Network Analyst - best route shapefile

1541
9
08-03-2016 03:52 PM
Highlighted
New Contributor


Hello,

I have a large dataset (15,000+ ) of Origin-Destination pairs. For each pair, I need to solve for the best route and determine whether the best route would include using a road segment with certain characteristics (e.g. the segment is serviced by a bus route). I have a separate shapefile for the bus route. How can I perform this analysis efficiently?

Thank you very much in advance.

Reply
0 Kudos
9 Replies
Highlighted
MVP Esteemed Contributor

one bus route is not going to help you assess whether it best serves the OD pairs.  Underlying this is your assumption that the road segment would even be a possible candidate for inclusion in the route.  Consider and origin and destination in the east end... but the bus route is in the west end of town... have you excluded those from your candidate pairs first?

Reply
0 Kudos
Highlighted
New Contributor

Hi Dan,

Thanks for your prompt response. I have 30+ bus routes in the separate

shapefile (separate from the network shapefile).

Cheers!

From: Dan Patterson

Sent: Wednesday, August 03, 2016 4:24 PM

Subject: Re: - Network Analyst - best route shapefile

GeoNet <https://community.esri.com/?et=watches.email.thread>

Network Analyst - best route shapefile

reply from Dan Patterson

<https://community.esri.com/people/Dan_Patterson?et=watches.email.thread>

in *Community

Help* - View the full discussion

<https://community.esri.com/message/625841?et=watches.email.thread#comment-625841>

Reply
0 Kudos
Highlighted
MVP Esteemed Contributor

so now your problem is more complicated, you still have to parse the origin destination pairs against possible bus routes otherwise your calculations will contain improbable options.  It is still unclear as to your intent since bus routing is an often well crafted study and the input parameters and requirements may have less to do with origins and destinations but more on territory inclusion and road type traversal.  If you could clarify a bit more as to what you are trying to accomplish, others may weigh in with their experience in a practical context

Reply
0 Kudos
Highlighted
New Contributor

Sorry about the confusion. I am not interested in routing the bus. The bus

route is a given – I am interested if some of the trips (best route)

overlaps with a predefined bus network.

The goal is to assess how many trips (or parts thereof) could have been

made using bus instead of driving.

I hope that clarifies the question.

Thanks again for your prompt response.

From: Dan Patterson

Sent: Wednesday, August 03, 2016 4:33 PM

To: Harya Dillon <hdillon@uci.edu>

Subject: Re: - Network Analyst - best route shapefile

GeoNet <https://community.esri.com/?et=watches.email.thread>

Network Analyst - best route shapefile

reply from Dan Patterson

<https://community.esri.com/people/Dan_Patterson?et=watches.email.thread>

in *Community

Help* - View the full discussion

<https://community.esri.com/message/625842?et=watches.email.thread#comment-625842>

Reply
0 Kudos
Highlighted
MVP Esteemed Contributor

closer... consider a scenario, whereby people 'might' consider using a bus if they are with 0.5 km of a bus stop/bus line.  Buffer you bus route to produce a polygon and select all of the 'origins' are within the buffer.  They are your candidates for using the bus (I have seen places use 1 km, but the choice is yours).

Similarly perform the same analysis for the destinations.  typically, any destination outside a comfortable walking distance is going to be a car destination (except for the committed bus commuter).

So my opinion would be to limit your candidates and do a round of analysis before you throw everything into the mix.  report back too

Reply
0 Kudos
Highlighted
New Contributor

Thanks. That would work with the bus scenario.

What if now I want to know if the best route would include a (potentially)

tolled segment?

From: Dan Patterson

Sent: Wednesday, August 03, 2016 5:26 PM

To: Harya Dillon <hdillon@uci.edu>

Subject: Re: - Network Analyst - best route shapefile

GeoNet <https://community.esri.com/?et=watches.email.thread>

Network Analyst - best route shapefile

reply from Dan Patterson

<https://community.esri.com/people/Dan_Patterson?et=watches.email.thread>

in *Community

Help* - View the full discussion

<https://community.esri.com/message/625846?et=watches.email.thread#comment-625846>

Reply
0 Kudos
Highlighted
MVP Esteemed Contributor

that would be a network "penalty" which you can model, but I will let you look at incorporating other costs into the analysis, like turns etc.

Reply
0 Kudos
Highlighted
New Contributor

I am not quite there yet. This is really preliminary and crude. I want to

get a (crude for now) an estimate about how many of the trips would

involve segments that might be tolled in the future. We have no prior

information about the toll level – hence assuming perfect inelasticity. At

this point, I am interested at how many trips would be impacted if tolls

are implemented.

I understand that a shapefile would be created for each O-D pair? Is this

correct?

From: Dan Patterson

Sent: Wednesday, August 03, 2016 5:37 PM

To: Harya Dillon <hdillon@uci.edu>

Subject: Re: - Network Analyst - best route shapefile

GeoNet <https://community.esri.com/?et=watches.email.thread>

Network Analyst - best route shapefile

reply from Dan Patterson

<https://community.esri.com/people/Dan_Patterson?et=watches.email.thread>

in *Community

Help* - View the full discussion

<https://community.esri.com/message/625849?et=watches.email.thread#comment-625849>

Reply
0 Kudos
Highlighted
MVP Esteemed Contributor

you can or a featureclass could be made, depending on your needs.  Just don't try to run all at once! You might consider clustering points so it is a zone to zone, the size of the zone could be a block, where you might have 10 origins (people) and 1 or 2 destinations.  Simplify before you expand otherwise you might be overwhelmed and patterns may be hidden

Reply
0 Kudos