seasonal-trends.Rmd

---
title: "Divvy usage over the year"
author: "Peter Carbonetto"
output: html_document
---

<!-- Define defaults shared by all workflowr files. -->
```{r read-chunk, include=FALSE, cache=FALSE}
knitr::read_chunk("chunks.R")
```

<!-- Update knitr chunk options -->
```{r knitr-opts-chunk, include=FALSE}
```

<!-- Insert the date the file was last updated -->
```{r last-updated, echo=FALSE, results='asis'}
```

<!-- Insert the code version (Git commit SHA1) if Git repository
 exists and R package git2r is installed -->
```{r code-version, echo=FALSE, results='asis'}
```

In this last analysis, I use the Divvy trip data to examine biking
trends in Chicago over the course of one year.

I begin by loading a few packages, as well as some additional
functions I wrote for this project.

```{r load-packages, message = FALSE}
library(data.table)
library(ggplot2)
source("../code/functions.R")
```
<br>

## Read the data

First, I read in the Divvy trip and station data from the CSV files.

```{r load-data}
divvy <- read.divvy.data()
```

I would like to analyze city-wide departures for each day of the year,
so I create a new "day of year" column.

```{r day-of-year}
divvy$trips <-
  transform(divvy$trips,
    start.dayofyear = factor(as.numeric(format(divvy$trips$starttime,"%j")),
                             1:366))
```

I also convert the "start week" column to a factor to make it easier
to compile trip statistics for each week in the year.

```{r convert-start-week}
divvy$trips <- transform(divvy$trips,start.week = factor(start.week,0:52))
```
<br>

## Plot departures per day and per week

Here, I create a new vector containing the number of trips taken in each
day of the year, and then I plot these numbers.

```{r plot-trips-by-day, fig.height = 3, fig.width = 6}
counts.day <- as.vector(table(divvy$trips$start.dayofyear))
ggplot(data.frame(day = 1:366,departures = counts.day),
       aes(x = day,y = departures)) +
  geom_point(color = "darkblue",shape = 19,size = 1) +
  geom_line(color = "darkblue") +
  scale_x_continuous(breaks = seq(0,350,25)) +
  theme_minimal() +
  theme(panel.grid.major = element_blank(),
        panel.grid.minor = element_blank())
```

This plot shows a sizeable increase in bike trips during summer days,
but since the number of trips varies widely from one day to the next,
I think the plot will look nicer if instead we count the number of
trips *per week*.

```{r plot-trips-by-week, fig.height = 3, fig.width = 6}
counts.week <- as.vector(table(divvy$trips$start.week))
ggplot(data.frame(week = 0:52,departures = counts.week),
       aes(x = week,y = departures)) +
  geom_point(color = "darkblue",shape = 19,size = 1.5) +
  geom_line(color = "darkblue") +
  theme_minimal() +
  theme(panel.grid.major = element_blank(),
        panel.grid.minor = element_blank())
```

Indeed, the seasonal trends are less noisy in this plot; the majority
of Divvy bike trips in Chicago are taken when the weather is warmer
(weeks 20--40), and very few people are using the Divvy bikes in the
cold winter months.

<br>

## Seasonal trends at the University of Chicago

When we analyze trips taken at the University of Chicago bike station,
the "bump" during warmer months flattens out. This is probably because
a large fraction of University of Chicago students leave during the
summer.

```{r plot-trips-by-week-uchicago, fig.height = 3, fig.width = 6}
dat <- subset(divvy$trips,from_station_name == "University Ave & 57th St")
counts.week.uchicago <- as.vector(table(dat$start.week))
ggplot(data.frame(week = 0:52,departures = counts.week.uchicago),
       aes(x = week,y = departures)) +
  geom_point(color = "darkblue",shape = 19,size = 1.5) +
  geom_line(color = "darkblue") +
  theme_minimal() +
  theme(panel.grid.major = element_blank(),
        panel.grid.minor = element_blank())

```

<br>

## Session information

This is the version of R and the packages that were used to generate
these results.

```{r session-info}
```