Interpolation function for finding total corridor width #28
Comments
|
A reprex would really help! This is like Latin to me lol! |
|
Illustration as a taster, fully reproducible but let me create a tidy example that generates this plot:
|
|
Reprex with input data. Geo challenge (sounds simple): find the widths of the target geometries based only on the source roads: net_target = sf::read_sf("https://github.com/nptscot/osmactive/releases/download/v0.1/net_target.gpkg")
net_source = sf::read_sf("https://github.com/nptscot/osmactive/releases/download/v0.1/net_source.gpkg")
plot(net_target) |
|
@wangzhao and I have a slow solution that answers the question with The answer is not quite the average: when the centerline target geoms have 2 lanes in the source the values should be summed (the value returned by |
|
interpolate_extensive() does the summation. Can you give that a shot?
…On Mon, Jan 6, 2025 at 09:39 Robin Lovelace ***@***.***> wrote:
@wangzhao <https://github.com/wangzhao> and I have a slow solution that
answers the question with rnet::join() and some post-processing. Can
share that, and can have a bash with anime (already have on larger
datasets, which show that {anime} is very fast and good at finding the
average values with interpolate_intensive(), this is not quite the
average we're after here: when the centerline target geoms have 2 lanes in
the source the values should be summed. This can be done using the
shared_len outputs in anime objects I think but have not figured out how.
I sense a function, e.g. interpolate_widths() (or more general name)
coming on. Need to do nursery run now but can provide more info a bit
later..
—
Reply to this email directly, view it on GitHub
<#28 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/ADHIKLCVMXK63EHMLYUM55T2JK5TRAVCNFSM6AAAAABUWAHK22VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDKNZTGU4TGMJRGQ>
.
You are receiving this because you commented.Message ID:
***@***.***>
|
|
Already tried. I'll provide a follow-up reprex with that plus more attempts. |
|
Reprex: library(dplyr)
#>
#> Attaching package: 'dplyr'
#> The following objects are masked from 'package:stats':
#>
#> filter, lag
#> The following objects are masked from 'package:base':
#>
#> intersect, setdiff, setequal, union
net_target = sf::read_sf("https://github.com/nptscot/osmactive/releases/download/v0.1/net_target.gpkg")
net_source = sf::read_sf("https://github.com/nptscot/osmactive/releases/download/v0.1/net_source.gpkg") |>
sf::st_cast("LINESTRING")
#> Warning in
#> st_cast.sf(sf::read_sf("https://github.com/nptscot/osmactive/releases/download/v0.1/net_source.gpkg"),
#> : repeating attributes for all sub-geometries for which they may not be
#> constant
# plot(net_target)
# plot(net_source["width"])
matches = anime::anime(
source = net_source,
target = net_target,
angle_tolerance = 35,
distance_tolerance = 10
)
widths_interpolated = net_source |>
reframe(width_intensive = anime::interpolate_intensive(width, matches))
widths_extensive = net_source |>
reframe(width_extensive = anime::interpolate_extensive(width, matches))
net_joined = bind_cols(net_target, widths_interpolated)
net_joined = bind_cols(net_joined, widths_extensive)
net_joined$diff = net_joined$width_extensive - net_joined$width_intensive
mapview::mapview(net_joined, zcol = "diff") +
mapview::mapview(net_source, zcol = "width")
#> file:////tmp/Rtmp85U4F6/file19c826a8d1244/widget19c82524a847a.html screenshot completedCreated on 2025-01-06 with reprex v2.1.1 |
|
Neither approach is right but
The centerline in the target object is composed of 2 roads in the source object, with values of 8 and 9 m. So the total width should be 17m. The |
Robinlovelace
changed the title
|
The issue with the extensive interpolation, I think, is that small sections have just as much an influence on the value returned as long sections. We can work out the correct width based on |
|
