrees of freedom
## Residual deviance:  5393.4  on 293  degrees of freedom
## AIC: 7384.2
## 
## Number of Fisher Scoring iterations: 5

Month_rel shows that injuries were increasing by ~1.2% each month prior to cameras being installed. Camera_active shows that injuries dropped by 24.2% immediately after cameras were installed. Time_after shows us that the trend flipped from injuries increasing by ~1.2% each month prior to cameras to a decrease in injuries by 2.9% each month after cameras were installed. Everything was statistically significant!

Trend in injuries before camera activation & after camera activation

inj_its$predicted <- predict(inj_its_model, type = "response")

ggplot(inj_its, aes(x = month_rel, y = injuries)) +
  geom_point(alpha = 0.3, color = "gray") +
  geom_line(aes(y = predicted, color = as.factor(camera_active)), size = 1.5) +
  geom_vline(xintercept = 0, linetype = "dashed", color = "red", size = 1) +
  scale_color_manual(values = c("0" = "blue", "1" = "darkgreen"), 
                     labels = c("Before", "After"), name = "Camera Status") +
  labs(title = "Impact of Speed Cameras on NYC Intersections",
       subtitle = "Interrupted Time Series Analysis (Month 0 = Activation)",
       x = "Months Relative to Activation",
       y = "Total Injuries") +
  theme_minimal()
## Warning: Using `size` aesthetic for lines was deprecated in ggplot2 3.4.0.
## ℹ Please use `linewidth` instead.
## This warning is displayed once every 8 hours.
## Call `lifecycle::last_lifecycle_warnings()` to see where this warning was
## generated.

Trend: Total monthly injuries (2-year window before & after camera installation)

inj_plot <- inj_its %>%
  filter(month_rel >= -24 & month_rel <= 24) %>%
  mutate(period = if_else(month_rel < 0, "Before", "After"))

ggplot(inj_plot, aes(x = month_rel, y = injuries, fill = period)) +
  geom_bar(stat = "identity", alpha = 0.8) +
  geom_vline(xintercept = -0.5, linetype = "dashed", color = "red", size = 1) +
  scale_fill_manual(values = c("Before" = "#d95f02", "After" = "#1b9e77")) +
  labs(
    title = "Total Monthly Injuries Relative to Camera Activation",
    subtitle = "Aggregated across 2,495 camera locations (Month 0 = Activation)",
    x = "Months Relative to Camera Install",
    y = "Total Injuries",
    fill = "Period"
  ) +
  theme_minimal() +
  theme(legend.position = "bottom")

Analysis: Deaths

# DEATHS ALONE
raw_deaths <- year_window %>%
  group_by(period) %>%
  summarize(total_deaths = sum(`NUMBER OF PERSONS KILLED`, na.rm = TRUE))

Percentage drop in deaths

deaths_before <- raw_deaths$total_deaths[raw_deaths$period == "Before"]
deaths_after <- raw_deaths$total_deaths[raw_deaths$period == "After"]
death_pct_drop <- ((deaths_after - deaths_before) / deaths_before) * 100
print(death_pct_drop)
## [1] -43.33333

There was an observed 43.3% drop in deaths after cameras were installed.

T-test

deaths_t_test <- t.test(total_deaths ~ treatment, data = year_window)
print(deaths_t_test)
## 
##  Welch Two Sample t-test
## 
## data:  total_deaths by treatment
## t = 1.2319, df = 23116, p-value = 0.218
## alternative hypothesis: true difference in means between group 0 and group 1 is not equal to 0
## 95 percent confidence interval:
##  -0.0004333119  0.0018993227
## sample estimates:
## mean in group 0 mean in group 1 
##     0.002362763     0.001629757

The average differences in deaths before & after camera activation is insignificant.

