Event

There are two events within this race, the dog version and the human version.

DRAT4AS = “Doggie Run Across Tennessee For Animal Shelters” and GVRAT = human version.

table(gv_roster$Event)

## 
## DRAT4AS   GVRAT 
##     295   19016

Gender

In lots of races there are more women than men, and this one is no different.

table(gv_roster$G, gv_roster$Event)

##    
##     DRAT4AS GVRAT
##   F     165 10738
##   M     130  8278

prop.table(table(gv_roster$G, gv_roster$Event), 2)

##    
##       DRAT4AS     GVRAT
##   F 0.5593220 0.5646824
##   M 0.4406780 0.4353176

ggplot(gv_roster, aes(x=G, fill = G)) + geom_bar()

Definitely more ladies than men, even doggy ladies, but interestingly enough the same percentage of doggy ladies as human ladies.

chisq.test(table(gv_roster$G, gv_roster$Event))

## 
##  Pearson's Chi-squared test with Yates' continuity correction
## 
## data:  table(gv_roster$G, gv_roster$Event)
## X-squared = 0.01565, df = 1, p-value = 0.9004

Age

Taking out the doggies, we can look at the age distribution.

gv_roster %>% filter(Event == "GVRAT") %>%
  ggplot(aes(x=A)) + 
    geom_histogram(binwidth = 1)

## Warning: Removed 2 rows containing non-finite values (stat_bin).

The mean participant age is 43.3639942. The median age of human participants is 43.

I am exactly the median age for this race.

Age vs. Gender

The distributions of ages by gender look remarkably similar.

gv_roster %>% filter(Event == "GVRAT") %>%
  ggplot(aes(x=A, fill = G)) + geom_histogram(binwidth = 1) +
  facet_grid(G ~ .)

## Warning: Removed 2 rows containing non-finite values (stat_bin).

Looks normally distributed enough to just use normal distribution tests but I went with a t-test anyway. Men were about 7 months older than women in this race.

t.test(gv_roster$A ~ gv_roster$G) # includes doggos

## 
##  Welch Two Sample t-test
## 
## data:  gv_roster$A by gv_roster$G
## t = -3.6332, df = 17470, p-value = 0.0002808
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  -0.9128226 -0.2730463
## sample estimates:
## mean in group F mean in group M 
##        43.10585        43.69879

I guess this shows that with a large enough sample, even small differences in age become significant.

Total Miles vs. Gender

gv_roster %>% filter(Event == "GVRAT") %>%
  ggplot(aes(x=Miles, fill = G)) + 
  geom_histogram(binwidth = 5) + 
  facet_grid(G~.)

t.test(Miles ~ G, dat = gv_roster)

## 
##  Welch Two Sample t-test
## 
## data:  Miles by G
## t = -14.472, df = 16078, p-value < 2.2e-16
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  -15.03826 -11.45047
## sample estimates:
## mean in group F mean in group M 
##        104.0505        117.2949

Interestingly men had run a little farther than women on average (including the 0 mile people).

Age vs. Mileage

gv_roster %>% filter(Event == "GVRAT") %>% 
  ggplot(aes(x = A, y = Miles)) + 
  geom_point(alpha = 0.01, color = "blue")

## Warning: Removed 2 rows containing missing values (geom_point).

There didn’t seem to be a significant correlation between age and miles run, although on the regression, the older runners were running more in a statistically significant way. The regression parameter estimate is 0.84 total miles/year of age.

summary(lm(Miles ~ A, data = gv_roster))

## 
## Call:
## lm(formula = Miles ~ A, data = gv_roster)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -149.21  -33.07   -2.92   25.98  523.93 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept) 73.18551    1.76606   41.44   <2e-16 ***
## A            0.84467    0.03944   21.41   <2e-16 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 61.13 on 19306 degrees of freedom
##   (4 observations deleted due to missingness)
## Multiple R-squared:  0.0232, Adjusted R-squared:  0.02315 
## F-statistic: 458.6 on 1 and 19306 DF,  p-value: < 2.2e-16

Runners from Hawaii

Here’s the runners from my home state of Hawaii:

gv_roster %>% filter(grepl("Hawaii", Home))

## # A tibble: 42 x 7
##      Bib name             Event G         A Home         Miles
##    <dbl> <chr>            <fct> <fct> <dbl> <fct>        <dbl>
##  1  9385 Duane Zitta      GVRAT M        39 USA - Hawaii  261.
##  2 11868 Sally Cravens    GVRAT F        61 USA - Hawaii  192.
##  3 17429 Aaron Reed       GVRAT M        28 USA - Hawaii  182.
##  4  6786 April Shrum      GVRAT F        42 USA - Hawaii  169.
##  5 16019 Kana Yamamoto    GVRAT F        45 USA - Hawaii  158.
##  6  6906 Robert Davidson  GVRAT M        37 USA - Hawaii  155.
##  7 16629 michael hee      GVRAT M        60 USA - Hawaii  154.
##  8 11636 Sean Thompson    GVRAT M        39 USA - Hawaii  150.
##  9 13815 Andrew Borrebach GVRAT M        29 USA - Hawaii  143.
## 10  7202 Chris Robinson   GVRAT M        39 USA - Hawaii  134.
## # … with 32 more rows

Miles by State

gv_roster %>% filter(Event == "GVRAT") %>%
  filter(grepl("USA", Home)) %>%
  ggplot(aes(x = reorder(Home, Miles, FUN = median),y = Miles)) +
  geom_boxplot() +
  theme(axis.text.x = element_text(angle = 90, vjust = 0.5, hjust = 1)) + 
  xlab("State")

Hawaii is the 49th place USA location by median miles. It’s hard to tell from the boxplots but the difference in miles between Hawaii and Nebraska is over 20 miles, so it’s a pretty significant percent difference.

gv_roster %>% filter(Event == "GVRAT") %>%
  filter(grepl("USA", Home)) %>%
  group_by(Home) %>%
  summarize(medianMiles = median(Miles)) %>%
  arrange(-medianMiles)

## # A tibble: 53 x 2
##    Home                medianMiles
##    <fct>                     <dbl>
##  1 USA - Nebraska             118.
##  2 USA - West Virginia        117.
##  3 USA - APO                  112.
##  4 USA - Alabama              112.
##  5 USA - North Dakota         112.
##  6 USA - Colorado             112.
##  7 USA - New Mexico           111.
##  8 USA - Connecticut          109.
##  9 USA - Kentucky             108.
## 10 USA - Florida              108.
## # … with 43 more rows