I'm not following why you would multiple the value by the length that would make some incredibly large numbers in some cases. The idea is that the value is equally distributed across the length of the source line and we're summing only the shared portion of that line. I think it may be handy to see the overlapping bounding boxes that are matched. Would you mind making an issue for that? I can work on a function that would be helpful for debugging these questions. I actually think the extensive value may be closer to correct than maybe you think? Partially because of the very forgiving angle tolerance you've provided which is picking up the roundabouts. George has an interpolated value of 30ish. The parallel line south has a value of 8, about all of which should be matched. North of it has 9. Then the northeastern roundabout has a value of 7 (which based on the angle tolerance might be matched almost entirely? hard to know without overlaying bounding boxes) 
Then south west has a value of 6 and northwest has a value of 7 
Adding all of that up gives us a value of |
|
George Street has a value of 16.3 according to the
# Solution with rnet_join
library(dplyr)
#>
#> Attaching package: 'dplyr'
#> The following objects are masked from 'package:stats':
#>
#> filter, lag
#> The following objects are masked from 'package:base':
#>
#> intersect, setdiff, setequal, union
net_target = sf::read_sf("https://github.com/nptscot/osmactive/releases/download/v0.1/net_target.gpkg")
net_source = sf::read_sf("https://github.com/nptscot/osmactive/releases/download/v0.1/net_source.gpkg") |>
sf::st_cast("LINESTRING")
#> Warning in
#> st_cast.sf(sf::read_sf("https://github.com/nptscot/osmactive/releases/download/v0.1/net_source.gpkg"),
#> : repeating attributes for all sub-geometries for which they may not be
#> constant
library(stplanr)
net_target$length = sf::st_length(net_target) |>
as.numeric()
target_joined_rnet = stplanr::rnet_join(
net_target,
net_source,
max_angle_diff = 35,
dist = 10,
segment_length = 10,
subset_x = TRUE,
crop = TRUE
)
#> Warning: attribute variables are assumed to be spatially constant throughout
#> all geometries
#> Warning in st_cast.sf(sf::st_cast(x, "MULTILINESTRING"), "LINESTRING"):
#> repeating attributes for all sub-geometries for which they may not be constant
#> Joining with `by = join_by(id)`
#> Linking to GEOS 3.12.1, GDAL 3.8.4, PROJ 9.3.1; sf_use_s2() is TRUE
#> Warning: st_centroid assumes attributes are constant over geometries
target_joined = dplyr::left_join(
target_joined_rnet,
net_target |>
sf::st_drop_geometry(),
by = "id"
)
widths = target_joined |>
sf::st_drop_geometry() |>
dplyr::group_by(id) |>
dplyr::mutate(
total_length_y = sum(length_y, na.rm = TRUE),
length_x_updated = if_else(total_length_y < length, total_length_y, length)
) |>
dplyr::transmute(
id,
width = width * length_y / length_x_updated
) |>
dplyr::summarise(avg_road_width = round(sum(width, na.rm = TRUE),2))
net_target_final = dplyr::left_join(
net_target,
widths,
by = "id"
)
mapview::mapview(net_target_final, zcol = "avg_road_width")
#> file:////tmp/Rtmp9X7zk5/file200095f6ea2e7/widget200097defb16a.html screenshot completedCreated on 2025-01-06 with reprex v2.1.1 |
Maybe but wouldn't that result change depending on how many short linestrings there that match? One could make the answer 100+ just by having many 1 mm linestrings anywhere along the line, for example. |
It would if there were 5 parallel lines with a value of 6 matching all the way along the target geometry but in this case there are 2 parallel lines that share the majority of the length of the target, and then some small links that share only a small portion that have a disproportionate impact on the result, right? |
Happy to do so, that would be useful in general, but not sure it's relevant for this issue. |
|
The question is more or less how you interpret the value in those portions. We assume with the extensive interpolation that each variable is equally distributed along the line. For example we have a source line that is 10 units long and a value of 100. Our target line overlaps 3 units. We extract the value 30 from the source and assign it to the target. Do we then want to scale that assigned value by something to assign the value equally over the whole distance of the target? |
