The future of data is geographically aware

OK, but why does a GIS user need to mess with projections?

Because if you give them the right commands, GIS programs are capable of merging digital maps made with different projections. This is just one of the ways in which they can help you to discover hidden connections between different GIS data.

What is this GIS data, anyway?

It can be digital versions of what you'd find on normal maps, generic geotagged data or something called 'shapefiles'. The first category includes things like rivers, buildings, park boundaries and so on, plus all their associated properties. Other data… well, by now you know that this could be anything from scanned historical maps to use as a background to average incomes sorted by postcode.

And what are these shapefiles?

A de facto standard format, consisting of several types of files, describing geographical shapes and points such as roads or lakes, together with their properties (name, forest coverage etc).

This sounds like information overload to me. How do I possibly make sense of a map that includes all this data?

You're not supposed to – you can group the information in layers. This way, when browsing a map you can alter the colour, presence and order of the layers to show only what you're interested in.

Is there any free software available for GIS?

Indeed there is! The FOSS heavyweight in this space is the Geographical Resources Analysis Support System. Grass is a bundle of several command line programs and a bare-bones GUI. It creates maps and geographical models of human and natural activities, enabling users to find spatial relationships between data.

You mentioned that normal relational databases aren't really well equipped for GIS work. Isn't there an open source solution?

Yes: PostGIS, which extends PostgreSQL with custom GIS primitives and functions, enabling it to be used as a back-end database for programs such as Grass.

Isn't there anything else apart from Grass for getting started with GIS under Linux?I'd like to choose – that's why I'm a Linux user!

Sure. The most popular choice here is Quantum GIS, shortened to QGIS. QGIS can display, edit or print maps, work as a graphical front-end to Grass or PostGIS databases and measure distances or path lengths. It has a lot of plugins listed at http://wiki.qgis.org/qgiswiki/ PluginRepositories.

Can I add GIS to my website?

Yes. MapServer offers interactive maps that you can zoom, pan around or customise by selecting layers.

Thanks! Now I can do everything Google Maps does for myself!

Feature-wise this may be true, but unfortunately you're forgetting something crucial. GIS software is pretty much useless without lots of data: you either have to load an existing GIS database or create it yourself from scratch.

No big deal – lots of government agencies already use GIS. I'll just ask them for a copy of their databases. It's all public data, right?

Theoretically, yes. In practice, you may be in for an unpleasant surprise. For example, Ordnance Survey gives out its data only according to its online pricing and licensing policy. Other public offices simply refuse to share their information.

No fair. Why does this happen?

Sometimes it's just bureaucracy for bureaucracy's sake. But frequently the answer is money. Building good, large GIS databases can cost a lot. More often than not, keeping that data locked or licensing it at the highest possible price is the only way for some public agencies to make it through next year's budget.

Wait a second. Can't we the people make our maps ourselves, working together through the net?

Indeed we can – have a look at OpenStreetMap, a Wikipedia-style map of the whole world. While it doesn't solve the root problem, OpenStreetMap is free, editable in a collaborative way by any volunteer and you don't even need a GPS device to contribute. What are you waiting for? Sign up and start mapping!

First published in Linux Format, Issue 112

TOPICS