## Tutorials > Geocoding > Coordinate Systems

**The Basics: What is Being Geocoded and What Type of Coordinates Suffice?**

If we think of geocoding as linking a digital object (whether a record in a table, a marked-up text, a digital image, etc.) to a spatial feature that can be depicted on a map such as a point, a line, a polygon, or a raster cell, our aim will be first to consider what type of feature will ultimately come to represent our digital object, and second to identify what type of geographic coordinate we need to find.

On the first issue, in most cases our answer will resolve to a single type of feature: a point. Certainly, gathered data can be linked to lines and polygons in already-geolocated GIS files. For example, we might link social data to a polygon in the form of a census district, producing a choropleth map. In this case, the geocoding consists of a table join (using a common field) between our table of social data and our table of census districts.

But features other than points—a line, polygon, or raster cell—will often be identified for geocoding purposes by a single point in the form of a centroid at the geographic center of that feature. Here, for example, would be the centroids of a polygon, a line, and a raster cell:

If we are often geocoding by linking our data point or digital object to a point in space, what coordinates do we need to locate that point? As you will quickly discover, in most cases it is a latitude-longitude point. Thus, for instance, in the most common case, if you have a table of street addresses, geocoding consists of determining lat-long coordinates for the centroid of the building structure located at each address.

That may be sufficient information to get your points on a map, but it pays to know a little about lat-long coordinates.

Even if we confine our geocoding to obtaining lat-long coordinates, not all lat-long coordinates are the same, nor are they always formatted in the same way. Latitude and longitude systems of locating a point on earth belong to what are known in GIS as geographic coordinate systems. These are systems built around modeling the earth as either an ellipsoid or a spheroid, and using the center point and an axis or radius respectively, plus the angle between that axis/radius and two baselines (the equator and the prime meridium), of that earth-as-ellipsoid in order to create a unique coordinate pair for each point on that ellipsoid or spheroid. Here is a nice visualization of this from Australia’s ICSM:

That may be sufficient information to get your points on a map, but it pays to know a little about lat-long coordinates.

**Coordinate Systems**Even if we confine our geocoding to obtaining lat-long coordinates, not all lat-long coordinates are the same, nor are they always formatted in the same way. Latitude and longitude systems of locating a point on earth belong to what are known in GIS as geographic coordinate systems. These are systems built around modeling the earth as either an ellipsoid or a spheroid, and using the center point and an axis or radius respectively, plus the angle between that axis/radius and two baselines (the equator and the prime meridium), of that earth-as-ellipsoid in order to create a unique coordinate pair for each point on that ellipsoid or spheroid. Here is a nice visualization of this from Australia’s ICSM:

So latitude and longitude, in summary, are the two angle measurements from the equator and prime meridian to the axis/radius line from the center of the spheroid/ellipsoid to the wanted point.

The selected spheroid/ellipsoid, which must approximate the (non-spherical) earth, and its accompanying baselines are known as the

*datum*. But since the earth is not a perfect sphere or ellipse, one’s selected spheroid/ellipsoid matters quite a bit, as does accurate measurement of the size of the earth so one knows the length of the radius/axis. Over the years, as more accurate measurements of the size of the earth have improved, the datum deployed for lat-long have changed. Moreover, there are different spheroids/ellipsoids selected in various datum systems. But the most common one is WGS84 (World Geodetic System, 1984 revision).

WGS84 works very well for lat-long coordinates that need to be globally unifying, since the spheroid selected is meant to work fairly well for much of the globe, making it a good unifying system. GPS systems, for example, translate their found location to WGS84 coordinates. Google Maps and Google Earth also provide lat-long coordinates for this system, and so do many other geocoders—hence the deep dive here into coordinate systems for the moment.

**A Few More Things on Coordinate Systems**

In addition to understanding the basics of geographic coordinate systems and WGS84 lat-long, it is good to be aware of variations on the themes above.

*Lat-Long Formats*

You may encounter lat-long listed in decimal degrees (DD), as with Google Maps: 40.730781, -73.997586, the coordinates for Washington Square Park. Note the negative, as lat-long systems also divide the world into Cartesian coordinate planes so that points in the the northern hemisphere have a positive latitude, while points to the west of the prime meridian have a negative longitude. But lat-long can also be expressed in degrees-minutes-seconds, or DMS. You may have to convert one form of unit to the other. Here is a quick converter, though you’ll need a batch converter if you need to convert many records without individually typing them in.

*Projected Coordinate Systems*

Geographic coordinate systems, with their reliance on angles and preservation of a 3D surface can make it difficult to work with distances and Cartesian space. Besides, once one is working with a small section of the globe, it is better to treat the surface as a two-dimensional plane, considerably easing visualization (i.e. on your flat computer screen!) and measurement of distance.

This is where projections come in: points on the earth’s surface are translated onto a tangent plane; that is, they are “projected” as if a light bulb has been placed inside the globe and the resultant projection of the surface of the earth onto a screen becomes the planar surface representing that surface. The flat, tangent plane can take the form of a cylinder or cone wrapped around the earth so that it is tangent at certain points on the sphere and then unwrapped and flattened to form the planar surface. Latitude and longitude lines are imagined as projected onto the flat surface, resulting in a grid that enables location and distance. Here (via UC-Santa Barbara geography department and Rice University) is what this looks like:

As with geographic coordinate systems, there are a variety of projected coordinate systems in use, most developed to suit particular locations on the globe (i.e. a U.S. state, or a country), or to work for the planet as a whole. Among the most common are the system for U.S. states called State Plane (in which each state gets its own projection), Mercator (based on a cylindrical projection) and an application of Mercator, Universal Transverse Mercator, that splits up the projected surface into 60 zones. There is a British National Grid used for the UK, and an Irish National Grid used for Ireland. These systems have a zero point (equivalent to the x,y value of 0,0 on a Cartesian grid) and an “x” and “y” axis known as an easting and a northing, respectively. The result is a system where a location on a map can be given a grid systems’ x,y value and thus a unique locational coordinate.

Here are some helpful resources to understanding the information above in greater depth:

- Overview of Coordinate Systems, both projected and geographic
- WGS84 specs on Wikipedia
- Explanation of Irish Grid System (but with helpful overview of all of the above)
- Listing of Coordinate Systems, both projected and geographic (see especially the list of local projected systems such as State Plane, Irish Grid, etc.)