|
I think we do want to distribute the value of 100 in that case. The question is more about what to do when you have many source geometries matching each target, with some source geometries having a large amount of overlap and some having small amounts of overlap. That's what's happening in the reprex above and explains the unwanted result stating that the corridor is 30 m wide despite there only being 2 lanes ~8 m wide each (not real data but has the same structure as the data we're using). Can show with another image building on that if that would be helpful, do you have a link on excalidraw/similar? |
|
It sums it up for each of the overlapping lines Using the same george street example, if you reduce the 
The challenge that you're running into is that the interpolation is including the widths of the roundabouts which it seems you don't want. |
|
That does look right in that case but imagine a small section that had exactly the same angle and a very short length, e.g. 1% of the length, that has a width of 20. We don't want the value 20 to be added to the entire section, but we don't want to ignore it completely. The solution we're after, which is implemented in the reprex in #28 (comment), is for the influence of the value of each source to vary depending on its length, if that makes sense. As far as I understand it, with |
|
|
|
I think it's just a question of multiplying the value by the target weight for our use case: library(dplyr)
net_target = sf::read_sf("https://github.com/nptscot/osmactive/releases/download/v0.1/net_target.gpkg")
net_source = sf::read_sf("https://github.com/nptscot/osmactive/releases/download/v0.1/net_source.gpkg") |>
sf::st_cast("LINESTRING")
# plot(net_target)
# plot(net_source["width"])
matches = anime::anime(
source = net_source,
target = net_target,
angle_tolerance = 35,
distance_tolerance = 10
)
matches_df = anime::get_matches(matches)
summary(matches_df$source_id)
net_source_matches = left_join(
net_source |>
mutate(source_id = row_number()),
matches_df
)
net_target_aggregated = net_source_matches |>
sf::st_drop_geometry() |>
group_by(row_number = target_id) |>
summarise(
v_weighted = sum(width * target_weighted, na.rm = TRUE),
v_max_shared = sum(width * shared_len) / max(shared_len)
)
net_target_joined = left_join(
net_target |>
mutate(row_number = row_number()),
net_target_aggregated
)
summary(net_target_joined$v_weighted)
summary(net_target_joined$v_max_shared)
mapview(net_target_joined, zcol = "v_weighted") +
mapview(net_source, zcol = "width")Which generates the correct value for George Street, even using the deliberately forgiving
|
|
See minimal example in #30 for how I got there, useful to break it down into the basics and starting from simple examples. Question for you on that, what should the value of the centreline be in this case?
And can you see that there are use cases in which the value needed is different from the simple sum of all matching source values (13 = 10 + 1 + 2)? |
Of course, but I'm trying to make sure that the math is interpretable. Can you try this: library(dplyr)
net_target = sf::read_sf("https://github.com/nptscot/osmactive/releases/download/v0.1/net_target.gpkg")
net_source = sf::read_sf("https://github.com/nptscot/osmactive/releases/download/v0.1/net_source.gpkg") |>
sf::st_cast("LINESTRING")
matches = anime::anime(
source = net_source,
target = net_target,
angle_tolerance = 35,
distance_tolerance = 10
)
matches_df = anime::get_matches(matches)
matches_df |>
mutate(
target_len = target_len[target_id]
) |>
left_join(mutate(net_source, source_id = row_number())) |>
group_by(target_id) |>
filter(shared_len > 0) |>
print(n = Inf) |>
summarise(v_ext_int = sum(width * source_weighted * target_weighted), source_widths = toString(width)) |>
inner_join(mutate(net_target, target_id = row_number())) |>
sf::st_as_sf() |>
mapview::mapview(zcol = "v_ext_int") +
mapview::mapview(net_source, zcol = "width")The important part here is : summarise(v_ext_int = sum(width * source_weighted * target_weighted))The idea here is two fold—which is what I think you're getting at.
