OpenSource Map Vector Data

s003apr · 02-08-2005, 08:51 PM

I was wondering if there is any interest in developing a database that can provide vector data for mapping and routing. I know a lot of projects have involved using a third party SDK like destinator, but with so many people looking for a good data source, why has nobody started working on a free solution? The tiger data is free and makes a good source for creating a database. It is not only of interest to me for a car PC but for my linux computers as well and my Zaurus Pocket PC.

I have started on some basic tests and it seems to me that it is very doable. I am using Gigabase 3.30 for database software along with the tiger data for my home state of Ohio. Gigabase stores its' database in a single file which is nice, and it is not a pain to set up like the big database softwares. Also, it will compile on both Linux and Windows (I have tried it with Mingw, but there is some make data for VC++).
All the roads in Ohio resulted in a 267 MB file containing roughly 1,000,000 line segments. I was careless with the size so I am confident that it shrunk to a large degree with minimal effort. I followed this with a spatial test query and was able to retrieve over 1000 records in fractions of a second (I did not officially time it). I am next interested in trying out sqlite. It is also crossplatform, keeps databases in a single file, is small, and easy to use. It does not support spatial queries but that can be worked around.

If anyone has interest in this type of project, reply and let me know.

Change · 02-09-2005, 12:08 AM

Downside to the Tiger data is that it lacks data on whether a road is one-way or not.

NiN^_^NiN · 02-09-2005, 12:20 AM

Hmm i thought tiger data showed 1 way roads but didn't show you which way it was?

I really need to find some software for me to make some maps that can be made routeable as i'd like to make a map of NZ.

Few carpc guys in NZ who have no GPS option at all unless they buy a garmin GPS to route with the handheld.

Do you guys know of software of how to actually make the maps?

RX7FDTT · 02-09-2005, 01:14 AM

The thing to do is, to get eveybody to log their GPS Data, and then start building a database of the routes, kind of like SETI@Home to piece together a map, the issue comes in maintianing it, and naming all the roads...

rando · 02-09-2005, 01:43 AM

Quote: Originally Posted by Change

Downside to the Tiger data is that it lacks data on whether a road is one-way or not.

It's also missing Z-order information to allow you to distinguish over/under passes from intersections.

AlienEclipse · 02-09-2005, 04:15 AM

Quote: Originally Posted by NiN^_^NiN

Hmm i thought tiger data showed 1 way roads but didn't show you which way it was?

I really need to find some software for me to make some maps that can be made routeable as i'd like to make a map of NZ.

Few carpc guys in NZ who have no GPS option at all unless they buy a garmin GPS to route with the handheld.

Do you guys know of software of how to actually make the maps?

I'm from NZ and having been doing a little research into GPS maps of NZ. I know there are 1 or 2 but they are proprietary for use with truck/courier tracking systems. The only NZ company that has NZ fully mapped is http://www.tumonz.co.nz/ . I'm not sure if they have vector street maps with all relevant detail.

I'm willing to do what i can to support this project, as NZ really needs something.

mpattonm · 02-09-2005, 04:47 AM

I will be glad to help in project going as long as its intention is to develop free (or better to say open source) navigation software. I can take over the GUI part.

s003apr · 02-09-2005, 06:07 AM

Quote: Originally Posted by rando

It's also missing Z-order information to allow you to distinguish over/under passes from intersections.

I believe the CFCC data can be used to figure this out. However the one-way street problem is a little more difficult.

s003apr · 02-09-2005, 09:48 AM

Quote: Originally Posted by mpattonm

I will be glad to help in project going as long as its intention is to develop free (or better to say open source) navigation software. I can take over the GUI part.

I would plan to only release software under the GPL or something very similar.

Getting others to write GUI's that interface with the database is probably the easy part. There are already a number of examples of this. My main goal is to get poeple to stop using the following solutions and use a common one:
-Tiger files (they are big and not good for query type work). In general users will have to limit themselves to small amounts of data, like one or two counties.
-Shapefiles. They are all around better than Tiger data but you are still limited to manually loading files of interest.
- Navteq or other proprietary database. Everyone has been begging for companies to distribute these databases for awhile. It is ridiculous that I can buy Rand Mcnally software which uses Navteq for under $20, but if I want to be able to make my own software, Navteq is unwilling to even sell the database. Maybe they are trying to guarantee there commercial buyers that they are not going to promote outside competition, or maybe they think that selling the database directly to end users will result in it being pirated. The best we have so far is the expensive and Window's only Destinator SDK. Either way, I have gotten tired of waiting for an affordable database solution to be provided.

