How do I combine XY coordinates and LATLONG data?

[young_spruce]

Hi,

I want to show markers on a map using the PROC GMAP statement (following this example). However, I think my map data uses a different type of data compared to my marker data.

How can I manipulate my data so that I can show markers on my map? I have tried the code below unsuccesfully. 

Thanks in advance for any help.

The code

proc gproject data=marker_data out=marker_data eastlong latlong degrees;
id pc4nr;
run;

My Netherlands map data uses the following XY coordinates

X Y
122146.69994 487964.19995
122192.39992 487945.20011
122265.2 487932.00006
122301.70003 487966.79998
122515.99996 487935.89999
122534.30007 487892.70001

My marker_data  uses LATLONG codes

lat long
52.1392 5.0417
52.1392 5.0417
52.025 4.8681
52.025 4.8681
52.1392 5.0417

My marker_data  after GJPROJECT 

X Y
-0.005838858 -0.005235953
-0.005838858 -0.005235953
-0.007718145 -0.007212797
-0.007718145 -0.007212797
-0.005838858 -0.005235953

[ChrisNZ]

My Netherlands map data uses the following XY coordinates

X Y
122146.69994 487964.19995

What do these numbers represent? In order to translate, you need a translation rule.

Do you have a couple of points in both coordinate systems?

[young_spruce]

Hi Chris,

Thanks for the quick reply.

Those numbers represent some type of coordinates per Dutch zipcode (4 digits). The map data is a set we had lying around at the office.

For zipcode 1011 and 1012, my map data contains:

X	Y	PC4NR
122146.69994	487964.19995	1011
122192.39992	487945.20011	1011
122265.2	487932.00006	1011
122301.70003	487966.79998	1011
122515.99996	487935.89999	1011
122534.30007	487892.70001	1011
122604.19995	487881.30011	1011
122623.79997	487682.30003	1011
122691.20003	487652.59997	1011
122762.20003	487253.29992	1011
122673.80005	487088.69993	1011
122249.49991	486641.70001	1011
121923.30006	486531.90005	1011
121837.09992	486639.69995	1011
121722.79996	486711.80009	1011
121524.39992	486751.09991	1011
121899.8999	487422.09993	1011
121940.80003	487820.99993	1011
122079.60006	487836.99992	1011
122146.69994	487964.19995	1011
121409.6	488033.1001	1012
121608.89998	488308.29993	1012
121624.40009	488294.20004	1012
121759.2	488197.30003	1012
122114.40001	487977.59992	1012
122146.69994	487964.19995	1012
122079.60006	487836.99992	1012
121940.80003	487820.99993	1012
121899.8999	487422.09993	1012
121524.39992	486751.09991	1012
121313	486706.00007	1012
121123.29994	486701.70009	1012
121010.60007	486854.39992	1012
120995.59993	486931.2001	1012
121098.4001	487516.5001	1012
121409.6	488033.1001	1012

My zipcode coordinate data set contains the following data for that same zipcode:

Longitude	Latitude	Postal Code
49.027	523.666	1011
49.027	523.666	1012

[ChrisNZ]

1. Both zip codes at the same location?

2. So many X Y to one LAT LON?

I'm getting lost here. Can you give more context?

You might need to use the centroids (which can be approximated using the averages for now) then.

1011:

122231.7799825

487475.77499

1012:

121566.149998125

487584.3000025

3. Why not use the SAS/GRAPH NL map?

[young_spruce]

Answers:

1. They are the same zipcode so they should represent the same location.

2. I'm unsure why there are so many XY values per zipcode. The data set was already present, but the person who created that dataset is no longer working here. I found the zipcode LATLON data online. It did not contain any further documentation.

3. I could use the SAS MAP if there's a way to create zipcode boundaries. The NL map only contains province boundaries. The data I have is zipcode based. I need to create that countains specific boundaries (administrative regions for Dutch longterm healthcare based on zipcodes). The map data we have does contain does those boundaries.

Questions:

A. In your second answer you mention using centroids. Would it be an option to use my current map data to create centroids per zipcode, geocode those centroids to my zipcode data set and then plot those on my map?

[ChrisNZ]

How are zip code 1011 and zip code 1012 the same?

This website disagrees http://www.geonames.org/postalcode-search.html?q=1011&country=NL

What happens when you run

proc gmap data=NLMAP(keep=X Y PC4NR) map=NLMAP;
  id PC4NR;
  choro X ;
run;
quit;

[young_spruce]

Hi Chris,

I'm sorry for the misunderstanding. Zipcode 1011 and 1012 are not the same. I thought you meant the same for both my data sets. I included two observations, to make it more clear. 

Last night (I could not let this rest) I had a breakthrough. It turns out my map data uses a Dutch system, known as the National Triangular System (Rijksdriehoekstelsel or EPSG:28992). It uses meters from a random zero point and allows all of Western Europe to have positive values so that there's no misunderstanding about left or right. The Dutch Navi provided a program to translate that system to latitude and longitude data, so I'll try to run that. 

I think this solves my problem for now. I like the boundaries of the SAS map more than I like those of my zip code data (water also has a zipcode), but that is a different question. 

Thanks for all your time and effort in solving this problem. 

PS running your code gave me the attached result.

[GraphGuy · Accepted solution]

Looks like the X & Y in your map dataset is already projected, or in a different coordinate system.

You'll need to find out what projection/coordinate-system was used for the map, and then either "unproject" the map X/Y's to get them into lat/long coordinates (same as your marker data), or project your marker data the same way. (Modern versions of Proc Gproject have a from= and to= option that can help in converting). I usually get both the map and marker data into lat/long, and then project them both, using a projection before plotting them together.

When you tried projecting your marker data, it just used SAS' "default projection" ... which is not the same projection that was used to project your map.