The "Where" Questions

In many of our other articles and resources about geographic information systems (GIS) and mapping for nonprofit organizations (NPOs), we focus more on the "NPO" than the "GIS" aspects: why should NPOs use GIS/mapping tools, how do these tools help fulfill nonprofit missions, when is it appropriate for NPOs to use these tools, what types of projects are other NPOs working on, etc.

In this article I'll be discussing some concepts squarely from the GIS/mapping side of the fence, but relax: we're going to keep it super simple and try to put in as many relevant community/NPO/NGO examples as we can :)

Every formal definition of GIS ultimately boils down to "managing information about location" -- so where something or someone is/was/will be.  Your GIS or mapping project might include information about who (a map of your organization's constituents), what (a map of local schools and parks), or when (historical community maps) -- maybe even how (a map showing streets, subway, and bus routes, for instance).  But the entire point of GIS is to present all of those other types of information in a "where context." 

One way to simplify GIS is to look at it as a way to get answers to our where questions.  In this post, I'd like to provide nonprofit-relevant examples of the types of where questions that GIS can provide.  In fancy-pants geographer-speak, these are frequently referred to as "spatial queries," but I'm going to just keep using "where questions" because seeing phrases like "spatial queries" tends to freak non-geographers out.

Time for a geometry refresher - no scary trig or equations, just some elementary concepts.  In GIS, we have three main "types" of specific location data that we can store.

• point - a single location, like a dot on a map.  When most people think of storing geographic data, they think of latitude and longitude coordinates referring to a point, like on a GPS unit.
• line - When we connect two points, we have a line.
• polygon - When we connect the endpoints of three or more lines, we have a polygon.  Polygons can be simple (a square) or incredibly complex (a map of an electoral district).

An interesting note about scale: on a map of the world, New York City looks like a single point.  However, on a New York City subway map, you can see that New York City isn't just a single point but an area (or a VERY complicated polygon).  A true point has no dimensions: it's as if we had a camera with an "infinite zoom in" feature aimed at a tiny point on the ground, zooming in forever.  For pragmatic purposes, like in our world map vs. subway map above, we depict points as dots depending on the scale of the map.  Similarly, roads are frequently depicted on maps as being lines but no road in the real world is ever an exact straight no-width line: just like our example above, as we zoom in on a road from above, we can see that it gains width and curves around.  A better example of a line in this context would be a straight-line border between neighboring states, such as Wyoming/Montana: no matter how thin you drew the border line, you could always zoom in more and make it even thinner.

Now that we've had a geometry refresher, let's look at how we can ask questions about the world around us -- our where questions -- and examine the ways that GIS tools can answer these questions.

• Where is Point A?  This is the absolute most basic "where" question and the most helpful answer is the most common: the GIS system draws a map and marks the point with a dot or star or other symbol.  After all, a picture is worth a thousand words :)  Note: if we're dealing with human input, this basic step usually involves the process of geocoding; that's where the GIS system converts a human-readable address ("123 Main Street in Happytown") into a specific geographic location suitable for marking on a map ("40.271 degrees north, 87.68 degrees west").  If you've ever typed an address into Google Maps, you've performed this type of spatial query.
• Where is the closest "X" to me?  This is essentially calculating distance - a relatively simple calculation that is used in a lot of more complex calculations in GIS applications.  For instance, let's suppose we have a website service that takes a user's address and displays the address of the nearest health clinic.  Before the system can determine the closest clinic, it must calculate the distance between Point A (the user) and Point B (a clinic) for every clinic in the area - then it can display the address of the clinic with the smallest distance from the user.
  
  
• Where is Point A with respect to X?  This is a very broad query, and it assumes we're trying to relate a location (Point A) to another location (such as a point or polygon).  GIS systems can understand several spatial concepts, such as adjacency (borders), within/not within, overlapping, or intersecting.  One of the most commonly used of these concepts is "within" or "contains."  For example, we might use a GIS system to create a list of all of the community centers within a certain congressional district: if we define the district as a polygon shape, and we have a list of points representing community centers, then it's easy for a GIS system to determine if each point is within the district and display a list or a map of those that are.  We can search in the other "direction" as well: consider the popular "enter your address to find out who your legislator" is applications.  They work by converting your address into a point and then determining which of their defined political districts -- which are just polygon shapes -- contains that point.  A site like Project Vote Smart can then easily list each of the districts - local, state, federal - containing your home and who the representative is for each district.  Taking it a step further, we can use our GIS system to analyze more and more data so that we're answering more sophisticated queries: "show all ZIP codes where the local juvenile asthma rate is above the state average AND are also in a congressional district where the senator voted No on our juvenile asthma legislation."

"But wait!" some of you are thinking.  We've defined points and lines and polygons - but it's all straight edges and jagged angles!  Where are our nice, soothing, organic circles?  What if we need to know where all the hospitals are within a 50-mile radius of a certain point as opposed to a predefined square or polygon containing that point?

Applying the concept of circles to our where questions involves the GIS concept of a "buffer," as in the phrase "buffer zone."  Think of a circle as a buffer zone around a central point - the size of the circle depends on the radius, the distance from the circle's center to its edge.  So in our example above, our list of health clinics would be sorted by whether or not they fall within a 50-mile-radius buffer zone of our searcher's address.  Buffers are useful because they can also exclude regions around points or polygons: for instance, "show me all of the Potential New Park Sites that are NOT within 1 mile of a Toxic Waste Dump."

See, that wasn't so bad, was it?  :)  Hopefully the knowledge of the common types of spatial queries that GIS applications can answer will help you plan your next nonprofit GIS project.  As always, comments can be posted below or sent privately via our contact page.