Smathermather's Weblog

Remote Sensing, GIS, Ecology, and Oddball Techniques

Archive for November, 2012

Using ST_Node in PostGIS (where once I used ST_Union)

Posted by smathermather on November 30, 2012

Just discovered ST_Node, a function in PostGIS just for noding geometries.  I used to use ST_Union for this, a potentially memory intensive operation.  Now I can use ST_Node.

For usage, see the PostGIS docs (requires 2.0 and GEOS 3.3.2 for bug free use).

Back to coding.  That is all.

Posted in PostGIS | Tagged: , | 2 Comments »

What is GIS? (continued again!)

Posted by smathermather on November 27, 2012

In previous posts [1] and [2] I talk about identity crises in GIS/Geospatial world related to real economic changes in the way the geospatial sector is structured.  Comments have been many, but it seems that (my fault here) it quickly devolved into Planning GIS vs. Spatial IT (after all, I was responding to Paul Ramsey’s meme), in other words IT vs. Planning, and the cultural identities of the various parties involved.  That dialectic I find tiring.  GIS/GeoSpatial is both, largely because I’m the type of guy who likes to hack around with code, and I like to do big picture stuff.  To the extent that those are separate things, I see as artificial and generational.  I’m (almost) of the digital native/GenY/Millennial Generation, so I lay claim to the possibility of being good at both.

But let’s get back to James Fee’s supposition of fundamental shifts in the geospatial industry and what they mean to GIS professionals.  To frame this more deeply, I’ll tap Clayton M. Christensen’s article in the New York Times entitled “A Capitalist’s Dilemma: Whoever Win’s on Tuesday” (written just before the presidential election).

Without getting into Mr. Christensen’s conclusions, Christensen frames the innovations associated with economy in three categories:

  • “empowering” innovations
  • “sustaining” innovations
  • “efficiency” innovations.

Empowering innovations are things like the Model T Ford relative to previous automobiles– industry creating efficiencies that result in innovations at scale.  “Cloud” computing he also places in this category, as it allows individuals, small, and medium businesses access to technologies once available only to enterprise level investment.  Empowering innovations create new sectors, new jobs, new career opportunities.

Sustaining innovations are self competing innovations– Toyota Prius as a competing product with the Toyota Camry, in order to prevent competition from without.

Finally efficiency innovations, like Geiko Insurance, are places where refinement of existing technologies results primarily in cost ratio efficiencies of great scale.  This is where labor and other costs are reduced significantly (jobs lost, capital gained).

The distinction between empowering innovations and efficiency innovations fascinates me, as both are about driving down cost ratios– the latter to drive up capital returns, the former to do the same, but by scaling to new markets.  In other words, both are about efficiency, but with a different vision and different outcomes.  Empowering innovations seek efficiency for the purpose of scaling horizontally.

To place this back in the geospatial geoid, and reference again James Fees “Goes without saying“, if you learn to program and engage in geospatial development, then you can hang on to the rope for the ride up that the empowering innovation of the larger field geospatial development allows for.  Yes, your work will become more efficient.  Embrace that.  There’s much more work to be done in the sector, so embrace the efficiencies.  If you choose not to engage in geospatial development, then you may get squeezed by efficiency part of the equation from others, and will not benefit from the horizontal scaling.

Learn a little Python; learn a little PostGIS.  And welcome!

Posted in Planning | Tagged: | Leave a Comment »

Tesselation of Hexagons in PostGIS

Posted by smathermather on November 25, 2012

Sadly, I cannot post the code for this one, but will post the images– playing with tesselation in PostGIS today:

Posted in Analysis, Database, PostGIS, PostgreSQL, SQL | Tagged: | 3 Comments »

A little texture that needs composited

Posted by smathermather on November 12, 2012

Posted in Other | Leave a Comment »

Compositing in TileStache

Posted by smathermather on November 9, 2012

TileStache (and some adaptation of great pre-built cartography from others more artful than I) results in this:


Posted in Cartography, Mobile, Trails | Tagged: | Leave a Comment »

Multi-line Comments in SQL for PostgreSQL

Posted by smathermather on November 3, 2012

Who knew? Not I! I shouldn’t confess, but just discovered multi-line comments in PostgreSQL:

-- This is a single line comment in SQL (I knew this already...)
/* Multi-line quotes use C
syntax.  Simple as that.  Surprised I hadn't googled (well duck duck goed) this before! */

Sadly, WordPress syntax highlighting doesn’t recognize this… . I think I can muddle through without it though.

Posted in Database, SQL | Tagged: | Leave a Comment »

What is GIS? (continued)

Posted by smathermather on November 2, 2012

In the previous post, I cite Paul Ramsey, Brian Timoney, and James Fee’s various posts on changes in the geospatial sector.  Paul responded with a great post refining his position on Spatial IT vs. GIS.  Due to my argumentative academic training, I can’t leave well enough alone.  Besides, I said I’d follow up on Brian’s post, and the following does so, at least indirectly.

Paul hits the problem on the nose with the following line:

“… (A)s we know, GIS courses are just the bait in the trap, to suck naïve students into a career where” they become either a “digitization monkey” or “map monkey”

Ah, so true, so true.  In brief (ha!), Brian Timoney and James Fee’s posts point to what Paul describes as the end of typical GIS grunt work– work which will “be folded into generic IT workflows, automated, and systematized”.

(Short, but relevant aside.  My wife’s training is in Political Economy and Political Ecology.  She out ranks me with a Ph and D after her name.  She critiques economic systems and is good at it.  Whatever follows of my analysis which is good can be attributed to her.  Whatever is bad, is mine and mine alone.)

So, why so many years of grunt work before the automation?  Why now the end that Brian and Paul declare?  Why must analysts be programmers as Fee declares?  Simple: when I wrote that ESRI was starting to directly directly compete with their a) resellers and b) users, I knew that wasn’t really true.  ESRI has almost always had a) resellers and b) software operators, with a few rare users.  The software operators were valuable to ESRI in order to sell licenses.  In other words, the software was the capital, the software operators the labor, even if that labor was employed by others.  As ESRI gets squeezed, they are in no position to keep all their excess labor.  They will streamline workflows and provide hosting services to drive down cost for the end user at the expense of labor and middlemen.

Your way out as a software operator, is to (in some small or medium or big way) make the software your own, to move from operator to engineer.  Thus, to be an analyst you must be a programmer.  Don’t worry– there’s plenty of space still for good geospatial software engineers.

Finally, Paul draws the distinction between GIS/Planning as

  • high touch and interpersonal;
  • qualitative and presentational;
  • ad hoc and unpredictable.

as contrasted with those flows that can and should be automated.  This is where my tendency to lump (and time in academia) gets in my way of agreeing 100%.  Almost all processes are repeated.  Spend some time doing the high touch, ad hoc, and qualitative work, the fun planning stuff.  As you do it, look for patterns, repetition, and opportunities for automation.  Then, automate the heck out of it.  You’ll thank yourself later.

Posted in Planning | Tagged: | 1 Comment »