[GRASSLIST:3810] Re: Newbie questions and problems

Tue Jul 6 02:57:35 EDT 2004

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

In case nobody has answered.
My personal advice is to throw away 5.0 and use 5.7 (far more powerful when it 
comes to database and vector).
1. you can do queries within grass - no need to split files
2. vector is generally better - conversion to and from raster is smooth, 
however, so do not worry
3. supervised classification is ok in grass
4. not sure about this
5. no need to worry about the choice between raster and vector (conversion is 
smooth)
6. I do not see your point - areas should be true
7 + 8 requires more thinking and possibly more work.
All the best.
pc

At 19:15, sabato 03 luglio 2004 you presumably wrote:
> Hi
> I'm a newbie to GRASS and linux. I have worked a bit in MapInfo (liked it
> very much) and ArcView (less impressed) before, as well as some image
> analysis in Idrisi. I'm not sure how GRASS compare.  I installed GRASS
> 5.0.3 (the version that came with Debian) and it is pretty stable and
> working nicely. I would like some advice on the capabilities of GRASS: 1. I
> have GPS point data together with behavioural data, species, numbers, dates
> and times (of animals I saw) in Excel format. What would be the best way to
> get it into GRASS? I would like to be able to do SQL queries on the Data so
> the the map will show for instance only animals of a certain sex or
> involved in certain behaviour. Should I import seperate files for each
> query or can I actually perform similar queries from within GRASS (I know
> that this is possible in MapInfo). I have already been able to get the data
> in a shape file (together with  the connected dbf). Should I import this
> one big file, or rather first do all my queries on the dbf and then import
> the resulting shape files? 2. For these kind of database queries:  should I
> import into vector format or rather into site data? (I have tried both and
> both work, but I don't know which will be better for analysis). 3. I also
> have a LANDSAT image (3 Bands: false color) that I want to use to identify
> the different habitat types (plant communities). I want to overlay this
> data (and others, like slope, aspect, fire history, distance from water
> etc.) with the point data of the animal distribution in order to make
> inferences about the habitat preferences of the animals (and possibly how
> it differs for the different age, sex or species or if different habitats
> are preferred for different behaviour. For this reason I plan to do
> supervised classification of the image. Can I do this in GRASS, or should I
> use something else? (I'm not sure that ImageMagic is the tool for this... I
> might need to get IDRISI if GRASS can't do it). 4. Is it possible to do the
> oposite of a SQL query in GRASS i.e. to select certain points on the Map
> and have it automatically selected in a database? Say I overlay my
> vector/site animal file on the new habitat map... is there a way for me to
> get a file with all the animals that occured in a specific habitat type?
> Will this file be in a format so that further statistical analysis can be
> done on it? 5. I made GPS track points of the borders of the different fire
> blocks. If I want to use the polygons between these track lines for
> analysis, what format would be best? Should I try to convert it into
> uniform raster blocks (and be able to do the same type of analysis as
> above, or even to combine the fire history and plant community maps in
> order to get a combination on which I can overlay the animal distribution)?
> Or is it possible to overlay vector polygons on raster blocks and somehow
> combine them (i.e. define areas where they overlap). 6. All my GPS data was
> converted into UTM WGS84 34H S (using Waypoint+) and imported as such. The
> area of study is about 3500ha. I used a closed transect  with a maximum
> width (not uniform everywhere) of about 450m, and about 20km long ( closing
> an area about 7,5km from W to E and about 4 km from N to S). I would like
> to get a density value for the animals in the different habitat types (as
> defined by my ovelays+ different buffer zones around the transect). I know
> that UTM is not area-true (doesn't conserve area). Would it make enough
> difference at this scale to distort the density (area) data? Would I need
> to reproject in order to do this kind of analysis? 7. From another source,
> the image (.img file) and some other vector maps (in ARC-View shape files)
> (contours, roads, streams etc.) were in Lamberts Equal Area Azimuthal,
> WGS84, with a nonstandard Datum 5N 20E. I imported into GRASS: they fitted
> beatifully. I also imported the image into UTM and used known ground points
> on the (GPS) vector maps to rectify it there. So then I tried to
> reprojecting both the raster and vector maps from the Lamberts projection
> to UTM. Both end up  (about 10 km) South of the GPS data (they still fit on
> each other though)! I think this is a bug in the reproject modules, because
> in the end I had to use Global Mapper (in Windows) to get the Lamberts AEA
> to fit correctly with the other UTM data. 8. I would like to get a DEM from
> the contour vector data. However, the shp file gives an error when imported
> (duplicate tags). When I try to convert it to raster, I can actually see
> that it must have come from different original map put together, because
> even though I can't see any places where there are gaps in the lines, they
> form squares of different colours. Is there a way to clean up the map
> on-screen? The same would be needed to convert my waypoint tracks into
> vector polygons... to make sure that lines properly connect, to delete
> lines that are wrong (e.g. to get rid of slivers where the lines coincide
> with  other lines), to tell the GIS somehow that the lines actually
> encompass polygons that can have different values specified. Can I do this
> in GRASS? I hope these questions are not too many at one time... I will
> seperate my questions in future if that is the prefered way. Thanks
> Chavoux Luyt

- -- 
Paolo Cavallini
cavallini at faunalia.it            www.faunalia.it
Piazza Garibaldi 5 - 56025 Pontedera (PI), Italy   Tel: (+39)348-3801953
GPG key @: www.faunalia.it/Public_key_Paolo.asc
Only free software: www.gnu.org / www.linux.org
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)

iD8DBQFA6k1f/NedwLUzIr4RAlXfAJ9MUVsrNdAgog5J1LjgmHnYCF6uDACgmSsu
7gtfNTxsYBKGaqv3K0CbSSE=
=cYBe
-----END PGP SIGNATURE-----