On GIS and their integration

Since I started working with spatial information, I’ve seen the term GIS used to define quite different things; ranging from mastering data input on QGIS to weird command line geeking. This variation of definition is very interesting, even though it results in misunderstandings and complicated situations at an organizational level. We’ll try to see where this can come from and maybe how to tackle it. This comes from my own personal experiences and the different working environments that I’ve observed.

Where does GIS come from?

We won’t re-write a post about GIS history here, other resources do this very well. You can have a look at ESRI’s website, this short introduction from the University of South Carolina, or GISGeography (first 3 search engine results huhu).

It’s interesting to note that GIS originate from environmental questions:

• How can I link my natural resource inventory data to spatial information?
• How can I visualize my resources?
• How do I plan my exploitation?
• etc.

Which makes sense, spatial is mostly about territory, and locating things is a very practical need, it’s also a very intuitive information. Even when we look at the origins of QGIS, which is the exact definition of a software, we see that it’s been initiated by Gary Sherman (interview), who was at the time a geologist in need of visualizing PostGIS data in a Linux environment.

At this point, GIS defines a tool, a software that helps you visualize, analyze, maybe process, spatial data. QGIS, ArcInfo/Map/GIS, MapInfo are examples of GIS in this context. You have some tables stored in files or a database, you connect your GIS to those data, and you have an information system that could manage the entirety of your company spatial information. It feels like this is still the most widespread definition of what GIS are. Even if this definition used to be accurate, the context we are in today is different, and I find it limiting. The amount of data we’re working with now is not the same as 20 years ago, and spatial data can’t be kept out of broader information systems.

The spatial exception

In spatial data there is spatial, and depending on the perspective, this can be a big deal. Coming from the applied side of data analysis, I used to think that spatial data had to be given a special treatment, as they are spatial. Obviously you can’t work with these data in the same way as you’d do with other types of normal data, right? Well they are spatial, but they are primarily data, and now that my head is most of the time in databases I’m focusing much more on that part of the expression. It feels weird to discover one day that the thing you thought special about what you are doing is basically a column in a table, and that it can be as simple as a text.

Obviously, geometries come with their very special and necessary set of tools and skills: input of spatial data, spatial indexing, spatial analysis and statistics, map creation and visualization etc. It does have its own set of challenges, and the developments of geospatial standards were necessary to respond to them. But now, the fact that it has its own protocols, catalogs, specialized management, all parallel to the world of data, makes it a type of information difficult to merge with the rest. It can even be very concrete, I’ve observed that in some companies, geospatial data units will exist out of the more global ‘data team’, using their own tools, standards and strategies, resulting in two worlds doing essentially the same job, but somehow uncoordinated. What is the point of using two different tools for transferring data between two tables?

If we considered spatial data separately, GIS could be an entire field, with its own unique and distinct rules. This definition, unlike the previous one, feels to broad and almost pretentious. Spatial data is not that special. It had to have its own standards to solve problems in its time, but with the general improvement of tools and computational capacities, the spatial field will probably gain from getting closer to the broader data community. The growth of the open source community and the speed and robustness of developed tools could benefit the spatial community.

A shift seems to be occurring though, the spatial world is finally colliding with what the rest of the domain has been doing (e.g. WFS3 following REST principles). Interoperability is becoming more and more present, and spatial standards joining broader standards will facilitate that.

OK so what?

Someone asked me recently, what is the role of a GIS specialist, and this stirred up some thoughts. Purely by experience, I can say that during recruitment processes and in day-to-day work, being specialized in GIS means a lot of things. I happened to set up Linux servers, install services, develop and schedule data pipelines, change map symbology in QGIS projects, all within the same job, with the same title: GIS specialist. To me, it doesn’t make much sense. GIS can’t be reduced to a single software like QGIS (amazing tool though). Neither can it realistically be its own independent network with specific rules.

Maybe we are forgetting a simple part, the information system part. If we take the Wikipedia definition we get:

An information system (IS) is a formal, sociotechnical, organizational system designed to collect, process, store, and distribute information.

Prefixing it with a G doesn’t change the essence of an information system. Many roles and skills are necessary to make an information system work, being geographic or not. From the definition it should make possible to collect, process, store and distribute data, geographic in our case. There need to be user interfaces developed for entering data, servers to be administered, databases to be architected, data to be analyzed, visualized, distributed etc.

GIS are information systems and should be treated as such. There is no reason to keep them out of more general policies and rules when it comes to organizational or technical decisions within a company.

What to do

I happened to be in the situation where, as a GIS specialist, I was completely out of the IT department, and I mean organizationally, physically and numerically, not even on the same network within the company. Trying to get closer to the IT seemed to be necessary in that case. Making sure to be aware of data policies, standards in data storage and processes, knowing about conventions, rules, tools etc. I tried to reach out for services responsible for data management, and it’s been very well received, and even encouraged. By doing so I could align on guidelines and finally start to take advantage from larger means.

It’s a simple question of using resources efficiently. GIS is now a completely digital field, while companies are more and more sensible to the importance of IT infrastructure, data management. It makes sense to bring spatial data into the equation. Resources will be given for the IT architecture, and GIS should be integrated.

Spatial data add something essential that other type of data don’t, but forgetting that they are firstly data is a mistake. I think a good example/metaphor of spatial integration would be how PostGIS works within PostgreSQL, it’s just an extension. Spatial data can make use of all the broader features of Postgres, and additionally they have their own set of tools for their special aspects, without being completely isolated.