|
POST
|
Hello Scott. I'm sorry for your trouble, but the good news is that there's an easy fix for you. For some reason, your street data did not end up with vertices at the locations where the StopsOnStreets intersect them: Without vertices at those locations, travelers cannot travel between the streets and the transit lines, hence why the transit lines are never being used. Vertices should be automatically created when you run the Connect Network Dataset Transit Sources To Streets tool. I will investigate further to see why this didn't happen. However, regardless of what happened in the first place, there is an easy way for you to fix your network. 1) If you made any changes to your network's configuration (I don't think you did...), run the Create Template From Network Dataset tool to preserve your network's configuration into a file. Skip this step if you didn't make any changes in that case you can just use the existing template that comes with the tutorial (the one you already used). 2) Delete your network dataset. 3) Run the Integrate geoprocessing tool with the inputs configured as follows: The Integrate tool will create vertices at the intersections of the input feature classes. Setting the XY Tolerance to 0 ensures that your input features will not be moved at all. 4) Re-create your network dataset using the template. 5) Run Build Network to re-build your network. After this procedure, I was able to see the vertices in the expected location, and the network used the transit lines. Assuming you're able to get up and running successfully and you have some time, could you tell me the following to help me investigate the cause of the original problem? What version of ArcGIS Pro are you using? Did you follow the tutorial steps exactly (using your own data), or did you deviate from that and do something else? Are you able to share your GTFS data with me so I can run some tests?
... View more
02-12-2021
09:12 AM
|
0
|
2
|
2767
|
|
POST
|
The Location-Allocation solver cannot display the actual paths taken through the network from Facilities to Demand Points, even though it uses the network paths under the hood. It's just a limitation of the solver and is done this way to improve performance. If you want to display the actual paths taken for visualization purposes, you will need to use the Demand Points and chosen facilities as input stops for a Route analysis (or maybe Closest Facility) and solve that analysis. If the chosen facilities are reliably the closest to each demand point, or you only have one chosen facility, Closest Facility is the easiest thing to do. Otherwise, you can use Route, and you must load a copy of the facility for each demand point. You need to use the RouteName field to designate a unique identifier for each Demand Point/Facility pair to ensure that a separate route is generated between each demand point and its associated facility.
... View more
02-08-2021
10:08 AM
|
0
|
6
|
5906
|
|
POST
|
This page explain the output of the Copy Traversed Source Features tool: https://pro.arcgis.com/en/pro-app/latest/tool-reference/network-analyst/copy-traversed-source-features-output.htm If the routes do not travel over edges, junctions, or turns, then the associated output table will be empty.
... View more
02-05-2021
12:33 PM
|
0
|
1
|
2407
|
|
POST
|
The problem is that your Closest Facility analysis is failing to solve because it has no incidents and facilities in it. You need to first configure your Closest Facility layer and add the desired inputs, then solve it. Once you are convinced that the results are as expected, then you can run Copy Traversed Source Features. Here is a tutorial to help you learn how to set up a Closest Facility analysis: https://desktop.arcgis.com/en/arcmap/latest/extensions/network-analyst/exercise-4-finding-the-closest-fire-stations.htm
... View more
02-05-2021
11:21 AM
|
0
|
0
|
2408
|
|
POST
|
Hello @DavisTurner. First of all, you did the right thing in removing your GTFS data from the network dataset. BetterBusBuffers is about measuring access to public transit, not about measuring access to other facilities by public transit. Thus, you don't want your network dataset to model travelers taking public transit, you just want to model them walking. Second, the size of your GTFS data is probably not the issue. The Preprocess GTFS tool does the heavy lifting and reads in all the data from the different agencies. Once it's passed that point, constructing the Service Area polygon buffers becomes the difficult part. It will definitely take longer if you have more transit stops, but anyway, the slowness is in accounting for overlapping polygons and trying to "flatten" them. Regarding your error: Are you getting this error on the "Step 1 – Preprocess Buffers" tool? That seems more likely, but please let me know if my assumption is incorrect. This tool can often be finicky, and it's because it's doing some fairly complex geometric procedures to handle the overlapping polygons. Something strange is happening under the hood, and it's kind of impossible to know why. The only thing to do is to try again by altering some of the settings to change the geometry of the resulting polygons slightly and hoping that the geometry error will be avoided. Here are a few things you can try changing to see if it avoids the problem: Change the Polygon Trim setting to a different value. Update your impedance attribute slightly, like changing the assumed pedestrian walk speed by a small amount. Changing the Buffer Size from 20 minutes to some other number (I know, not ideal...) This last suggestion applies only if you are using ArcMap and not ArcGIS Pro: Calculate a Service Area Index on your network dataset. Basically, we introduced a newer, better algorithm for creating Service Area polygons which fixed a lot of weird geometry problems and made all-around better polygons. This is the default in ArcGIS Pro, but if you want to use it in ArcMap, you have to explicitly choose to calculate a Service Area Index in your network dataset settings. You just have to check on the check-box and run Build Network again. Note: If you are not using ArcGIS Pro, I recommend switching to Pro anyway to get overall better functionality. The tool might take less time to run or at the very least be less likely to run out of memory. Hope this helps!
... View more
02-05-2021
08:02 AM
|
0
|
1
|
1712
|
|
POST
|
After calculating your routes, run the Copy Traversed Source Features tool. This tool creates an individual feature for each segment of the network dataset that is traveled upon, and it has a field indicating which route or facility it is used by. You can do some joins to copy over the number of people from the original data. It also has the ObjectID of the network dataset source it is associated with. You can find all the traversed segments with that ObjectID and sum the number of people associated with each one.
... View more
02-04-2021
04:35 PM
|
0
|
5
|
2415
|
|
POST
|
This tutorial describes how to create a Model using Network Analyst tools: https://pro.arcgis.com/en/pro-app/latest/help/analysis/networks/create-a-model-for-route-analysis.htm The example uses the Route solver, but the basic procedure would be the same for Service Area or Location-Allocation.
... View more
02-02-2021
07:43 AM
|
0
|
0
|
5912
|
|
POST
|
Hello Seunghoon. The GTFS specification only started to allow blank values for stop_lat and stop_lon fairly recently when it adopted support for platforms.txt. The Add GTFS to a Network Dataset toolbox you are referring to is now deprecated and has not been updated since that change. That's why you're getting an error. In general, I do not recommend using the old Add GTFS to a Network Dataset toolbox anymore. Instead, you should use the newer, better tools available out-of-the box in ArcGIS Pro. Learn more here. However, switching to Pro will not actually solve this problem until ArcGIS Pro 2.8 is released (later this year). For now, you will need to manually update your GTFS data. Since the blank stop_lat and stop_lon rows probably represent non-geographic location_type values of 3 or 4, I suggest you just delete the rows. They will not be used in the analysis anyway.
... View more
01-30-2021
02:52 PM
|
0
|
0
|
2020
|
|
POST
|
That build error indicates that the transit stop has not successfully connected to a street feature. There are a couple of likely reasons for this: The street dataset you're using does not cover the entire geographic area that your transit system covers, so there are some stops that cannot find a street to snap to. There are some transit stops in your GTFS data that are in the wrong geographic place. For example, you might have a few stray stops in the middle of the ocean or in Antarctica or something because the lat/lon coordinates are wrong. Your street data covers the correct geographic extent, but it is missing some smaller roads or sections where there are transit stops, so the stops cannot find anything to snap to within the search tolerance. The search distance you used when running the Connect Network Dataset Transit Sources To Streets tool was too small, and for some reason, your transit stops are offset significantly from the street centerlines. For this reason, they failed to snap to the streets. I think you need to take a close look at both the locations of your transit stops and also your street data to determine which of the above cases is causing your problem. Once you've determined that, then you can figure out how to fix it. If the problem is only for a few stops, and they're in an area where you don't expect them to be used for your analysis, then you can safely ignore the build errors. The rest of the network will continue to function. Otherwise, you will need to resolve the issue. You might need to edit the stop locations (or correct the original GTFS and start over), update your street data or use a different street dataset, or re-run Connect Network Dataset Transit Sources To Streets with a larger search tolerance.
... View more
01-27-2021
08:25 AM
|
1
|
2
|
5097
|
|
POST
|
Are you doing the workflow manually at this point? You could reduce the amount of work you need to do by automating the procedure using Python or Model Builder. I don't know that this would reduce the overall time, but at least you could press start and then take a lunch break. Does this seem like something you might want to do?
... View more
01-11-2021
07:58 AM
|
0
|
0
|
5922
|
|
POST
|
Hello SakshiKosta. Unfortunately, the Add GTFS To A Network Dataset tool is now deprecated, so I am not able to provide extensive assistance with it. From the error message, it looks like there might be a problem with the spatial reference you have chosen for your feature dataset. I suggest starting over with a new file geodatabase and feature dataset. Make sure to carefully select a valid spatial reference for the feature dataset, and then try again. In general, though, I recommend you switch to using ArcGIS Pro for your transit-related analyses. We have made equivalent tools available in the core product, and these tools are more modern and thoroughly tested, and you can get help on them from Esri Support. Learn more.
... View more
01-04-2021
07:58 AM
|
1
|
0
|
2882
|
|
POST
|
I don't completely understand your workflow, but basically this is what you need to do: In ArcGIS Pro, go to the Backstage area (hit the blue Project button), and click the Python option on the left: Click Manage Environments to clone your default environment. Once that's done and the new environment is activated, click the Add Packages button. Search for networkx. You should be able to click the item for networkx in the list and install it. After that, I think you should be good to go.
... View more
12-17-2020
07:59 AM
|
1
|
0
|
7214
|
|
POST
|
Yes, that's exactly what it means. The OD Cost Matrix calculation determined that the fastest route between the origin and the destination was just to walk, so it returned the walk time.
... View more
12-16-2020
08:06 AM
|
0
|
0
|
861
|
|
POST
|
It is not unexpected to get different results with ArcGIS Online and your own network. The network dataset used by ArcGIS Online is a comprehensive dataset using the highest-quality data available from our data vendors. It includes detailed street and path geometry as well as an extensive set of attributes defining the characteristics of the streets, which are used in calculating impedances and restrictions. Your own data may have some differences in geometry - roads in slightly different places, certain roads missing, etc. Thus, the optimal paths between places on the network may differ slightly. You might also have some connectivity errors - places where the streets seem topologically connect but aren't actually logically connected in the network. Even if the geometry of the streets is identical, there is probably a difference in attributes between your data and the ArcGIS Online data. The impedance calculations may be different, and ArcGIS Online might be using some restrictions that you are not. Finally, it should be noted that the ArcGIS Online data contains information about turn restrictions - specific turns that are not allowed or which incur a delay. If you use the "Driving Distance" travel mode, these turn restrictions will be used, even though, for your use case, this probably doesn't make sense. If your own network doesn't have any turn restrictions, it may allow you to build your pipes following the streets around a corner where vehicles are not allowed to turn, like an intersection where you cannot turn left. But for pipes, I presume this doesn't matter. Thus, your data may actually be more appropriate for this situation. Note that there are ways you can adjust the travel modes even when using ArcGIS Online to turn off this restriction, so the issue is not insurmountable. As far as testing your network, it's pretty hard to interpret the results of a Location-Allocation problem. It might be better to do some Service Areas where you turn on the Lines output. This will help you look for holes in your network where things might be cut off. You can also use Route to test some specific point-to-point routes to see if the results make sense. If you are concerned about topological connectivity in your network, you can try using this downloadable sample tool to identify topologically disconnected areas: https://www.arcgis.com/home/item.html?id=b2227f745a6c4e1c94dd57810729d2a9
... View more
12-16-2020
08:05 AM
|
1
|
2
|
7219
|
|
POST
|
There are a few steps/levels that need to be considered, but it seems like everything is mostly set up correctly already. Below is a sort of reverse-order list of things that need to be considered. I'm starting with the part you select when doing your analysis, and it's sort of backing up all the way through the network configuration. You would have to do the steps in the reverse order, but I'm explaining it this way because I think it's easier to understand conceptually. When you do your analysis, make sure to select your "automobil_distance" travel mode on the ribbon. That ensures that you are using the travel mode you specifically configured for your purpose. You need to make sure the "automobil_distance" travel mode is configured the way you want it. A travel mode is essentially just a bucket of other settings that model a particular mode of travel. You want to use an impedance based on distance, so your Impedance property in the Costs section needs to be set to a cost attribute that models distance. In your case, it's set to a cost attribute called "Length", which shows units of meters in your screenshot, so this seems good. You need to make sure that the "Length" cost attribute is correctly calculating the length of the streets. To do that, switch to the Costs tab in the network dataset properties Travel Attributes page, find "Length" in the list of cost attributes, and examine its properties. My guess is that this is the automatically-generated distance attribute you always get when you create a new network dataset, so it's probably just using the shape length of the roads to calculate the length. If this is good enough for you, just leave it. Otherwise, you might want to read more about cost attributes and evaluators so you can properly configure the way it calculates distance.
... View more
12-04-2020
08:10 AM
|
1
|
5
|
7241
|
| Title | Kudos | Posted |
|---|---|---|
| 1 | a week ago | |
| 1 | 2 weeks ago | |
| 1 | 2 weeks ago | |
| 1 | 06-12-2026 01:53 PM | |
| 1 | 04-21-2026 08:39 AM |
| Online Status |
Offline
|
| Date Last Visited |
a week ago
|