Page 4 of 5 FirstFirst 12345 LastLast
Results 31 to 40 of 42

Thread: OpenSource Map Vector Data

  1. #31
    Low Bitrate
    Join Date
    Mar 2004
    Posts
    89
    I did not realize that it was so innaccurate. They give coordinates in degrees down to 6 decimal places, that is resolution that I calculate to be roughly 0.4 feet using the formulas on this page:
    http://www.meridianworlddata.com/Dis...alculation.asp
    Maybe they have high resolution and very low accuracy, I had just guessed that this was not the case based on the number of digits they were including.

    I have had no trouble understanding the tiger data. I use the tiger line ID (TLID) in the RT1 file to find any information I need from the other files. I was using the CFCC codes to group roads into 5 seperate categories based on size, that way when you are zoomed out you can limit your query to just the big roads instead of having to query all roads and getting a bunch of results where you have to sort out the data to find just the bigger roads.

  2. #32
    Newbie
    Join Date
    Oct 2003
    Location
    Tampa, FL
    Posts
    38
    When you draw out the line segments, you will see how inaccurate the data it.

    For polygons (like water areas, parks and landmarks), I believe you need to read in RTP, RTI and some other records. I started doing it until it turned into a royal pain. I switched to shapefiles from ESRI which are 100 times easier to read. But they only have 2000 TIGER data.

  3. #33
    Low Bitrate
    Join Date
    Mar 2004
    Posts
    89
    Okay, back from vacation. I have some files in CVS although I need to update them.
    SumitB: if you or anyone else are interested in taking a look at Gigabase and how it performs you can get gigabase from:
    http://www.garret.ru/~knizhnik/gigabase.html
    In the cvs archive I have everything I am currently working with in a module called "dbtestbed" which contains some subprojects.
    tig2gigA is a script that can be used to create a database from tiger record type 1 files and record type 2. To run the script:
    >tig2gigA TigerFile.rt1 TigerFile.rt2 DatabaseName.dbs
    The other subproject you will probably be interested in is gigAq, it shows an example query.
    You probably will not want to actually use this query on the database created with tig2gigA because there are no indexes on any of the columns.
    You can add indexes and do queries with the "SubSQL" program that comes with Gigabase.

    I will add an index on the spatial column soon to tig2gigA so that you can see how it performs with indexing.

    I am currently working on the subproject called tig2gigB. This is a simple library that I used swig to create a python wrapper for. Then I can do a lot of testing and analysis quickly in python so that I can learn how to best organize the database.

  4. #34
    Constant Bitrate
    Join Date
    Mar 2004
    Location
    California
    Posts
    135

    Hey

    Hey guys, I don't know if you remember me, but I was working on the XCar project as well before we dropped the ball and let it go stagnant. I've got quite a bit of experience doing spacial development in the past using MapInfo data. One of the companies I worked for in the past I was developing traffic mapping with hopes to sign deals with satnav companies.

    Anyway, I'm interested in all things carputer, and aside from maybe releasing a product or two from a company I'm starting, I want to contribute some development to some free apps.

    I'm onboard for this project. I've done some work with Tiger data directly, and am quite aware of its deficiencies. I was considering setting up a web services server for receiving remote updates of locational information and embed the client into some sort of mapping application (i.e. Map Monkey) that would help keep the data up to date.

    What do you guys think?

    Also, what's the status of NavDB at the moment?
    Kevin
    Audi A4 Carputer, 80% setup.

  5. #35
    Low Bitrate
    Join Date
    Mar 2004
    Posts
    89

    NavDB status

    As of now I am the only developer for NavDB. Right now it is pre-alpha. I am doing different tests to analyze the tiger data and possible database backends. Right now my main concern is trying to reduce the size of the database. Once I actually figure out how I want to implement the database, then I think it will progress rapidly into something that can be used. If you are interested in helping, one thing I could use assistance on is in testing other database backends. I am working with Gigabase right now, but I am unsure how others like SQLite or something I am not even aware of, may be better.

    I also have considered how a server could be used with clients to constantly improve the data. It is one of a number of features that will hopefully be implemented in a database, just not NavDB. Requirements creep means that development time will be extended and since there are currently no solutions similar to NavDB out there, I would like for NavDB to be released in a shorter amount of time. With that in mind, my vision is that there will be a future development which implements features that NavDB does not have. It might be called NavDB II. I want NavDB to have fewer requirements that way it will be developed quickly.

  6. #36
    Constant Bitrate
    Join Date
    Mar 2004
    Location
    California
    Posts
    135
    OK, I'm in. I've had good successes with SQLite, MySQL, and PostgreSQL. I like the GIS extensions in MySQL and PostgreSQL.

    I've have some experience with BerkeleyDB, and a few other embedded databases, but most of them have bad api's.

    I'll look into other ones and reply back.
    Kevin
    Audi A4 Carputer, 80% setup.

