I'd like to to summarise the road lengths within the parish of Burford.
I'm logged into ArcGIS Online's Map Viewer and using the Summarize Within tool:
- Features to summarize: OS Open Roads Vector Map Data - SP RoadLink (Burford is within the SP Area)
- Summary areas: Parishes May 2023 Boundaries EW BFC - PAR MAY 2023 EW BFC (Covers England & Wales)
- Calculate statistics:
Summarized geometry units - Miles
Field statistics - length > Sum - Environment settings:
Processing extent - Full extent
Obviously, this gives me results for the entire SP area. The other Environment settings are also not what I want.
So, how do I clip the features within Burford to Burford's boundaries?
After 48 hours of trying and failing, any help greatly appreciated!
Accepted Solutions
I agree with @BobBooth1.
1. It sounds like you have correctly filtered your polygon layer, great! Remaining decisions are mostly about excluding features in your lines layer to lower the number of credits consumed.
2. For the environment settings you have a few viable options:
- Processing extent: You have a few options:
- Layer - This will use the extent defined on the layer (even if you have a filter, it's pulling it from the original layer), and isn't filtering down the features. This will produce the same result as below, but will cost more credits than if you exclude some features for analysis (via other env settings).
- Display extent - this will help help cut down on credit consumption. Note that it won't "clip" the input features, it will use anything that intersects with your display extent. I would recommend using this option, and make sure when you run the analysis you are "viewing" the area of interest. If you don't want think about what's covered in the extent when you run the data, you can choose "Coordinates" and then "Set coordinated from display" - this will set the extent when you click the button, and won't change even if you zoom or pan around before clicking run.
- Output coordinate system - I would keep it as the default unless you know your result should be in a different coordinate system than the inputs.
Another option, if you have a way to filter the roads based on region, you could also do that. Then when you run the analysis, it would only use your roads within Burford. Currently, you have > 260,000 features of roads, and it looks like you probably only care about a very small subset of them. So if you can filter them on an attribute, I would recommend that over the environment settings.
A future note - if this is data you continuously use in the future, and you only care about Burford you could make a view of the data. Then, you could use the view in the future (which lets you apply spatial and attribute filters on a layer) and they look like they are a layer in your org. Doing that would mean you wouldn't need to filter the next time you open it in the map, nor would you mistakenly run analysis on all the features. https://doc.arcgis.com/en/arcgis-online/manage-data/create-hosted-views.htm
I don't think you need to make a view for this analysis, it's more to make long term maintenance + analysis easier once you've figured out your future needs + complete this analysis.
Hi @GrayElkington you can either "clip" the features using the overlay layers tool (union option) or put a filter on the polygon features.
Option 1: Use the Overlay tool
- Add the line layer and polygon layer to the map
- Create a filter on the polygon layer (I'm assuming your Parishes Boundaries layer has a field you can filter on). Something like "Area == Burford". See screenshot on filtering below.
- Run Overlay Layers -> Intersects with the lines and filtered polygons. Make sure the output geometry is set to "Line"
- This should result in the lines that are within the Burford Area.
Another option requires fewer tool runs, and is similar:
- Add the point and line layers to the map
- Apply the filter to the polygon layer (as you did before)
- Run Summarize Within using the filtered layer
I think those should both work. Here is what setting a filter looks like:
- Select the layer you want to filter
- Click the filter button
- Set the condition to specify the features of interest
- Save it
Once you save it, you will only see the filtered features on the map.
Here are a few references to help:
- https://doc.arcgis.com/en/arcgis-online/create-maps/apply-filters-mv.htm
Please let me know if that doesn't address your needs,
Sarah Ambrose
Product Engineer
Web Analysis
I agree with @BobBooth1.
1. It sounds like you have correctly filtered your polygon layer, great! Remaining decisions are mostly about excluding features in your lines layer to lower the number of credits consumed.
2. For the environment settings you have a few viable options:
- Processing extent: You have a few options:
- Layer - This will use the extent defined on the layer (even if you have a filter, it's pulling it from the original layer), and isn't filtering down the features. This will produce the same result as below, but will cost more credits than if you exclude some features for analysis (via other env settings).
- Display extent - this will help help cut down on credit consumption. Note that it won't "clip" the input features, it will use anything that intersects with your display extent. I would recommend using this option, and make sure when you run the analysis you are "viewing" the area of interest. If you don't want think about what's covered in the extent when you run the data, you can choose "Coordinates" and then "Set coordinated from display" - this will set the extent when you click the button, and won't change even if you zoom or pan around before clicking run.
- Output coordinate system - I would keep it as the default unless you know your result should be in a different coordinate system than the inputs.
Another option, if you have a way to filter the roads based on region, you could also do that. Then when you run the analysis, it would only use your roads within Burford. Currently, you have > 260,000 features of roads, and it looks like you probably only care about a very small subset of them. So if you can filter them on an attribute, I would recommend that over the environment settings.
A future note - if this is data you continuously use in the future, and you only care about Burford you could make a view of the data. Then, you could use the view in the future (which lets you apply spatial and attribute filters on a layer) and they look like they are a layer in your org. Doing that would mean you wouldn't need to filter the next time you open it in the map, nor would you mistakenly run analysis on all the features. https://doc.arcgis.com/en/arcgis-online/manage-data/create-hosted-views.htm
I don't think you need to make a view for this analysis, it's more to make long term maintenance + analysis easier once you've figured out your future needs + complete this analysis.
Given the number of parishes, I think a different approach is warranted.
It looks as though there are around 1,280 parishes that cover the SP_Roadlink block of road segments.
I tried the analysis using the Summarize Within tool in ModelBuilder in ArcGIS Online.
The process ran for about 30 minutes (less actually, but I wasn't watching when it finished) and the ModelBuilder session cost 26 credits. You pay for Analysis Session time rather than features processed (mostly) when using ModelBuilder in ArcGIS Online.
The awkward thing is, you said 14,000 parishes. The way the road data is tiled, you'll need to run the analysis on each block of road segments (so, roughly 30 times?).
There may be some strange edge effects, for parishes on the edges of blocks of roads.
Notice the light colored edge parishes with low road length sum values.
If it were my analysis, I'd do it in ArcGIS Pro with local geodatabase feature class data.
I'd get a local copy of the road segments and parishes data, merge all of the road segments back into a single feature class, and use an ArcGIS Pro overlay tool (Intersect or Union or Summarize Within) to make a single massive feature class of all road segments with the parish name attribute on them. Then (unless I used Summarize Within) I'd calculate summary statistics to get the Sum of the length of the features by Parish name (or better, Parish ID code, if there is one in the dataset, because names might repeat).
Best,
Bob
