One-way ANOVA

The ANOVA test is based on the cleaned data from regression step, dedicated to show the association between variables and the outcome.

showing = function(x){
  x %>% 
  broom::tidy() %>%
  mutate(
    p.value = format(p.value, scientific = TRUE, digits = 3)
  ) %>% 
  select(term, p.value) %>% 
  rows_delete(tibble(term = "Residuals")) %>% 
  rename(variable = term) %>% 
  knitr::kable()
}

Year

res_year = aov(number_shoot ~ factor(year), data = nyc_fit_lm)
showing(res_year)

variable	p.value
factor(year)	7.81e-39

The p-value of the year variable is smaller than 0.05, thus we reject the null hypothesis and regard the year as statistically significant.

Month

res_month = aov(number_shoot ~ factor(month), data = nyc_fit_lm)
showing(res_month)

variable	p.value
factor(month)	1.04e-13

The p-value of the month variable is smaller than 0.05, thus we reject the null hypothesis and regard the month as statistically significant.

Borough

nyc_boro = nyc_fit_lm %>% 
  rename(borough = boro)
res_boro = aov(number_shoot ~ factor(borough), data = nyc_boro)
showing(res_boro)

variable	p.value
factor(borough)	3.08e-115

The p-value of the borough variable is smaller than 0.05, thus we reject the null hypothesis and regard the borough as statistically significant.

Statistical_murder_flag

res_smf = aov(number_shoot ~ factor(statistical_murder_flag), data = nyc_fit_lm)
showing(res_smf)

variable	p.value
factor(statistical_murder_flag)	6.97e-32

The p-value of the Statistical_murder_flag variable is smaller than 0.05, thus we reject the null hypothesis and regard the Statistical_murder_flag as statistically significant.

Perp_sex

res_perpsex = aov(number_shoot ~ factor(perp_sex), data = nyc_fit_lm)
showing(res_perpsex)

variable	p.value
factor(perp_sex)	2.5e-41

The p-value of the sex of perpetrator variable is smaller than 0.05, thus we reject the null hypothesis and regard the sex of perpetrator as statistically significant.

Perp_race

res_perprace = aov(number_shoot ~ factor(perp_race), data = nyc_fit_lm)
showing(res_perprace)

variable	p.value
factor(perp_race)	6.74e-93

The p-value of the race of perpetrator variable is smaller than 0.05, thus we reject the null hypothesis and regard the race of perpetrator as statistically significant.

Vic_sex

res_vicsex = aov(number_shoot ~ factor(vic_sex), data = nyc_fit_lm)
showing(res_vicsex)

variable	p.value
factor(vic_sex)	1.6e-34

The p-value of the sex of victims variable is smaller than 0.05, thus we reject the null hypothesis and regard the sex of victims as statistically significant.

Vic_race

res_vicrace = aov(number_shoot ~ factor(vic_race), data = nyc_fit_lm)
showing(res_vicrace)

variable	p.value
factor(vic_race)	1.35e-131

The p-value of the race of victims variable is smaller than 0.05, thus we reject the null hypothesis and regard the race of victims as statistically significant.

Perp_age_group

res_perp_age_group = aov(number_shoot ~ factor(perp_age_group), data = nyc_fit_lm)
showing(res_perp_age_group)

variable	p.value
factor(perp_age_group)	7.24e-56

The p-value of the perp_age_group variable is smaller than 0.05, thus we reject the null hypothesis and regard the perp_age_group as statistically significant.

Vic_age_group

res_vic_age_group = aov(number_shoot ~ factor(vic_age_group), data = nyc_fit_lm)
showing(res_vic_age_group)

variable	p.value
factor(vic_age_group)	5.3e-41

The p-value of the vic_age_group variable is smaller than 0.05, thus we reject the null hypothesis and regard the vic_age_group as statistically significant.