[OSGeo-Discuss] Data format frustration - too general!

[Homan, Thomas]

Good Morning berryman,

Google 'gis traffic analysis' for some possible answers

Your message really has 2 parts as I see it so I'll give a shot towards
each.

First:

You might want to start by rereading the manuals for your GIS software
package in use and other GIS references specifically areas on
projections. Your 'unitless locations' statement suggests to me a big
hole is your knowledge. It's only unitless because you have not defined
the units in the software or do not know the units of the source data
that you are working from. GIS software packages typically support many
different projections and once a dataset's projection is defined
reprojection can occur with ease. Everything has some level of error and
a projection is a way of managing that error. Just as an example, a
quick check of my ESRI ArcGIS installation lists 4000+ different
predefined projections. I don't know how many are defined in OSGeo.
Then, because it might be relevant to your project, you have surveyor
defined coordinate systems that you might come across working with
traffic alignments. These can be problematic because they might not be
able to be converted easily.  

I'm curious how you are going about doing the encoding. Are you using
road construction plans or aerial imagery for your base or something
else? Without knowing the established projection of your source data and
how different data sets interrelate then moving forward will be
impossible because you cannot correlate something like vehicle stopping
distance to your generated data.

Living in the US and working for a local government, I do 99% of my work
in either a State Plane or UTM projection. It's much easier to do
length/area calcs under these conditions than with Lat/Long but I can
export to LL with a few clicks of the mouse provided the software knows
my starting projection.


Second:

For my personal opinion, any piece of software can be considered a GIS
if it allows the spatial relevance of 2 or more objects/entities to be
evaluated. This covers a map of land ownership, your traffic analysis
and the screen my dentist brings up that shows graphically where all my
fillings are.

That being said, are you attempting to utilize 'GIS software' in place
of a commercial analysis package due to cost? If so then you are
definitely taking the 'hard road' because decisions such as database
format and other analytical constants will have to be made by you and
the engineer certifying the results not to mention so custom software
development will possibly have to occur. It is very common to have one
database for the spatial features and another for the analytical
features with a field or 2 that link them together.

Looking at the larger picture, 10% of the project will be the line work.
90% will be the data/intelligence applied to the line work.

Good Luck

Tom

-----Original Message-----
From: discuss-bounces at lists.osgeo.org
[mailto:discuss-bounces at lists.osgeo.org] On Behalf Of berryman
Sent: Tuesday, August 03, 2010 5:08 PM
To: discuss at lists.osgeo.org
Subject: [OSGeo-Discuss] Data format frustration - too general!


Hello,

I am hope to use GIS for a traffic analysis program that I am involved
with. 
After having looked at a couple weeks worth of GIS resources, I still
feel
like I haven't found what I'm looking for and it's time to get some
human
input.

The main source of my frustration is that all of the standard data
formats
seem to be too general for my purpose.  Common formats provide the
ability
to encode points, lines, polylines, and polygons (where the locations
are
unitless).  But I need the ability to encode the geographic location
(e.g.
lat/long, NAD83) of high-level constructs.  For instance, I'm not just
interested in the location of streets, I'm interested in the location of
each lane.  I also need to know how street topology and traffic
direction so
that I can make path planning determinations.  Finally, it would even be
ideal if I could encode the location of traffic signals, stop signs,
bus-only lanes, etc.

I know that these types of encoding must exist somewhere , because my
Garmin
GPS understands the rules of the streets and can anticipate things
coming
up.  For instance, it might tell me "stay in the right lane."  Are there
standard data formats for encoding this information?  Or is the ability
to
encode this a non-public extension of some more generic standard?  If
so,
then how do I go about extending some other standard for my application?

Thanks all!
-- 
View this message in context:
http://osgeo-org.1803224.n2.nabble.com/Data-format-frustration-too-gener
al-tp5370620p5370620.html
Sent from the OSGeo Discuss mailing list archive at Nabble.com.
_______________________________________________
Discuss mailing list
Discuss at lists.osgeo.org
http://lists.osgeo.org/mailman/listinfo/discuss