Interrupted time series (ITS)

death_its <- analysis_df %>%
  mutate(month_rel = floor((as.numeric(crash_date_clean - activation_date_clean) / 30.44))) %>%
  group_by(month_rel) %>%
  summarize(deaths = sum(total_deaths, na.rm = TRUE)) %>%
  mutate(
    camera_active = if_else(month_rel >= 0, 1, 0),
    time_after = pmax(0, month_rel)
  )

death_its_model <- glm(deaths ~ month_rel + camera_active + time_after, 
                     family = poisson, 
                     data = death_its)

summary(death_its_model)
## 
## Call:
## glm(formula = deaths ~ month_rel + camera_active + time_after, 
##     family = poisson, data = death_its)
## 
## Coefficients:
##                Estimate Std. Error z value Pr(>|z|)    
## (Intercept)    1.193688   0.115506  10.334   <2e-16 ***
## month_rel      0.014996   0.001805   8.309   <2e-16 ***
## camera_active -0.442536   0.186389  -2.374   0.0176 *  
## time_after    -0.030771   0.003078  -9.999   <2e-16 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for poisson family taken to be 1)
## 
##     Null deviance: 427.38  on 296  degrees of freedom
## Residual deviance: 291.89  on 293  degrees of freedom
## AIC: 722.21
## 
## Number of Fisher Scoring iterations: 5

Month_rel shows that deaths were increasing by ~1.4% each month prior to cameras being installed. Camera_active shows that deaths dropped by 44.3% immediately after cameras were installed. Time_after shows us that the trend flipped from deaths increasing by ~1.4% each month prior to cameras to a decrease in deaths by ~3% each month after cameras were installed. Everything was statistically significant!

Trend in deaths before camera activation & after camera activation

death_its$predicted <- predict(death_its_model, type = "response")

ggplot(death_its, aes(x = month_rel, y = deaths)) +
  geom_point(alpha = 0.3, color = "gray") +
  geom_line(aes(y = predicted, color = as.factor(camera_active)), size = 1.5) +
  geom_vline(xintercept = 0, linetype = "dashed", color = "red", size = 1) +
  scale_color_manual(values = c("0" = "blue", "1" = "darkgreen"), 
                     labels = c("Before", "After"), name = "Camera Status") +
  labs(title = "Impact of Speed Cameras on NYC Intersections",
       subtitle = "Interrupted Time Series Analysis (Month 0 = Activation)",
       x = "Months Relative to Activation",
       y = "Total Deaths") +
  theme_minimal()

Trend: Total monthly deaths (2-year window before & after camera installation)

death_plot <- death_its %>%
  filter(month_rel >= -24 & month_rel <= 24) %>%
  mutate(period = if_else(month_rel < 0, "Before", "After"))

ggplot(death_plot, aes(x = month_rel, y = deaths, fill = period)) +
  geom_bar(stat = "identity", alpha = 0.8) +
  geom_vline(xintercept = -0.5, linetype = "dashed", color = "red", size = 1) +
  scale_fill_manual(values = c("Before" = "#d95f02", "After" = "#1b9e77")) +
  labs(
    title = "Total Monthly Deaths Relative to Camera Activation",
    subtitle = "Aggregated across 2,495 camera locations (Month 0 = Activation)",
    x = "Months Relative to Camera Install",
    y = "Total Deaths",
    fill = "Period"
  ) +
  theme_minimal() +
  theme(legend.position = "bottom")

Analysis: Casualties (Injuries + Deaths)

# INJURIES + DEATHS COMBINED (CASUALTIES)
raw_casualties <- year_window %>%
  group_by(period) %>%
  summarize(total_casualties = sum(`NUMBER OF PERSONS INJURED`, `NUMBER OF PERSONS KILLED`, na.rm = TRUE))

Percentage drop in casualties

casualties_before <- raw_casualties$total_casualties[raw_casualties$period == "Before"]
casualties_after <- raw_casualties$total_casualties[raw_casualties$period == "After"]
cas_pct_drop <- ((casualties_after - casualties_before) / casualties_before) * 100
print(cas_pct_drop)
## [1] -16.9169

Total casualties (injuries + deaths) dropped by 16.7% after cameras were installed.