this is keeping same angle tolerance etc 
|
|
Looks promising and nice and simple, will give it a bash and report back. |
|
I get this error: |
|
Did you mean |
|
The results after fixing the error look better than those returned by
Reprex in details. library(dplyr)
#>
#> Attaching package: 'dplyr'
#> The following objects are masked from 'package:stats':
#>
#> filter, lag
#> The following objects are masked from 'package:base':
#>
#> intersect, setdiff, setequal, union
net_target = sf::read_sf("https://github.com/nptscot/osmactive/releases/download/v0.1/net_target.gpkg")
net_source = sf::read_sf("https://github.com/nptscot/osmactive/releases/download/v0.1/net_source.gpkg") |>
sf::st_cast("LINESTRING")
#> Warning in
#> st_cast.sf(sf::read_sf("https://github.com/nptscot/osmactive/releases/download/v0.1/net_source.gpkg"),
#> : repeating attributes for all sub-geometries for which they may not be
#> constant
matches = anime::anime(
source = net_source,
target = net_target,
angle_tolerance = 35,
distance_tolerance = 10
)
matches_df = anime::get_matches(matches)
matches_df |>
mutate(
target_len = shared_len[target_id]
) |>
left_join(mutate(net_source, source_id = row_number())) |>
group_by(target_id) |>
filter(shared_len > 0) |>
print(n = Inf) |>
summarise(v_ext_int = sum(width * source_weighted * target_weighted), source_widths = toString(width)) |>
inner_join(mutate(net_target, target_id = row_number())) |>
sf::st_as_sf() |>
mapview::mapview(zcol = "v_ext_int") +
mapview::mapview(net_source, zcol = "width")
#> Joining with `by = join_by(source_id)`
#> # A tibble: 266 × 9
#> # Groups: target_id [107]
#> target_id source_id shared_len source_weighted target_weighted target_len
#> <int> <int> <dbl> <dbl> <dbl> <dbl>
#> 1 2 63 30.7 0.848 0.579 31.4
#> 2 3 46 31.4 0.502 0.955 6.11
#> 3 4 60 6.11 0.833 0.0846 32.7
#> 4 4 124 32.7 1.00 0.452 32.7
#> 5 6 23 147. 0.975 0.900 74.5
#> 6 7 80 74.5 0.929 0.882 0
#> 7 7 180 4.55 0.375 0.0539 0
#> 8 9 36 70.0 0.976 0.638 70.0
#> 9 9 31 19.1 0.799 0.174 70.0
#> 10 10 149 0.675 0.0576 0.0269 19.1
#> 11 14 114 7.78 0.0796 0.996 19.7
#> 12 14 33 19.7 0.200 2.53 19.7
#> 13 15 107 0.338 0.0155 0.00626 0.338
#> 14 15 114 54.2 0.555 1.00 0.338
#> 15 16 126 4.14 0.593 0.432 54.2
#> 16 16 63 5.50 0.152 0.574 54.2
#> 17 18 80 5.70 0.0711 0.0852 5.50
#> 18 19 173 1.89 0.249 0.00927 5.70
#> 19 19 7 186. 1 0.914 5.70
#> 20 19 178 4.15 0.440 0.0204 5.70
#> 21 19 8 188. 1 0.922 5.70
#> 22 20 99 152. 0.991 0.816 1.89
#> 23 20 106 33.7 0.935 0.181 1.89
#> 24 20 112 33.9 1.00 0.182 1.89
#> 25 20 109 4.45 0.0707 0.0239 1.89
#> 26 23 81 9.62 0.170 0.692 188.
#> 27 23 69 4.29 0.0584 0.309 188.
#> 28 23 3 11.8 0.208 0.846 188.
#> 29 23 74 2.02 0.0272 0.146 188.
#> 30 24 147 10.8 0.897 0.0870 0
#> 31 25 13 73.4 0.649 0.896 152.
