Vacant land inventory

Working on a large 4 county area around Houston, TX. Wanted a single image to analyze parcel data to do a vacant land analysis.  This is a test to mosaic the 4 county individual NAIP images into a single ECW image.  I told this program to make the NoData areas transparent...not sure how this will turn out.

More to come...
The resulting imagery of the 4 county mosaic is about 4.3gigabytes compressed and would be about 120gigabytes uncompressed.  The lime green are unimproved parcels clipped out any areas within the 100 year floodplain.  This is a rough quick analysis done using only vector data from the appraisal districts and the associated attribute data.  An image classification analysis would help weed out false results from bad attribute data.

I had to go through these results and do a lot of manual removal.  State/County/School/Hospital/Cemetery parcels and for some reason airports, don't have declared improvement values in the CAD data.  This results in many false positives for "vacant land".

Filling gaps in your data

After mosaicing all the Travis county DEMs (10meter) you can see the abrupt white squiggles that represent areas that have NoData.

After running this through the Raster Calculator using the conditional formula FilledDEM= Con (isnull([traviscty]),Focalmean([traviscty],rectangle,4,4),[traviscty])  through several iterations, the result was a gapless digital elevation model for all of Travis county.

Basic Slope Analysis

This is a slope analysis. It's fairly simply to produce in ESRI ArcGIS.  Download the elevation data and process it using the Spatial Analyst tool under the Surface Analyst context menu.  This is showing a single tile DEM west of Austin near Mansfield dam.

You should make sure your data frame is in the same units as your z-values for your DEM or you'll get wonky results that may look right but will have bad slope values.  Most DEMs that I have used have been in meters.

If you want to Mosaic several DEMs to a new Raster dataset there is a tool called Mosaic to New Raster.  If the resulting raster has zeros or NoData cells (at the seams between the tiles) you need to use the Raster Calculator and this formula
FilledDEM = Con(isnull([Your DEM]),Focalmean([Your DEM], rectangle,4,4),[Your DEM])

This is a Conditional formula that looks for null values in your DEM when it finds this situation it will fill these cells with the mean value of a 4x4 rectangle of the surrounding cells.

Open Source geographic data viz

QGIS is open source GIS at it's most user friendly state.  GRASS is a more powerful open source GIS suite that's been in development since 1982, but it's lacking in the user interface department.  Luckily the two programs work together nicely so you can use QGIS as a frontend to GRASS.

Using QGIS with GRASS to prepare a single Digital Elevation Model(DEM) from the National Elevation Database(NED) downloaded from the Texas Natural Resources Information System(TNRIS).... whew there are so many acronyms.

Preparing this DEM to learn how to use nviz(open source 3d visualization tool)

(to be continued)

Texas appraisal data

A few counties in Texas are very user friendly and keep their our data online and update frequently. Tarrant County is very good about this as is Harris County. These are probably the two best examples of how all public entities should share our data with us.

This is a quick example of a 5 year appraisal value trend I threw together with data I already had prepared. It shows the percent change in appraised value by subdivision in Lovejoy ISD.