Sheds and Crime
Brook Frye
2024-01-29
Sheds and Crime
- Number of sheds (subset permit dates within +/- 8 months of complaints data & sheds that had a sidewalk adjacent) 4,402
- Total number of complaints 2019 - present: 638,828
- Total complaints that happened within 15’ building boundary of shed: 64,030
We analyzed the relationship between the presence of sheds and crime, using the available data on sheds from DOB and approximating crime using the complaint data from the NYPD.
We used building data (BIN) info to approximate the location of the sheds; we created a 10’ buffer around the area of the building adjacent to sidewalks. We subset complaints that were geolocated within the 10’ buffer 4 months before and after the erection of the shed.
To account for the changes in complaints over time (complaints have increased since pre-covid), we multiplied a ‘crime inflation factor’ to the total number of complaints/segment each year/month.
Crime Inflation Factor := S_i_t^ = S_i_t(B_t/C_t)
- S_i_t = # of complaints for a shed i at time t,
- B_t = baseline # complaints at time t
- C_t = citywide total for time t
Number of complaints per shed (*adjusted for change in crime over time)
## [1] "Total Complaints"## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 0.000 0.000 0.000 1.342 1.000 111.000## [1] "Total Complaints Before Shed"## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 0.000 0.000 0.000 0.654 1.000 55.000## [1] "Total Complaints After Shed"## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 0.000 0.000 0.000 0.709 1.000 89.000
Difference in number of complaints*
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## -21.00000 0.00000 0.00000 0.07605 0.00000 84.00000
Simple Model
- outcome is number of complaints
- predictor is whether pre or post shed
- model output is at the log scale - exponentiate coefficients
- the shedExist coefficient is interpreted as sheds are correlated with a 20% increase in crime, however the baseline amount of crime on average is about .08 (practically 0) - so a 20% increase is still 0. In places where crime is higher, that 20% increase would be impactful.
##
## Call:
## glm(formula = adj_n_bin ~ shed_exst, family = "poisson", data = bb_dt4)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -0.4458 -0.4458 -0.4065 -0.4065 7.3764
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -2.49339 0.01932 -129.080 < 0.0000000000000002 ***
## shed_exstShed 0.18453 0.02658 6.944 0.00000000000382 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for poisson family taken to be 1)
##
## Null deviance: 32541 on 62642 degrees of freedom
## Residual deviance: 32493 on 62641 degrees of freedom
## AIC: 41465
##
## Number of Fisher Scoring iterations: 6## [1] 0.08256495## [1] 1.189473## [1] 0.09820875## [1] 1.788885
Subset to only areas with high crime
##
## Call:
## glm(formula = adj_n_bin ~ shed_exst + ntaname, family = "poisson",
## data = high_crime)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -1.1650 -0.5763 -0.4626 -0.3250 6.0109
##
## Coefficients:
## Estimate Std. Error z value
## (Intercept) -4.03110 0.70740 -5.698
## shed_exstShed 0.34028 0.03531 9.636
## ntanameBedford Park-Fordham North 1.18691 0.73598 1.613
## ntanameBelmont 0.99577 0.76871 1.295
## ntanameBrownsville 0.55037 0.73485 0.749
## ntanameBushwick North 0.50715 0.79057 0.642
## ntanameCentral Harlem North-Polo Grounds 1.08999 0.71568 1.523
## ntanameCentral Harlem South 1.43190 0.71941 1.990
## ntanameChinatown 1.12388 0.72375 1.553
## ntanameClaremont-Bathgate 0.98635 0.75277 1.310
## ntanameClinton 2.23573 0.71058 3.146
## ntanameCrotona Park East 2.38081 0.71700 3.321
## ntanameCrown Heights North 0.90446 0.72956 1.240
## ntanameEast Concourse-Concourse Village 2.16964 0.71459 3.036
## ntanameEast Harlem