032_thematic_mapping_geom_sf.Rmd

---
title: "Thematic Mapping with geom_sf"
output: html_document
---

from:  https://map-rfun.library.duke.edu/032_thematic_mapping_geom_sf.html

```{r load-libraries, message=FALSE, warning=FALSE}
library(tidyverse)       # Tidyverse for Tidy Data
library(readxl)
library(tigris)
library(sf)
library(viridis)
library(tigris) 
```


## Shapefiles as sf 

Repeating steps from the [previous section](031_thematic_mapping.html), using the `tigris` package, get Census Tiger shapefiles for census geographies.  Coastal boundaries can be gathered with the tigris argument:  `cb = TRUE`.

```{r getuscb, message=FALSE, warning=FALSE, include=FALSE}
us_geo <- tigris::states(class = "sf", cb = TRUE)
```
``` r 
us_geo <- tigris::states(class = "sf", cb = TRUE)
```

## Get BLS data

As mentioned before, the data are from the Bureau of Labor Statistics.  These data are stored in an excel file in the `data` directory of the [repository](https://github.com/libjohn/mapping-with-R):  `data/OES_Report.xlsx`.  

Again, from the previous section you will see how these data were gathered, loaded, transformed, and joined. 

```{r getblsdata, message=FALSE, warning=FALSE}
Salary4Helpers <- 
  read_excel("data/OES_Report.xlsx",
             col_types = c("text", "numeric"), 
             skip = 4)

Salary4Helpers
```

## Wrangle the data

As before...  

```{r wrangleblsdata, message=FALSE, warning=FALSE}
BlsWage_ToJoin <- Salary4Helpers %>% 
  rename(wages = "Annual mean wage(2)") %>% 
  mutate(State = str_extract(`Area Name`, "\\w+.*(?=\\()")) %>% 
  drop_na(wages) %>% 
  select(State, wages)
```


```{r}
as_tibble(us_geo) %>% 
  arrange(NAME)
```


## Join data

Use the `dplyr::left_join` function to append BLS variable to the `sf` tibble (data frame). 

```{r leftjoin_usgeo, warning=FALSE}
HelperShapeObject <- us_geo %>% 
  left_join(BlsWage_ToJoin,
            by = c("NAME" = "State"))
```


## Contiguous 48 states

Filter to only the contiguous 48 states + D.C.

```{r filterlower48, message=FALSE, warning=FALSE}
contiguous_states <- HelperShapeObject %>% 
  filter(GEOID < 60) %>% 
  filter(STUSPS != "AK") %>% 
  filter(STUSPS != "HI")
```



## ggplot2 with geom_sf and viridis

In this section we introduce making shapefiles with ggplot2.  ggplot2 is one of the more popular and broadly distributed graphics packages used in the R community. ([Learn more](https://rfun.library.duke.edu/#portfolio) about ggplot2. 

In this plot I reversed the direction of the color scale.  After consulting with my visualization colleagues it seems this may have been a non-standard action on my part.  But, I leave the `direction` argument here for the curious.  

Use a pleasing projection, in this case assigned the crs projection to '5070' (`coord_sf(crs = 5070)`), and removed the gridlines (i.e [graticules](https://en.wikipedia.org/wiki/Graticule)) `coords_sf(datum = NA)` 

```{r with-ggplot-geom_sf, message=FALSE, warning=FALSE}
contiguous_states %>% 
  ggplot(aes(fill = wages, color = wages)) +
  geom_sf() +
  coord_sf(crs = 5070, datum = NA) +
  scale_fill_viridis(direction = -1, label = scales::dollar) + 
  scale_color_viridis(direction = -1, label = scales::dollar) +
  labs(title = "Annual Mean Wages by State",
       subtitle = "Mental Health and Substance Abuse Social Workers(SOC Code211023)", 
       caption = "Data Source: BLS.gov - Occupational Employment Statistics ; 2016")
  
```


## End Notes

This session inspired by  https://www.computerworld.com/article/3175623/data-analytics/mapping-in-r-just-got-a-whole-lot-easier.html