#> 32 26 15 10.1 0.200 0.222 33.7
#> 33 26 164 37.2 0.740 0.821 33.7
#> 34 27 25 26.7 1 0.316 33.9
#> 35 28 96 56.7 0.735 1.05 0
#> 36 29 87 78.7 1 0.648 4.45
#> 37 30 15 40.3 0.800 0.971 9.62
#> 38 30 150 0.641 0.0559 0.0154 9.62
#> 39 31 137 74.1 0.992 0.372 4.29
#> 40 31 38 17.6 1 0.0886 4.29
#> 41 31 167 12.8 1 0.0643 4.29
#> 42 31 12 94.2 0.941 0.474 4.29
#> 43 32 43 4.58 0.311 0.0762 11.8
#> 44 32 105 3.62 0.210 0.0602 11.8
#> 45 32 116 34.0 1 0.565 11.8
#> 46 32 209 2.22 0.0437 0.0369 11.8
#> 47 35 12 5.94 0.0593 0.675 10.8
#> 48 36 90 35.5 1 0.951 73.4
#> 49 36 94 3.17 0.0495 0.0849 73.4
#> 50 37 16 40.1 0.593 0.859 10.1
#> 51 38 91 85.8 0.547 0.972 37.2
#> 52 38 41 3.45 0.344 0.0391 37.2
#> 53 39 72 419. 0.995 0.999 0
#> 54 39 122 5.15 0.336 0.0123 0
#> 55 39 19 413. 1 0.984 0
#> 56 39 4 0.948 0.00512 0.00226 0
#> 57 41 9 21.1 0.109 0.871 56.7
#> 58 43 125 22.4 1 0.109 40.3
#> 59 43 9 194. 1 0.947 40.3
#> 60 43 169 2.10 0.307 0.0103 40.3
#> 61 43 175 157. 1 0.767 40.3
#> 62 43 10 12.5 0.836 0.0608 40.3
#> 63 44 59 49.1 0.884 0.785 0.641
#> 64 44 62 8.07 1 0.129 0.641
#> 65 44 126 0.841 0.120 0.0134 0.641
#> 66 45 77 71.9 0.955 0.958 74.1
#> 67 45 89 2.66 0.187 0.0355 74.1
#> 68 46 50 16.4 0.311 0.301 17.6
#> 69 48 82 11.5 0.494 0.676 94.2
#> 70 48 29 3.33 0.0279 0.195 94.2
#> 71 49 91 71.0 0.453 0.849 4.58
#> 72 50 160 2.95 0.0834 0.0739 3.62
#> 73 50 92 35.1 0.798 0.880 3.62
#> 74 50 28 27.1 1.00 0.680 3.62
#> 75 50 98 0.457 0.0121 0.0115 3.62
#> 76 50 95 11.1 1 0.279 3.62
#> 77 50 137 0.594 0.00795 0.0149 3.62
#> 78 50 151 4.34 0.252 0.109 3.62
#> 79 51 55 0.318 0.00635 0.00676 34.0
#> 80 51 141 33.1 1.00 0.704 34.0
#> 81 51 142 14.0 0.429 0.299 34.0
#> 82 52 46 30.2 0.484 0.994 2.22
#> 83 56 113 15.4 1 0.398 40.1
#> 84 56 99 1.45 0.00944 0.0373 40.1
#> 85 57 33 78.9 0.800 0.822 85.8
#> 86 58 152 47.7 0.978 0.971 3.45
#> 87 58 147 1.24 0.103 0.0253 3.45
#> 88 59 39 32.0 0.677 0.708 419.
#> 89 59 30 12.3 0.317 0.271 419.
