The Spatial Relationship Between Campsites, Homelessness Services, and Drug Offenses by Portland, OR Neighborhoods between 2025-2025
Group 5 - Logan Rossell & Talia Walderman
The number of homeless individuals in Portland, OR has increased over time from 11,430 at the start of 2024 to 18,260 individuals, according to the Homeless Services Department. We sought to explore the spatial relationship between homelessness in a neighborhood, available services, and documented crime in an effort to understand where areas of Portland, OR are underserving its unsheltered residents.
Regression
Overly Simplistic Regression
Throwing all our variables into a linear regression produced a model with a high R-squared value, but it does not satisfy the assumptions of linear regression, thus making this a misleading model.
##
## Call:
## lm(formula = clean_site_count ~ `Assault Offenses` + Burglary +
## `Extortion/Blackmail` + shelter_count + crime_count + Embezzlement +
## `Counterfeiting/Forgery` + `Homicide Offenses` + `Drug/Narcotic Offenses` +
## `Human Trafficking Offenses` + `Stolen Property Offenses`,
## data = combined_counts)
##
## Residuals:
## Min 1Q Median 3Q Max
## -156.656 -58.857 -3.524 46.199 289.891
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -18.18396 19.28548 -0.943 0.35019
## `Assault Offenses` 1.55444 0.52509 2.960 0.00465 **
## Burglary 2.89956 0.54994 5.273 2.75e-06 ***
## `Extortion/Blackmail` -73.88830 22.58881 -3.271 0.00192 **
## shelter_count 19.57547 16.54793 1.183 0.24231
## crime_count -0.12075 0.05049 -2.392 0.02051 *
## Embezzlement -26.09480 17.40983 -1.499 0.14008
## `Counterfeiting/Forgery` -15.99550 4.56316 -3.505 0.00096 ***
## `Homicide Offenses` 39.26796 11.86486 3.310 0.00172 **
## `Drug/Narcotic Offenses` 3.49768 1.67977 2.082 0.04236 *
## `Human Trafficking Offenses` -84.89174 43.24575 -1.963 0.05511 .
## `Stolen Property Offenses` 39.12280 22.10132 1.770 0.08268 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 88.08 on 51 degrees of freedom
## (59 observations deleted due to missingness)
## Multiple R-squared: 0.9038, Adjusted R-squared: 0.883
## F-statistic: 43.55 on 11 and 51 DF, p-value: < 2.2e-16Nuanced Regression
Accounting for population variation between neighborhoods and for the non-linearity of the data produces a more accurate but less intuitive model, which serves as a reminder to question the preconceived notions we all hold about homelessness.
##
## Call:
## lm(formula = log_clean_sites_per_capita ~ Vandalism_per_capita +
## Burglary_per_capita + `Weapon Law Violations_per_capita` +
## library_count_per_capita + pov, data = model_data)
##
## Residuals:
## Min 1Q Median 3Q Max
## -1.46182 -0.43756 -0.01352 0.48763 1.69611
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -5.3353 0.2324 -22.956 <2e-16 ***
## Vandalism_per_capita 33.7920 15.4288 2.190 0.0329 *
## Burglary_per_capita 36.0906 13.5229 2.669 0.0101 *
## `Weapon Law Violations_per_capita` 283.0522 110.1695 2.569 0.0130 *
## library_count_per_capita 1400.8274 1387.6714 1.009 0.3173
## pov -0.5597 1.8190 -0.308 0.7595
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.7402 on 53 degrees of freedom
## (63 observations deleted due to missingness)
## Multiple R-squared: 0.61, Adjusted R-squared: 0.5733
## F-statistic: 16.58 on 5 and 53 DF, p-value: 7.876e-10Visualizations
Choropleth Map - Homeless Campsites Cleaned by the IRP
What is the IRP?
“The Impact Reduction Program is designed to be a stopgap measure that addresses the issues of hygiene, waste management, and public access that arise as a result of unsheltered homelessness. The program treats the symptoms of the problem of homelessness, but is not designed to be a root cause solution to homelessness.” Source: https://www.portland.gov/homelessness-impact-reduction/overview
Homelessness Services Offered in Portland, OR
