4

I am trying to identify the zip code for each set of lat/lon coordinate using a shapefile.

Lat Lon data is extracted from: https://data.cityofchicago.org/Public-Safety/Crimes-2017/d62x-nvdr (Crimes_-_2001_to_present.csv)

Shapefile: https://www2.census.gov/geo/tiger/PREVGENZ/zt/z500shp/ zt17_d00.shp (zip code definitions for the state of Illinois)

library(rgeos)
library(maptools)

ccs<-read.csv("Crimes_-_2001_to_present.csv")
zip.map <- readOGR("zt17_d00.shp")
latlon<-ccs[,c(20,21)]
str(latlon)
   'data.frame':   6411517 obs. of  2 variables:
    $ Latitude : num  42 41.7 41.9 41.8 42 ...
    $ Longitude: num  -87.7 -87.6 -87.7 -87.6 -87.7 ...
coordinates(latlon) = ~Longitude+Latitude
write.csv(cbind(latlon,over(zip.map,latlon)),"zip.match.csv")

This is the error I get:

Error in (function (classes, fdef, mtable) : unable to find an inherited method for function ‘over’ for signature ‘"SpatialPolygonsDataFrame", "data.frame"’

What am I missing? Any help is appreciated!

IsisDorus
  • 67
  • 2
  • 10
  • You are trying to create a comma delimited file from a CSV and SpatialPolygonsDataFrame, they have completely different dimensions to them. You need to combine data with that SPDF file differently. Try sharing `str(zip.map)` output and what the purpose of binding is with us. I understand what you are looking for in the very end, but how is cbind with over your method for achieving this? I might be able to help if I can see the data. – sconfluentus Aug 30 '17 at 04:50
  • I haven't tried this for your data, but you might want to look up `?inout` in the `splancs` package. It's a way of testing if a set of points falls within a polygon (shapefile), which might get you closer to what you need. Sorry this isn't the exact solution. – Rich Pauloo Aug 30 '17 at 05:01

1 Answers1

13

From the error message, it looks like your coordinates(latlon) = ~Longitude+Latitude line did not successfully transform the data frame into a Spatial object. You may wish to check that first with a class(latlon) call after the transformation.

It would also be helpful to plot out the shapefile & overlay it with the latlon layer, just to make sure that your datasets actually do overlap. If they don't, check whether they share the same projection (sp::identicalCRS).

The following is an example using dummy data, since the shapefile link in the question doesn't work.

library(rgdal)

# load Scotland shapefile, which came with the package
dsn <- system.file("vectors", package = "rgdal")[1]
shapefile <- readOGR(dsn=dsn, layer="scot_BNG")
shapefile <- spTransform(shapefile, CRS("+proj=longlat +datum=WGS84")) #change CRS

# create dummy data frame with coordinates in Scotland (I clicked randomly on Google Maps)
csvfile <- data.frame(lat = c(-4.952, -4.359, -2.425),
                      long = c(57.57, 56.59, 57.56))

# convert data frame to SpatialPoints class
coordinates(csvfile) <- ~lat+long

# make sure the two files share the same CRS
csvfile@proj4string <- shapefile@proj4string

# visual check
plot(shapefile, border = "grey")
points(csvfile, col = "red", cex = 5)
axis(1) # showing the axes helps to check whether the coordinates are what you expected
axis(2)

visual check

# if everything works out so far, the following should work
points_in_shape <- over(csvfile, shapefile)

> points_in_shape
  SP_ID          NAME ID_x COUNT   SMR  LONG  LAT     PY EXP_ AFF   X_COOR   Y_COOR ID_y
1    50 Ross-Cromarty    5    15 352.1 57.71 5.09 129271  4.3  10 220678.6 870935.6    5
2    11 Perth-Kinross   29    16 111.3 56.60 4.09 346041 14.4  10 291372.7 746260.5   29
3     3  Banff-Buchan    2    39 450.3 57.56 2.36 231337  8.7  16 385776.1 852378.2    2

> cbind(csvfile, points_in_shape["NAME"])
     lat  long          NAME
1 -4.952 57.57 Ross-Cromarty
2 -4.359 56.59 Perth-Kinross
3 -2.425 57.56  Banff-Buchan
Z.Lin
  • 28,055
  • 6
  • 54
  • 94
  • Thank you for the help Z. Lin! You were right, my coordinates() transformation was not transforming correctly to a spatial object - I had missed some nulls in the lat/lon data... Once I got that fixed and followed all the other helpful steps you laid out so clearly it worked. Thank you!!! – IsisDorus Aug 30 '17 at 22:42