#> 90 60 160 32.4 0.917 0.870 0
#> 91 60 98 37.0 0.975 0.992 0
#> 92 60 64 4.62 0.0404 0.124 0
#> 93 61 5 26.0 0.937 0.127 5.15
#> 94 61 200 19.2 1 0.0938 5.15
#> 95 61 61 171. 1 0.838 5.15
#> 96 61 159 4.76 0.250 0.0233 5.15
#> 97 61 130 171. 1 0.836 5.15
#> 98 61 88 2.06 0.158 0.0101 5.15
#> 99 61 181 5.99 0.392 0.0293 5.15
#> 100 63 89 11.6 0.813 0.251 0
#> 101 64 154 4.96 0.0868 0.143 0.948
#> 102 64 86 12.6 0.291 0.362 0.948
#> 103 64 156 11.0 0.413 0.316 0.948
#> 104 65 13 23.2 0.205 0.562 21.1
#> 105 65 185 27.0 0.581 0.654 21.1
#> 106 65 165 9.33 0.201 0.226 21.1
#> 107 67 70 29.3 0.550 0.716 194.
#> 108 67 111 23.1 0.436 0.564 194.
#> 109 67 120 8.52 0.751 0.208 194.
#> 110 68 88 6.69 0.514 0.0328 2.10
#> 111 68 6 190. 1 0.933 2.10
#> 112 68 181 2.43 0.159 0.0119 2.10
#> 113 68 79 189. 1.00 0.928 2.10
#> 114 68 148 2.20 0.250 0.0108 2.10
#> 115 68 178 2.92 0.309 0.0143 2.10
#> 116 68 182 5.28 0.601 0.0259 2.10
#> 117 70 52 98.6 0.969 0.811 12.5
#> 118 71 133 5.85 0.157 0.0774 49.1
#> 119 73 2 116. 1.00 0.411 0.841
#> 120 73 68 83.6 0.914 0.297 0.841
#> 121 73 57 4.23 0.0244 0.0150 0.841
#> 122 73 73 9.32 0.0410 0.0331 0.841
#> 123 73 71 11.3 0.715 0.0400 0.841
#> 124 73 4 12.9 0.0697 0.0458 0.841
#> 125 76 123 30.0 1.00 0.319 16.4
#> 126 76 154 11.5 0.201 0.122 16.4
#> 127 76 155 51.7 1.00 0.549 16.4
#> 128 76 70 23.9 0.450 0.254 16.4
#> 129 76 111 4.43 0.0838 0.0470 16.4
#> 130 76 156 9.11 0.343 0.0966 16.4
#> 131 77 139 3.08 0.0936 0.0461 11.5
#> 132 77 138 42.0 1 0.628 11.5
#> 133 77 206 22.2 1 0.332 11.5
#> 134 77 34 0.526 0.126 0.00787 11.5
#> 135 79 73 10.1 0.0446 0.851 0
#> 136 79 71 4.10 0.261 0.344 0
#> 137 79 4 0.283 0.00153 0.0237 0
#> 138 79 13 4.39 0.0388 0.368 0
#> 139 79 18 9.56 0.0445 0.803 0
#> 140 81 209 13.4 0.264 0.719 0
#> 141 81 129 4.53 0.146 0.243 0
#> 142 82 29 102. 0.855 0.714 0
#> 143 84 36 1.71 0.0239 0.190 35.1
#> 144 84 201 7.12 0.410 0.790 35.1
#> 145 85 106 2.35 0.0651 0.105 27.1
#> 146 85 20 20.2 0.681 0.899 27.1
#> 147 85 109 22.3 0.355 0.997 27.1
#> 148 87 204 81.8 0.941 0.891 11.1
#> 149 87 192 5.56 0.546 0.0605 11.1
#> 150 88 75 98.8 0.992 0.984 0.594
#> 151 89 207 133. 1 0.363 4.34
#> 152 89 66 68.6 1.00 0.188 4.34
#> 153 89 27 9.40 1.00 0.0257 4.34
#> 154 90 114 35.7 0.365 0.911 0.318