  7. #37
    Low Bitrate
    Join Date
    Mar 2004
    Posts
    89
    I am interested in seeing how MySQL works out. Do you plan to use the 5.x developement release? I believe that it is the only one that supports spatial indeces. I am interested in seeing how it performs. I like the fact that it seems a little easier to administer than Postgres and that it runs well on Windows. Also it has an embedded component, so that if we go with it we might be able to just embed it into an app so that users have no problems with administration.

    I have also thought about the possibility that you suggested of having a main web database that recieves info from clients and is constantly updated. Postgres or MySQL would probably work better for this, so there is the possibility that in the end we support multiple databases. For instance we might find that Gigabase is the best solution for PDAs and small device type things. We could have programs that download regional maps from the web MySQL database and create databases that fit well on a PDA and future cell phones in gigabase format or our own custom r-tree format.

    BTW, what platform are you developing on? Windows or Linux?

  8. #38
    Constant Bitrate
    Join Date
    Mar 2004
    Location
    California
    Posts
    135
    Great points. GigaBase/SQLite/SomeOtherObjectEmbeddedDB would definately perform better on smaller devices (PDAs, smartphones, ancient laptops, low-end car computers). But an option for the embedded MySQL sounds interesting as well.

    I was thinking about using 5.x for the spatial indeces, it seems promising. PostGIS is nice, but I don't like the idea of being tied down to a specific implementation in the future. So, initially we should support one, but make room for the other, and then performance test (let's see which one finally owns it).

    Another idea is that we can build off of a combination of MySQL/PostgreSQL with MapServer and the custom TIGER data sets, using MapServer to actually serve up maps. The maps could be generated in MapServer (atleast until we come up with some sort of custom functionality) and stored to the database upon a first request for that tile, and subsequent requests would just pull it from our db. "Lite" users (PDAs, etc) can ask our web services for bundles of data and maps. What do you think of that?

    I'm currently developing on Windows, but I have a ton of Linux servers here that I can test on as well. How about yourself?
    Kevin
    Audi A4 Carputer, 80% setup.

  9. #39
    Low Bitrate
    Join Date
    Mar 2004
    Posts
    89
    I am on linux using kdevelop/qmake. I also have windows and a sharp zaurus which runs linux with Qtopia which I want to eventually have something for. I have used PostGIS. I started with it but really learned to dislike it because it requires PostgreSQL Source Code and therefore is not good for distributing as a binary and would be hell for end users and it requires Cygwin on windows. Do you need any help understanding how to read the Tiger Data and do you know where to find it? I have a bash script made by someone else that can be used to download whole states if you want it.

  10. #40
    Constant Bitrate
    Join Date
    Mar 2004
    Location
    California
    Posts
    135
    I have KDevelop/qmake setup on one of my Linux servers, and VS.NET 2003 on a Windows XP box here. So I'm covered on both ends. QT sounds fine for doing cross-platform (Windows, Linux, OSX), and even nicer for embedded (Qtopia aka QT/Embedded), although I don't believe Qtopia works with VS.NET with SDE (Smart Device Extension), but I can take a look.

    I own many PDAs and PDA Phones running Windows Media 2003 (and Windows Media 2003 Phone Edition) as well as some tablet pcs (I'm doing some home automation software development that should be pretty cool), but I don't have any Linux-based PDAs unfortunately. I've found the VS.NET w/SDE to be the best way to develop once and deploy on most any windows device, but I can't find a way to encompass embedded Linux devices yet in the same code base. I was hoping that Java J2ME/MIDP would have solved that for me, but not-so-surprising that Pocket PCs do not have a JVM, and the only quality JVM-like platform for them seems to be a commercial version of SuperWaba, which doesn't follow J2ME spec at all!!! Maybe you have some ideas about what we can do for the mobile platforms.

    Send me that script for the TIGER data retrieval, I had to retrieve mine "the hard way" by writing a parser for the TIGER website, and then using wget to fetch all the linked files. But I'm sure I've mislocated it by now Have you written anything to parse and condense the data into a database yet?

    Do you think we should write a semi-abstract TIGER parser, so we have the flexibility to write it to any database (i.e. both MySQL _and_ GigaBase in one swoop)? Or would it be better to just write a MySQL parser only, and then extract the data to GigaBase once it's put into MySQL?
    Kevin
    Audi A4 Carputer, 80% setup.

Page 4 of 5 FirstFirst 12345 LastLast

Similar Threads

  1. RELEASE: Map Monkey v1.1.3
    By Ninja Monkey in forum Map Monkey
    Replies: 163
    Last Post: 02-12-2009, 10:27 AM
  2. Replies: 106
    Last Post: 10-20-2008, 09:09 AM
  3. Replies: 3
    Last Post: 02-03-2005, 12:11 PM
  4. Delorme Street Atlas 05 quick review
    By cash_Money in forum GPS
    Replies: 7
    Last Post: 12-13-2004, 07:31 PM

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •