Test of Rpubs

rpubs-test.rmd

This document is just a test of RPubs platform for publishing R Markdown. I'll use the `survey` dataset from the `MASS` package to explore a few features.

```{r get_data, message=FALSE}
library(MASS)
library(psych)
library(Hmisc)
library(xtable)
data(survey)
```


# Overview of the survey dataset
So let's have a look at this dataset

```{r}
head(survey)
```

As you can see, it contains an assortment of measurements for `r nrow(survey)` individuals.


```{r describe_data}
Hmisc::describe(survey)
```

# Are males taller?

```{r}
boxplot(Height ~ Sex, survey)
fit1 <- lm(Height ~ Sex, survey)
summary(fit1)
```

Yes, as expected, males are taller. The expected difference is `r round(coef(fit1)['SexMale'], 1)` centimetres or  `r round(coef(fit1)['SexMale'] /  2.54, 1) ` inches.

# Is smoking related to exercise?

```{r results='asis'}
survey$ExerSome <- survey$Exer %in% c("Freq", "Some")
tab1 <- table(smoke=survey$Smoke, exercise=survey$ExerSome)

print(xtable(tab1, caption='Cell counts'), 'html')
print(xtable(round(prop.table(tab1, 1), 2), caption='Proportion within smoker category'), 'html')
```

```{r}
chisq.test(tab1)
```

No, it doesn't look like it.

# What are the correlations between the different physical measurements?
```{r}
pairs.panels(survey[, c("Wr.Hnd", "NW.Hnd", "Height")])
```

This above figure shows the pearson correlation coefficient between writing hand span (`Wr.Hnd`), non-writing hand span (`NW.Hnd`) and actual height (`Height`).


# Conclusion
As I said at the start, this is just my first test of the RPub publishing platform for R Markdown.