#> 155 90 77 3.36 0.0447 0.0859 0.318
#> 156 90 33 41.4 0.419 1.06 0.318
#> 157 91 21 115. 0.561 0.885 33.1
#> 158 91 22 2.16 0.0482 0.0166 33.1
#> 159 93 81 46.9 0.830 0.576 30.2
#> 160 93 3 44.9 0.792 0.551 30.2
#> 161 93 64 34.8 0.304 0.427 30.2
#> 162 93 65 37.0 0.331 0.454 30.2
#> 163 95 165 33.0 0.711 0.774 15.4
#> 164 95 185 14.8 0.319 0.346 15.4
#> 165 96 157 36.5 0.974 0.230 1.45
#> 166 96 158 36.0 0.937 0.227 1.45
#> 167 96 14 61.5 1.00 0.388 1.45
#> 168 98 58 132. 0.997 0.993 0
#> 169 99 211 2.42 0.0963 0.0153 47.7
#> 170 100 21 84.9 0.414 0.945 0
#> 171 104 45 99.6 0.996 0.982 32.4
#> 172 105 166 0.222 0.0165 0.000795 37.0
#> 173 105 213 14.4 1 0.0515 37.0
#> 174 105 190 13.5 1 0.0485 37.0
#> 175 105 212 0.521 1 0.00186 37.0
#> 176 106 30 26.2 0.677 0.419 4.62
#> 177 106 198 23.3 1 0.373 4.62
#> 178 106 197 6.24 0.217 0.0997 4.62
#> 179 107 72 1.84 0.00436 0.0100 0
#> 180 107 57 173. 1 0.942 0
#> 181 107 4 184. 0.993 1.00 0
#> 182 107 73 9.32 0.0410 0.0507 0
#> 183 108 50 35.0 0.663 0.315 26.0
#> 184 109 69 69.2 0.942 0.947 19.2
#> 185 109 73 3.54 0.0156 0.0485 19.2
#> 186 109 74 72.1 0.970 0.987 19.2
#> 187 110 203 125. 1.00 0.317 0
#> 188 110 194 18.3 0.603 0.0463 0
#> 189 110 202 30.3 1 0.0768 0
#> 190 110 193 37.7 0.582 0.0956 0
#> 191 110 49 5.87 0.126 0.0149 0
#> 192 110 82 7.43 0.318 0.0188 0
#> 193 112 22 42.6 0.952 0.780 4.76
#> 194 113 73 215. 0.943 1.00 171.
#> 195 113 4 0.283 0.00153 0.00132 171.
#> 196 113 18 215. 1 1.00 171.
#> 197 113 74 0.209 0.00281 0.000973 171.
#> 198 114 11 9.17 0.0687 0.0958 2.06
#> 199 116 96 20.5 0.265 0.612 5.99
#> 200 117 201 10.2 0.590 0.0839 11.6
#> 201 117 32 26.6 0.953 0.218 11.6
#> 202 117 128 86.4 1 0.708 11.6
#> 203 118 64 75.1 0.656 0.997 0
#> 204 118 65 74.9 0.669 0.995 0
#> 205 118 98 0.479 0.0126 0.00636 0
#> 206 120 129 26.5 0.854 0.408 12.6
#> 207 120 37 22.5 1.00 0.347 12.6
#> 208 120 127 13.1 1.00 0.201 12.6
#> 209 120 60 1.22 0.167 0.0189 12.6
#> 210 123 142 18.7 0.571 0.184 23.2
#> 211 123 161 33.2 1.00 0.327 23.2
#> 212 123 210 14.9 1.00 0.147 23.2
#> 213 123 40 20.9 1.00 0.206 23.2
#> 214 123 104 7.99 1 0.0786 23.2
#> 215 124 139 29.8 0.906 0.210 27.0
#> 216 126 126 2.00 0.286 1.02 29.3
#> 217 131 209 35.2 0.692 1.07 0
#> 218 132 85 28.7 0.437 0.274 2.43
#> 219 133 11 124. 0.931 0.940 189.
#> 220 135 199 20.6 1 0.489 2.92
#> 221 135 20 9.43 0.319 0.223 2.92
#> 222 135 109 36.1 0.574 0.856 2.92
#> 223 135 121 12.5 0.644 0.296 2.92
#> 224 135 26 6.27 0.0897 0.149 2.92
#> 225 137 120 2.82 0.249 0.0385 98.6
#> 226 137 67 66.3 1 0.905 98.6
#> 227 137 121 6.91 0.356 0.0943 98.6
#> 228 137 26 63.7 0.910 0.870 98.6
#> 229 137 72 0.331 0.000785 0.00451 98.6
#> 230 137 115 4.07 0.178 0.0555 98.6
#> 231 137 122 10.2 0.664 0.139 98.6
#> 232 138 119 73.2 1 0.445 0
#> 233 138 186 46.5 0.968 0.283 0
#> 234 138 145 28.7 1 0.174 0
#> 235 138 187 14.8 0.952 0.0901 0
#> 236 141 164 13.1 0.260 0.597 0
#> 237 141 59 6.44 0.116 0.293 0
#> 238 145 16 17.4 0.257 0.906 9.32
#> 239 145 52 3.11 0.0306 0.162 9.32
#> 240 146 55 49.7 0.994 0.670 11.3
#> 241 146 1 13.4 0.587 0.181 11.3
#> 242 147 94 59.1 0.923 0.952 12.9
#> 243 148 24 172. 1 1.00 30.0
#> 244 149 184 13.2 0.570 0.339 11.5
#> 245 149 183 5.82 0.217 0.149 11.5
#> 246 151 17 118. 0.894 0.954 23.9
#> 247 151 152 1.07 0.0220 0.00863 23.9
#> 248 151 189 6.16 1 0.0497 23.9
#> 249 153 179 4.24 0.333 0.0505 9.11
#> 250 153 146 27.9 1.00 0.332 9.11
#> 251 154 107 21.4 0.984 0.795 3.08
#> 252 154 108 7.18 0.657 0.267 3.08
#> 253 156 154 5.06 0.0887 0.0755 22.2
#> 254 156 86 30.5 0.704 0.455 22.2
#> 255 156 208 29.8 0.701 0.445 22.2
#> 256 156 162 4.92 1 0.0735 22.2
#> 257 158 50 1.36 0.0258 0.0253 10.1
#> 258 158 51 52.2 0.284 0.972 10.1
#> 259 159 211 22.7 0.904 0.180 4.10
#> 260 159 54 15.2 1.00 0.120 4.10
#> 261 160 56 17.1 1 0.251 0.283
#> 262 160 97 11.9 0.573 0.174 0.283
#> 263 160 35 19.2 1.00 0.282 0.283
#> 264 161 204 0.615 0.00708 0.00943 4.39
#> 265 161 118 12.1 1.00 0.186 4.39
#> 266 161 47 4.07 1.00 0.0624 4.39
#> # ℹ 3 more variables: osm_id <chr>, width <dbl>, geom <LINESTRING [m]>
#> Joining with `by = join_by(target_id)`
#> file:////tmp/RtmpdgHUKP/file6d9512ddb8860/widget6d9515483be73.html screenshot completedCreated on 2025-01-07 with reprex v2.1.1 |
|
The result of the width = sum(width * target_weighted, na.rm = TRUE),method I used isn't much different:
|
|
Another avenue that I'm interested in trying: dividing the total shared length by the the length of the target geometries. Tried but failed that somewhere else but can have another bash. In terms of the theory underlying this approach, to the limited extent that there is any, is this interpolation approach somewhat of a halfway house between extensive and intensive, or is it just length-weighted extensive? |
|
Are there any cases where using sum(width * source_weighted * target_weighted)Rather than just leads to a substantially different and potentially better/worse result? |
Full description + reprex incoming. I think this is an ideal use case highlighting the flexibility of
animeso should lead to good docs + content for write-up.The text was updated successfully, but these errors were encountered: