Demystifying geospatial - the basics - O'Reilly Digital Media Blog

location based services

Related link: http://www.oreilly.com/catalog/webmapping/index.html

Are you a web developer, data analyst or some other geek who wants to jump on the “location” bandwagon? Here are some of the “geospatial” basics you need to know. Want to learn more? My new book, Web Mapping Illustrated, will get you up to speed in no time.

What is geospatial anyway?
Geospatial data refers to information about the geographic location of an entity. This often involves the use of a geographic coordinate, like a latitude or longitude value. Spatial data is another commonly used term, as are: geographic data, geographic information system (GIS) data, map data, location data, coordinate data and spatial geometry data.

Applications using geospatial data perform a variety of functions. Map production is the most easily understood function of geospatial applications. Mapping programs take geospatial data and render it in a form that is viewable, usually on a computer screen or printed page. Applications can present static maps (a simple image) or dynamic maps that are customised by the person viewing the map through a desktop program or a web page.

Many people mistakenly assume that geospatial applications just produce maps, but geospatial data analysis is another primary function of geospatial applications. Some typical types of analysis include computing:
# distances between geographic locations
# the amount of area (e.g., square metres) within a certain geographic region
# what geographic features overlap other features
# the amount of overlap between features
# the number of locations within a certain distance of another
# and so on…

These may seem simplistic, but can be applied in all sorts of ways across many disciplines. The results of analysis may be shown on a map, but are often tabulated into a report to support management decisions.

The recent phenomena of location-based services promises to introduce all sorts of other features, but many will be based on a combination of maps and analysis. For example, you have a cell phone that tracks your geographic location. If you have the right software, your phone can tell you what kind of restaurants are within walking distance. While this is a novel application of geospatial technology, it is essentially doing geospatial data analysis and listing the results for you.

Why is all this so new?
Well, it’s not. There are many new hardware devices that are enabling mobile geospatial services. Many open source geospatial applications are also available, but the existence of geospatially focused hardware and software is nothing new. Global positioning system (GPS) receivers are becoming commonplace, but have been used in various industries for more than a decade. Likewise, desktop mapping and analysis tools have also been a major commercial market, primarily focused on industries such as natural resource management.

What is new, is how the latest hardware and software is being applied and who is applying it. Traditional users of mapping and analysis tools were highly trained GIS Analysts or digital mapping technicians trained to use CAD-like tools. Now, the processing capabilities of home PC’s and open source software packages have enabled an army of hobbyists, professionals, web developers, etc. to interact with geospatial data. The learning curve has come down. The costs have come down. The amount of geospatial technology saturation has increased.

How is geospatial data stored?
In a nutshell, there are two types of geospatial data in widespread use today. This is in additional to traditional tabular data that is also widely used by geospatial applications.

1) Raster Data
One type of geospatial data is called raster data or simply “a raster”. The most easily recognised form of raster data is digital satellite imagery or air photos. Elevation shading or digital elevation models are also typically represented as raster data. Any type of map feature can be represented as raster data, but there are limitations.

A raster is a regular grid made up of cells, or in the case of imagery, pixels. They have a fixed number of rows and columns. Each cell has a numeric value and has a certain geographic size (e.g. 30×30 metres in size).

Multiple overlapping rasters are used to represent images using more than one colour value (i.e. one raster for each set of red, green and blue values is combined to create a colour image). Satellite imagery also represents data in multiple “bands”. Each band is essentially a separate, spatially overlapping raster where each band holds values of certain wavelengths of light. As you can imagine, a large raster takes up more file space. A raster with smaller cells can provide more detail, but takes up more file space. The trick is finding the right balance between cell size for storage purposes and cell size for analytical or mapping purposes.

2) Vector Data
Vector data is also used in geospatial applications. If you stayed awake during trigonometry and coordinate geometry classes, you will already be familiar with some of the qualities of vector data. In its simplest sense, vectors are a way of describing a location by using a set of coordinates. Each coordinate refers to a geographic location using a system of x and y values.

This can be thought of in reference to a Cartesian plane - you know, the diagrams from school that showed an x and y-axis. You might have used them to chart declining retirement savings or increasing compound mortgage interest, but the concepts are essential to geospatial data analysis and mapping.

There are various ways of representing these geographic coordinates depending on your purpose. This is a whole area of study for another day - map projections.

Vector data takes on three forms, each progressively more complex and building on the former.
# Points - A single coordinate (x y) represents the discrete geographic location
# Lines - Multiple coordinates (x1 y1, x2 y2, x3 y4, … xn yn) strung together in a certain order. Like drawing a line from Point (x1 y1) to Point (x2 y2) and so on. These parts between each point are considered line segments. They have a length and the line can be said to have a direction based on the order of the points. Technically, a line is a single pair of coordinates connected together; whereas, a line string is multiple lines connected together.
# Polygons - When lines are strung together by more than two points, with the last point being at the same location as the first, we call this a polygon. A triangle, circle, rectangle, etc. are all polygons. The key feature of polygons is that there is a fixed area within them.

How can I start using geospatial data now?
If you are considering using geospatial data and building applications, this is a great time to get started. The historical foundation of geospatial data management and applications is opening up for others to capitalise on.

Many open source geospatial software projects exist. More are being started every year. Among these are both new and mature tools for handling data, making maps and doing geospatial analysis.

The work of the Open Geospatial Consortium (OGC) has produced many great interoperability specifications, allowing you to share data using standard formats and programming interfaces. Hundreds of organisations provide access to their data using these specifications, giving you easy access to data, at no cost.

Continuing Education
Want to know more? There are a variety of ways you can dig deeper into this subject:
# My book, Web Mapping Illustrated, discusses many aspects of geospatial data management and application development. It is due out next month (June 2005).
# Mapping Hacks is another O’Reilly book that is focused on geospatial technologies, also due out next month.

Consider attending one of these two conferences next month:
# Open Source Geospatial ‘05 - June 16-18th in Minneapolis, Minnesota, U.S.A.
# O’Reilly’s Where 2.0 Conference - June 29-30th in San Francisco, California, U.S.A.