Does the layering construct of GIS stand in the way of more holistic approaches?

by Matt Ball on March 11, 2011

Perspectives Header

A recent conversation with a marketing professional at a geospatial data provider sparked some thought about the construct of layers within GIS, and how layers have their limits. That conversation revolved around vendors and organizations that are so focused on the individual layers that they produce (whether its imagery, vector data, elevation, 3d buildings, roads, etc.) that they don’t see the forest for the trees. The creators see their data type as the essential element, often to the detriment of their business, rather than as a piece that needs to mesh more readily into a system where the combination of layers are the key to better understanding.

The layered data approach is central to GIS, but does that construct also cause problems in seeing the holistic whole? The layers of GIS provide the means to present information that reveal relationships and patterns, but layers also breed ownership issues that cause problems for data integration and collaborative approaches. While the layers in GIS certainly won’t go away, it’s constructive to consider these issues while aiming for a new degree of transdisciplinary collaboration.

Complexity in Combining

There is a great deal of opinion that goes into every map and mapped location, with varying levels of accuracy and authority. Given the variability of map data, it gets very complicated to bring different spatial data sets together, where the layers are all not equal and require a level of transformation and normalization before being merged.

Digital data integration is a long-standing issue that is not unique to geospatial data alone. The constant tweaking to data and data models creates issues when the changes get disconnected from systems and users. Progress on data and exchange standards have aided this issue. The creation of local data clearinghouses has also helped by forming communities of shared ownership. Yet, conflation and deconfliction of like data sets of the same location are mind-numbing processes full of compromises, and more must be done to address this.

Shared Missions

Perhaps the best example of layer ownership issues lies with the local, regional and federal government disconnect. Investment in geospatial data creation has not been vertical in terms of data type, but horizontal in terms of mission. The mission approach, where the particular flavor of a data type is deemed superior for a specific purpose, often leads to repeated data creation even within the same organization for such shared data layers as roads and data sources such as imagery. Layer ownership leads itself naturally to stovepipes and isolated development efforts. Ongoing efforts to rally around data for the nation initiatives have been promising, but slow to take hold.

When it comes to issues of great societal importance, particularly around events that are unfolding, individual missions melt away, and it’s all about collaboration to get the job done. The rising importance of geospatial tools for crisis management fuels the need for shared understanding for quick and effective action. The trick is to morph our approaches and systems to be able respond in a crisis without rebuilding them on the fly in the heat of the battle. Positive steps toward more integrated systems have resulted from crisis operations, where layers quickly come together from disparate sources, but there’s still a ways to go before the full promise of a geospatially-coordinated disaster response is reached. The integration that is honed in crisis response will benefit ongoing collaborative approaches for long-range planning and crisis avoidance.

Projects or Systems

Ongoing updates and collaborative improvement of data layer quality and currency are the ultimate goal.  The current inclination is toward packaged data to create services and solutions of a project-based nature. While this approach is important for engaging newcomers to technology with simplified interfaces, if there’s a disconnect between the layered data components of these solutions and the systems where the data originated, then the benefits of the system are lost.

A project-based approach addresses an issue in time, but the work often gets shelved upon completion event though the underlying issue is ongoing. The advantage of a system-based approach is that the overall system improves over time. As new issues are addressed, the ongoing and cumulative analysis yields greater understanding of a location. Only through continued development of data and queries on shared systems does the intelligence about a location improve.

As geospatial technology continues to mature, and more geospatial data becomes available, normalizing our data layers and providing improved integration methods will go a long way toward an improved understanding of our world. No map can be comprehensive and up-to-date, although we’re certainly getting closer. With more and more data creators, the real advances in our quest for real-time map updates will come with greater collaboration.



Congressional Bill Could Reduce Fragmented Government Data Activities, NSGIC News, Feb. 23, 2011

What Makes Location Data Messy, Chris Hutchins, O’Reilly Radar, March 9, 2011


{ 3 comments… read them below or add one }

Atanas Entchev March 11, 2011 at 5:26 pm

Does the layering construct of GIS stand in the way of more holistic approaches? Of course, by its very nature.

In order to understand them, we analyze things by breaking them apart (e.g., layers). This is the exact opposite of synthesizing, the holistic approach to understanding.

Matt Ball March 13, 2011 at 6:29 pm

Thanks for the comment. I agree that the question is simple. What I hope to have addressed are the implications of layer ownership thinking, and the need for a continued systems approach.

Atanas Entchev March 20, 2011 at 10:25 am

You did address all that, and I am with you 100%. I was just entertaining a tangential thought about the differences between analyzing and synthesizing. We GISers are married to analysis, of course, so this is just an academic exercise.

Leave a Comment


Previous post:

Next post: