Hack Night on July 5 - TPS/MTTA partnership

[James Wagner]

Hey Code for Tulsa. I don't know how many of you I've met (I'll post in introductions too).

I have a fun project to work on Tuesday night. It involves using Google Maps trip planner to plan a trip for all TPS high school students for this upcoming school year. I've created an outline for the project (attached). Would anyone want to work this project with me? I have some officials from TPS IT and Tulsa Transit committed to come. This could be a big impact project. I have more details to share on Tuesday night.

[Paul Johnson]

Sadly I work nights Tue-Sat, but I'm curious, is there a reason why it's using Google Maps instead of OpenStreetMap?  For one, there's more tools that can work with a more flexible dataset, and a lot of the routing engines available to work with that data are more mature and can assume modes such as truck or bus.

As an aside, getting a feedback loop in place with incog about any big projects that are going to require long-term closure is useful for me, I can update OSM.

On Tue, Jun 28, 2016 at 4:45 PM, James Wagner <jwa...@incog.org> wrote:

Hey Code for Tulsa. I don't know how many of you I've met (I'll post in introductions too).

I have a fun project to work on Tuesday night. It involves using Google Maps trip planner to plan a trip for all TPS high school students for this upcoming school year. I've created an outline for the project (attached). Would anyone want to work this project with me? I have some officials from TPS IT and Tulsa Transit committed to come. This could be a big impact project. I have more details to share on Tuesday night.

--
You received this message because you are subscribed to the Google Groups "Code for Tulsa" group.
To unsubscribe from this group and stop receiving emails from it, send an email to code-for-tuls...@googlegroups.com.
To post to this group, send email to code-fo...@googlegroups.com.
Visit this group at https://groups.google.com/group/code-for-tulsa.
For more options, visit https://groups.google.com/d/optout.

[Phillip Burger]

Looks very interesting! I do not have time right now, but would love to stay informed and maybe provide ideas or little pieces of input as you move on the project.

-Phillip

[Tom White]

I use pgrouting and open street maps data at work for fleet management. Getting the data from OSM into a usable format in postgres is a lot of work, but in general it works very well and the dataset is getting cleaner all the time. In fact, this last time I didn't have to clean up the dataset at all, and that was in 2015. But we are talking about a lot of data, meaning a little bit of horsepower would be required to use the data. A $5 VPS from digital ocean most likely wouldn't cut it.

Wish I could have made it to the meeting last night.

--

[John Whitlock]

Tom, there was no meeting last night. The meeting is next Tuesday, July 5th.

[Paul Johnson]

On Wed, Jun 29, 2016 at 8:32 AM, Tom White <tom....@gmail.com> wrote:

I use pgrouting and open street maps data at work for fleet management. Getting the data from OSM into a usable format in postgres is a lot of work, but in general it works very well and the dataset is getting cleaner all the time. In fact, this last time I didn't have to clean up the dataset at all, and that was in 2015.

I really, really tried, and my boyfriend can attest to my frustration with bad maps, to the point where I'd be taking notes on work expeditions and editing furiously for hours on end after work to actually have something accurate and up to date for work.  Pretty sure a similar effort I spearheaded with MESA both got me a free ride to next year's Wild Nights in volunteer hours and resulting in OSM having about the most complete set of data out of the big providers for Wilburton, Robbers Cave and Quinton.  Might not be the most complete in terms of addressing, but if you're also on okgis, you've probably heard my frustration about getting lot centroids with addresses.  But I've tried to make it not suck as much as possible for people actually using the data.

Sucks I don't get paid to be a cartographer.

 

But we are talking about a lot of data, meaning a little bit of horsepower would be required to use the data. A $5 VPS from digital ocean most likely wouldn't cut it.

What kind of specs are we talking?  I maaaay be able to use my connections to wrangling something like the entire planet.osm, just a chunk roughly the bbox of INCOG.  Granted, my idea of what it takes is relatively skewed by working with much smaller chunks in a java-based editor on my laptop...

 

Wish I could have made it to the meeting last night.

Me too (though, in general, just not last night, because that would have been reeealy lonely), but it's just about midnight relative to the clock I have to operate on right now... 

[Tom White]

Ok, I guess I'm glad I didn't make it last night then.

It's been a year and a half since I updated my routing data, so I'll go ahead and do that and see what size it is.

--

[James Wagner]

I have no problem with using OSM. I think all TPS is interested in is a simple step by step route for students to access the bus route AND the time they would need to leave to walk to the bus stop in time to catch it. The means to the end is not as important. The key is creating something like I illustrated in the project brief that shows a map and timed step-by-step directions on how to catch the bus.

[Luke Crouch]

FWIW, Code for Tulsa has lots of AWS credit we can use.

--

[Tom White]

An estimate of the amount of space needed for postgis storage plus indexes is 15GB.  This is based on a subset I'm using (everything but small streets) for large truck routing that takes up about 2GB in the database. It is able to route between any two points in the US in about 6 seconds. This is a single virtual CPU with 8GB RAM.

In contrast, the google routing API allows up to 150K requests per day for free if you put in a credit card, for publicly available websites.