The hard part is providing a good database and releasing it in stages that do not constantly break compatibility with frontends. My limited testing has at least convinced me that products like Gigabase or Sqlite will provide good backends to store and serve the data quickly and efficiently. The other hard part is developing cooperation among the many GUI programmers to decide upon how to structure this database so that it can meet their requirements and actually be used in their frontends.

god_of_cpu · 02-09-2005, 10:12 AM

Quote: Originally Posted by s003apr

- Navteq or other proprietary database. Everyone has been begging for companies to distribute these databases for awhile. It is ridiculous that I can buy Rand Mcnally software which uses Navteq for under $20, but if I want to be able to make my own software, Navteq is unwilling to even sell the database.

Navteq is perfectly willing to sell you the data, they are not a retail outlet though so you have buy in quantity. Saying that its ridculous that Navteq doesn't sell to end consumers is the same as saying its ridculous that intel won't directly sell you a single processor. Yeah I'm sure all distributors would love to sell directly to consumers, however, when you do that you take half the buisiness away from your resellers with the people who prefer cheapness over convienience. Your resellers then go out of buisiness cause half thier customers are gone so you then lose your resellers and half of your business cause most people don't want to deal with buying raw data and figuring out how to make it work.

You can't just give away map data, its expensive to make and expensive to maintain. Unless the goverment releases data good for navigation, you will always have to pay for good up to date map data.

s003apr · 02-09-2005, 12:50 PM

Quote: Originally Posted by god_of_cpu

Navteq is perfectly willing to sell you the data, they are not a retail outlet though so you have buy in quantity. Saying that its ridculous that Navteq doesn't sell to end consumers is the same as saying its ridculous that intel won't directly sell you a single processor. Yeah I'm sure all distributors would love to sell directly to consumers, however, when you do that you take half the buisiness away from your resellers with the people who prefer cheapness over convienience. Your resellers then go out of buisiness cause half thier customers are gone so you then lose your resellers and half of your business cause most people don't want to deal with buying raw data and figuring out how to make it work.

You can't just give away map data, its expensive to make and expensive to maintain. Unless the goverment releases data good for navigation, you will always have to pay for good up to date map data.

Great Points in relation to Navteq's business strategies. I eluded to some of the possible reasons nobody sells data to end users in my post. But we 100% agree on the major point. If we're waiting for someone to provide us with a cross platform and affordable map database, we will be waiting for a long time. Best solution: make our own.

I think everyone underestimates the quality of the Tiger data. It sucks from a routing usability standpoint, but the information needed to do the routing is there. It just needs to be in a more usable format. I can guarantee that the majority of Navteqs routing information comes from this data. It would be insanity to drive all around the U.S. and attempt to characterize routing information for every intersection. They do attempt to do this type of stuff for some high traffic and confusing areas like around airports, but for the most part, their data has a large number of routing problems as well.

pablosbrain · 02-09-2005, 01:04 PM

I would love to have the map data available in a more usable format... some database format that can be moved and translated to other database formats would be great... for instance MySQL for web applications. With the correct data it would be totally possible to create a flash based mapping software or other. It would open up a lot of possibilities! I tried a long time ago to look through and try to decipher the tiger map data. Its just a very confusing format. If a script can be written to convert/import it into a database that would be great!

god_of_cpu · 02-09-2005, 01:27 PM

Quote: Originally Posted by s003apr

It sucks from a routing usability standpoint, but the information needed to do the routing is there.

What are you going to do to get around the lack of directional information for one-way streets etc?

meddler · 02-09-2005, 04:30 PM

Quote: Originally Posted by god_of_cpu

What are you going to do to get around the lack of directional information for one-way streets etc?

If you can edit the database to give that type of info, why not have the users update the information. When they find something that is different to the DB send the required info to the person in charge of the database to get it updated. Then the DB info gets updated say every couple of months. It is in the users interest to give the correct info. Just a thought

s003apr · 02-09-2005, 07:43 PM

Quote: Originally Posted by pablosbrain

I would love to have the map data available in a more usable format... some database format that can be moved and translated to other database formats would be great... for instance MySQL for web applications. With the correct data it would be totally possible to create a flash based mapping software or other. It would open up a lot of possibilities! I tried a long time ago to look through and try to decipher the tiger map data. Its just a very confusing format. If a script can be written to convert/import it into a database that would be great!

I initially tried my projects with mysql and postgresql. In many ways they are overkill for this project from both a administration and features standpoint. Gigabase and sqlite both of support for limited subsets of SQL, however they do not have the heavy admin burden, they are small and crossplatform.

The tiger data is not that difficult, the documents are full of lots of fluff so you just have to narrow it down to the important info.