Use Census geographic data in R with tigris

class: center, middle, inverse, title-slide

# Use Census geographic data in R with tigris
### Kyle Walker, Texas Christian University
### April 4, 2017

---

## Follow along!

* https://walkerke.github.io/tigris-webinar

* https://github.com/walkerke/tigris-webinar

* _R Journal_ article: [tigris: an R Package to Access and Work with Geographic Data from the US Census Bureau](https://journal.r-project.org/archive/2016-2/walker.pdf)

---

## Get ready:

To follow along, open `R/examples.R` in the GitHub repository and install the required packages (some from GitHub)

---

## Disclaimer

This webinar, and the __tigris__ package, uses Census Bureau Data but is not endorsed or certified by the Census Bureau.

---

## TIGER/Line shapefiles

.footnote[Source: Wikimedia Commons]

---

## The Census hierarchy

.footnote[Source: [US Census Bureau](http://www2.census.gov/geo/pdfs/reference/geodiagram.pdf)]

---

## Types of TIGER/Line datasets

* Legal entities

* Statistical entities

* Geographic features

---

## Accessing TIGER/Line shapefiles

[Files accessible from dropdown menu interface](https://www.census.gov/cgi-bin/geo/shapefiles/index.php)

---

## Accessing TIGER/Line shapefiles

[Files can be downloaded directly from web or FTP interfaces](http://www2.census.gov/geo/tiger/TIGER2015/)

---

## The __tigris__ package

From CRAN:

```r
install.packages("tigris")
```

Or from GitHub (with __sf__ support):

```r
devtools::install_github("walkerke/tigris")
```
Let's get started!

```r
library(tigris)
```
---

## Basic functionality

```r
us <- states()

plot(us)
```

---

## How __tigris__ works

When you call a __tigris__ function, it does the following:

* _Downloads_ your data from the US Census Bureau website

* _Stores_ your data in a user cache directory (via __rappdirs__) or in a temporary directory

* _Loads_ your data into your R session with `rgdal::readOGR()` _or_ `sf::st_read()`

---

### TIGER/Line vs. cartographic boundary files

```r
ri <- counties("RI")

ri20 <- counties("RI", cb = TRUE, resolution = "20m")

plot(ri)
plot(ri20, border = "red", add = TRUE)
```

---

## Example: Zip Code Tabulation Areas (ZCTAs)

```r
fw_zips <- zctas(cb = TRUE, starts_with = "761")

plot(fw_zips)
```

---

## Example: roads in Loving County, TX

```r
loving <- roads("TX", "Loving")

plot(loving)
```

---

### Combining objects with `rbind_tigris`

```r
sts <- c("DC", "MD", "VA")

combined <- rbind_tigris(
  lapply(sts, function(x) {
    tracts(x, cb = TRUE)
  })
)

plot(combined)
```

---

### Combining objects with `rbind_tigris`

---

### Merging data with `geo_join`

```r
df <- read.csv("http://personal.tcu.edu/kylewalker/data/txlege.csv",
              stringsAsFactors = FALSE)
              
districts <- state_legislative_districts("TX", house = "lower", 
                                         cb = TRUE)

txlege <- geo_join(districts, df, "NAME", "District")

txlege$color <- ifelse(txlege$Party == "R", "red", "blue")

plot(txlege, col = txlege$color)
legend("topright", legend = c("Republican", "Democrat"),
      fill = c("red", "blue"))
```

---

## Merging data with `geo_join`

---

### Thematic mapping: the __tmap__ package

Let's get some data:

```r
library(tigris)
library(censusapi)
library(tidyverse)
library(tmap)

chi_counties <- c("Cook", "DeKalb", "DuPage", "Grundy", "Lake", 
                  "Kane", "Kendall", "McHenry", "Will County")

chi_tracts <- tracts(state = "IL", county = chi_counties, 
                     cb = TRUE)

key <- "YOUR KEY GOES HERE"

data_from_api <- getCensus(name = "acs5", vintage = 2015, 
                           key = key, vars = "B25077_001E", 
                           region = "tract:*", 
                           regionin = "state:17")
```

---

### Static mapping with __tmap__

```r
values <- data_from_api %>%
  transmute(GEOID = paste0(state, county, tract), 
            value = B25077_001E)

chi_joined = geo_join(chi_tracts, values, by = "GEOID")

tm_shape(chi_joined, projection = 26916) +
  tm_fill("value", style = "quantile", n = 7, palette = "Greens", 
          title = "Median home values \nin the Chicago Area") + 
  tm_legend(bg.color = "white", bg.alpha = 0.6) + 
  tm_style_gray()
```

---

### Static mapping with __tmap__

---

### Interactive mapping with __tmap__

For an interactive Leaflet map, run:

```r
ttm()
```

then run the __tmap__ code as before.

---

## Simple features for R

* The future of spatial data in R

* Spatial data represented as R data frames, with geometry stored in a list-column

* Learn more: http://edzer.github.io/sfr/

---

## Simple features in __tigris__

Add `class = "sf"` to your __tigris__ function call

or:

At the beginning of your script, specify `options(tigris_class = "sf")`

---

### Exploratory spatial analysis pipelines with __sf__

Question: how do home values in the Chicago area vary with distance from downtown?

```r
library(sf)

city_hall <- c(-87.631969, 41.883835) %>%
  st_point() %>%
  st_sfc(crs = 4269) %>%
  st_transform(26916)
```
---

### Exploratory spatial analysis pipelines with __sf__

```r
options(tigris_class = "sf")
chi_tracts_sf <- tracts(state = "IL", county = chi_counties, 
                        cb = TRUE)

chi_tracts_sf %>%
  st_transform(26916) %>%
  left_join(values, by = "GEOID") %>%
  mutate(dist = as.numeric(
    st_distance(
      st_centroid(.), city_hall
      )
    )) %>%
  ggplot(aes(x = dist, y = value)) + 
  geom_smooth(span = 0.3, method = "loess")
```

---

### Exploratory spatial analysis pipelines with __sf__

---

## Simple features and __ggplot2__

---

## Questions?

* Web: <http://personal.tcu.edu/kylewalker>

* Consulting/blog: <https://walkerke.github.io>

* Twitter: @kyle_e_walker

<style>

h1, h2, h3 {
  color: #035004; 
}

a {
  color: #1a730f; 
}

.inverse {
  background-color: #035004;

}
</style>

<!--

Outline:

* Basics of Census shapefiles
* How tigris works
* Core functionality - loading datasets as spatial data frames
* Helper functions: geo_join and rbind_tigris
* Making maps with tigris data: tmap and leaflet
* Simple features for R
* tigris and simple features
* Spatial analysis with sf using tigris data
* Mapping with tigris data and geom_sf()

-->