r-transport-applications.Rmd

---
title: "Data carpentry for sustainable transport planning: the Propensity to Cycle Tool"
author: "Robin Lovelace, University of Leeds, \n `r Sys.Date()`"
date: "Slides: [rpubs.com/RobinLovelace](http://rpubs.com/RobinLovelace)"
output:
  ioslides_presentation:
    widescreen: true
    transition: slower
    logo: ../figures/its-logo-square.png
    keep_md: true
---

```{r}
knitr::opts_knit$set(root.dir = "..")
```


## The nature of transport data

- The same as handling any other kind of data in R
- But: dynamic, complex (e.g. GTFS) reliant on official feeds
- That's why using a language that is powerful and flexible is good

## 

## Worked example: mode of travel in Bristol

- We'll load and visualise some transport data
- Get set-up with the pkgs:

```{r}
devtools::install_github("ropensci/stplanr", ref = "sfr") # dev version of stplanr
library(tidyverse)
library(stplanr)
library(sf)
```

## Load and inspect the zone data

```{r}
z = readRDS("gis4ta-data/z.Rds")
z = st_as_sf(z) # convert to an sf object
plot(z[c("all", "car_driver", "bicycle", "foot")])
```

## Download and preprocess the OD data

```{r, eval=FALSE}
od_data = read_csv("https://media.githubusercontent.com/media/npct/pct-outputs-regional-notR/master/commute/msoa/avon/od_attributes.csv")
od_data = od_data %>% filter(all > 50, geo_code1 %in% z$geo_code,
                             geo_code2 %in% z$geo_code) 
od_data$id = NULL
write_csv(od_data, "gis4ta-data/od_attributes.csv")
```

## Convert OD data to geographical 'desire lines'

```{r, message=FALSE}
od = read_csv("gis4ta-data/od_attributes.csv")
l = od2line(flow = od, zones = z)
plot(l$geometry)
```

## Analysing where people walk

```{r}
sel_walk = l$foot > 19
l_walk = l[sel_walk,]
plot(l$geometry)
plot(l_walk, add = T, col = "red")
```

## Subsetting in the tidyverse

```{r}
l_walk2 = l %>% 
  filter(bicycle > 19)
plot(l$geometry)
plot(l_walk2$geometry, add = T, col = "green")
```

## Exercises 

- Find and plot all desire lines where cycling is more common than walking 
- Find all desire which are less than 2.5 km 
- Find all lines in which driving is more common than walking AND are 
- Find and plot all desire lines that are longer. Where are they?
- Use the following code to download bus stops in the area

## Solution

```{r}
l$distsf = as.numeric(st_length(l))
summary(l$distsf)
l_drive_short = l %>% 
  filter(distsf < 1000 & car_driver > foot)
l_drive_short2 = l %>% 
  filter(distsf < 1000) %>% 
  filter(car_driver > foot)
l_short = l %>% 
  filter(distsf < 1000)
plot(l_drive_short[6])
```


```{r, eval=FALSE}
devtools::install_github("osmdatar/osmdata")
```


## Advanced challenge: download transport data with osmdata 

- You'll need to read-up on the osmdata vignette

```{r}
library(osmdata)
q = opq(bbox = "Leeds") %>% 
  add_feature(key = "railway", value = "station")
stations = osmdata_sf(q = q)
plot(l[6])
plot(stations$osm_points, add = T)
```