T-test

casualties_t_test <- t.test(total_casualties ~ treatment, data = year_window)
print(casualties_t_test)
## 
##  Welch Two Sample t-test
## 
## data:  total_casualties by treatment
## t = -0.43729, df = 22459, p-value = 0.6619
## alternative hypothesis: true difference in means between group 0 and group 1 is not equal to 0
## 95 percent confidence interval:
##  -0.02311096  0.01467992
## sample estimates:
## mean in group 0 mean in group 1 
##       0.3724502       0.3766657

The average differences between total casualties is insignificant (the number of injuries/crash went up, even though crashes went down).

Interrupted time-series (ITS)

cas_its <- analysis_df %>%
  mutate(month_rel = floor((as.numeric(crash_date_clean - activation_date_clean) / 30.44))) %>%
  group_by(month_rel) %>%
  summarize(casualties = sum(total_casualties, na.rm = TRUE)) %>%
  mutate(
    camera_active = if_else(month_rel >= 0, 1, 0),
    time_after = pmax(0, month_rel)
  )

cas_its_model <- glm(casualties ~ month_rel + camera_active + time_after, 
                    family = poisson, 
                    data = cas_its)

summary(cas_its_model)
## 
## Call:
## glm(formula = casualties ~ month_rel + camera_active + time_after, 
##     family = poisson, data = cas_its)
## 
## Coefficients:
##                 Estimate Std. Error z value Pr(>|z|)    
## (Intercept)    6.2903498  0.0085785  733.27   <2e-16 ***
## month_rel      0.0117561  0.0001237   95.03   <2e-16 ***
## camera_active -0.2429359  0.0137019  -17.73   <2e-16 ***
## time_after    -0.0298525  0.0002285 -130.62   <2e-16 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for poisson family taken to be 1)
## 
##     Null deviance: 28729.1  on 296  degrees of freedom
## Residual deviance:  5413.3  on 293  degrees of freedom
## AIC: 7405.7
## 
## Number of Fisher Scoring iterations: 5

Month_rel shows that total casualties were increasing by ~1.2% each month prior to cameras being installed. Camera_active shows that casualties dropped by 24.3% immediately after cameras were installed. Time_after shows us that the trend flipped from casualties increasing by ~1.2% each month prior to cameras to a decrease in casualties by ~3% each month after cameras were installed. Everything was statistically significant!

Trend in casualties before camera activation & after camera activation

cas_its$predicted <- predict(cas_its_model, type = "response")

ggplot(cas_its, aes(x = month_rel, y = casualties)) +
  geom_point(alpha = 0.3, color = "gray") +
  geom_line(aes(y = predicted, color = as.factor(camera_active)), size = 1.5) +
  geom_vline(xintercept = 0, linetype = "dashed", color = "red", size = 1) +
  scale_color_manual(values = c("0" = "blue", "1" = "darkgreen"), 
                     labels = c("Before", "After"), name = "Camera Status") +
  labs(title = "Impact of Speed Cameras on NYC Intersections",
       subtitle = "Interrupted Time Series Analysis (Month 0 = Activation)",
       x = "Months Relative to Activation",
       y = "Total Casualties") +
  theme_minimal()

Trend: Total monthly casualties (2-year window before & after camera installation)

cas_plot <- cas_its %>%
  filter(month_rel >= -24 & month_rel <= 24) %>%
  mutate(period = if_else(month_rel < 0, "Before", "After"))

ggplot(cas_plot, aes(x = month_rel, y = casualties, fill = period)) +
  geom_bar(stat = "identity", alpha = 0.8) +
  geom_vline(xintercept = -0.5, linetype = "dashed", color = "red", size = 1) +
  scale_fill_manual(values = c("Before" = "#d95f02", "After" = "#1b9e77")) +
  labs(
    title = "Total Monthly Casualties Relative to Camera Activation",
    subtitle = "Aggregated across 2,495 camera locations (Month 0 = Activation)",
    x = "Months Relative to Camera Install",
    y = "Total Casualties (Injured + Killed)",
    fill = "Period"
  ) +
  theme_minimal() +
  theme(legend.position = "bottom")