North 2.56871 0.71094 3.613
## ntanameEast New York 1.68167 0.72032 2.335
## ntanameEast New York (Pennsylvania Ave) 1.93549 0.73030 2.650
## ntanameEast Tremont 1.99002 0.71743 2.774
## ntanameEast Williamsburg 0.26710 0.76376 0.350
## ntanameErasmus 1.92064 0.76376 2.515
## ntanameFordham South 1.86886 0.73485 2.543
## ntanameHighbridge 1.60604 0.72415 2.218
## ntanameHunts Point 2.70608 0.71253 3.798
## ntanameJamaica 0.89161 0.75593 1.179
## ntanameLongwood 1.87740 0.74755 2.511
## ntanameManhattanville 2.22117 0.72501 3.064
## ntanameMelrose South-Mott Haven North 3.30318 0.70957 4.655
## ntanameMidtown-Midtown South 2.55211 0.70799 3.605
## ntanameMorrisania-Melrose 1.27781 0.72457 1.764
## ntanameMott Haven-Port Morris 1.79604 0.71297 2.519
## ntanameMount Hope 1.82454 0.72232 2.526
## ntanameNorth Side-South Side 1.81915 0.71880 2.531
## ntanameNorwood 1.67761 0.74536 2.251
## ntanameOcean Hill 0.29603 0.86603 0.342
## ntanamepark-cemetery-etc-Bronx -10.42691 140.02130 -0.074
## ntanamepark-cemetery-etc-Brooklyn 1.22314 0.91288 1.340
## ntanamepark-cemetery-etc-Manhattan -10.43828 191.73767 -0.054
## ntanameQueensbridge-Ravenswood-Long Island City 1.13668 0.74536 1.525
## ntanameSoHo-TriBeCa-Civic Center-Little Italy 0.91259 0.71414 1.278
## ntanameSoundview-Bruckner 2.03319 0.71700 2.836
## ntanameStapleton-Rosebank 1.50679 0.80179 1.879
## ntanameVan Nest-Morris Park-Westchester Square 0.28092 0.83666 0.336
## ntanameWest Concourse 1.78192 0.71880 2.479
## ntanameWest Village 1.81091 0.70923 2.553
## ntanameWilliamsbridge-Olinville 1.49102 0.75277 1.981
## Pr(>|z|)
## (Intercept) 0.0000000121 ***
## shed_exstShed < 0.0000000000000002 ***
## ntanameBedford Park-Fordham North 0.106811
## ntanameBelmont 0.195187
## ntanameBrownsville 0.453884
## ntanameBushwick North 0.521198
## ntanameCentral Harlem North-Polo Grounds 0.127755
## ntanameCentral Harlem South 0.046547 *
## ntanameChinatown 0.120456
## ntanameClaremont-Bathgate 0.190100
## ntanameClinton 0.001653 **
## ntanameCrotona Park East 0.000898 ***
## ntanameCrown Heights North 0.215072
## ntanameEast Concourse-Concourse Village 0.002396 **
## ntanameEast Harlem North 0.000303 ***
## ntanameEast New York 0.019564 *
## ntanameEast New York (Pennsylvania Ave) 0.008043 **
## ntanameEast Tremont 0.005540 **
## ntanameEast Williamsburg 0.726552
## ntanameErasmus 0.011913 *
## ntanameFordham South 0.010984 *
## ntanameHighbridge 0.026566 *
## ntanameHunts Point 0.000146 ***
## ntanameJamaica 0.238203
## ntanameLongwood 0.012025 *
## ntanameManhattanville 0.002187 **
## ntanameMelrose South-Mott Haven North 0.0000032364 ***
## ntanameMidtown-Midtown South 0.000312 ***
## ntanameMorrisania-Melrose 0.077809 .
## ntanameMott Haven-Port Morris 0.011766 *
## ntanameMount Hope 0.011538 *
## ntanameNorth Side-South Side 0.011379 *
## ntanameNorwood 0.024402 *
## ntanameOcean Hill 0.732478
## ntanamepark-cemetery-etc-Bronx 0.940639
## ntanamepark-cemetery-etc-Brooklyn 0.180287
## ntanamepark-cemetery-etc-Manhattan 0.956584
## ntanameQueensbridge-Ravenswood-Long Island City 0.127256
## ntanameSoHo-TriBeCa-Civic Center-Little Italy 0.201291
## ntanameSoundview-Bruckner 0.004572 **
## ntanameStapleton-Rosebank 0.060205 .
## ntanameVan Nest-Morris Park-Westchester Square 0.737048
## ntanameWest Concourse 0.013174 *
## ntanameWest Village 0.010669 *
## ntanameWilliamsbridge-Olinville 0.047625 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for poisson family taken to be 1)
##
## Null deviance: 16654 on 22804 degrees of freedom
## Residual deviance: 15072 on 22760 degrees of freedom
## AIC: 19952
##
## Number of Fisher Scoring iterations: 12## [1] 0.1693344Negative Binomial
- Used to model over-dispersed count data - an extra parameter is used to model variance
## [1] 1.004449