Let’s Lend Amazon A Hand

Amazon has announced that it will be building a second headquarters in the United States but has left it up to America’s cities to put forward their best bid to woo the retail giant. Amazon laid out some of the criteria they are looking for, including a business friendly environment, located near a large commercial airport, and a population over 1 million.

Many cities have put their names forward including some big locations such as Boston, Chicago, San Francisco, and Atlanta whom all seem to check the boxes Amazon is looking for. I figured we’d use some of what we’ve learned in class to help Amazon find the best candidate.

It’s All About Leadership

Using the Bureau of Labor Statistics copious amounts of data and shapefiles provided by The US Census Bureau I created a map showing the number of management jobs (based on NAICS Code) in each state. To get a true measurement that discounts state size the chloropleth’s shading shading weights are based on jobs per 1,000.

Management Jobs By US State

st_data <- read_excel("./oesm16st/state_M2016_dl.xlsx")
st_data_m <- st_data %>% filter(OCC_TITLE == "Management Occupations") %>% select(STATE,JOBS_1000, TOT_EMP)
st_data_m$JOBS_1000 <- as.numeric(st_data_m$JOBS_1000)
st_data_m$TOT_EMP <- as.numeric(st_data_m$TOT_EMP)


st_shape <- readOGR("./BLS_Shape4", "cb_2016_us_state_500k",verbose = FALSE)


st_shape <- spTransform(st_shape, CRS("+proj=longlat +datum=WGS84"))


st_mgmt <- merge(st_shape, st_data_m, by.x = "NAME", by.y = "STATE", duplicateGeoms = TRUE)

st_popup <- paste("<strong>State: </strong>",
                  st_mgmt$NAME,
                  "<br><strong>Management Jobs (per thousand): </strong>",
                  st_mgmt$JOBS_1000,
                  "<br><strong>Total Management Jobs: </strong>",
                  st_mgmt$TOT_EMP)
                  
pal <- colorQuantile("GnBu", NULL, n = 4)

leaflet() %>%
  addProviderTiles("CartoDB.Positron") %>%
  addPolygons(data = st_mgmt,
              fillColor = ~pal(JOBS_1000),
              color = "#BDBDC3",
              weight = 3,
              popup = ~st_popup)

Deep Dish Anyone?

Looking at our results we can see that both Massachussetts and Illinois have a substantial amount of management jobs strictly based on density (77.7 and 72.1 per thousand respectively). The state of Illinois however, contains almost twice as many management jobs within the entire state. It’s clear that Chicago has its eye on the prize and is taking steps to be as competitive as possible.
Let’s take a look at some of the metropolitan areas of Illinois to see if Chicago is the best area in the state for the new headquarters. We’ll use the same metrics as before, but in this chloropleth we only have a few areas visible. This is due to the scattered naming conventions of metro/micropolitan areas. We’ve isolated some of the largest areas by population and applied the shading based on management jobs per 1,000.

Management Jobs By US Metro/Micropolitan Statistical Area

msa_data <- read_excel("./oesm16ma/MSA_M2016_dl.xlsx")
msa_data <- msa_data %>% filter(PRIM_STATE == "IL", OCC_TITLE == "Management Occupations") %>% select (PRIM_STATE,AREA_NAME, JOBS_1000, TOT_EMP) %>% group_by(AREA_NAME)
msa_data$JOBS_1000 <- as.numeric(msa_data$JOBS_1000)
msa_data$TOT_EMP <- as.numeric(msa_data$TOT_EMP)

msa_shape <- readOGR("./BLS_Shape", "cb_2016_us_cbsa_500k",verbose = FALSE)
msa_shape <- spTransform(msa_shape, CRS("+proj=longlat +datum=WGS84"))
msa_shape <- msa_shape[c(22,374,628,707,877),]

msa_mgmt <- merge(msa_shape, msa_data, by.x = "NAME", by.y = "AREA_NAME", duplicateGeoms = TRUE)

msa_popup <- paste("<strong>MSA: </strong>",
                   msa_mgmt$NAME,
                  "<br><strong>Management Jobs (per thousand): </strong>",
                  msa_mgmt$JOBS_1000,
                  "<br><strong>Total Management Jobs: </strong>",
                  msa_mgmt$TOT_EMP)

pal <- colorQuantile("GnBu", NULL, n = 5)

leaflet() %>%
  addProviderTiles("CartoDB.Positron") %>%
  addPolygons(data = msa_mgmt,
              fillColor = ~pal(JOBS_1000),
              color = "#BDBDC3",
              weight = 4,
              popup = ~msa_popup)

Conclusions

Now looking at one variable (management jobs) is not an effective way to predict which city Amazon will choose for its headquarters. It certainly can’t hurt Chicago’s chances however, and it gave us a fun little exercise to explore using the tools we’ve learned. Some of the offhand insights learned were quite interesting as well, check out Maine’s density of management jobs (57 per thousand! Not too shabby!)