Vector Data Visualization in R#494
Vector Data Visualization in R#494HarshiniGirish wants to merge 22 commits intoMAAP-Project:developfrom
Conversation
|
Check out this pull request on See visual diffs & provide feedback on Jupyter Notebooks. Powered by ReviewNB |
docs/source/technical_tutorials/working_with_r/vector data visualization.ipynb
Outdated
Show resolved
Hide resolved
docs/source/technical_tutorials/working_with_r/vector data visualization.ipynb
Outdated
Show resolved
Hide resolved
docs/source/technical_tutorials/working_with_r/vector data visualization.ipynb
Outdated
Show resolved
Hide resolved
docs/source/technical_tutorials/working_with_r/vector data visualization.ipynb
Outdated
Show resolved
Hide resolved
|
Overall I think the use of GEDI here makes the tutorial more complicated that it needs to be. I think we should try to use some data out of MAAP STAC and focus on the different visualization options since using Jupyter with R is a little unusual.
Source formats, we should probably include:
https://docs.maap-project.org/en/develop/technical_tutorials/visualization/visualize_lonboard.html is the closest existing example. <- something is wrong with this notebook the plots are missing. @hrodmn @chuckwondo can you think of good vector data to visualize? https://github.com/MAAP-Project/maap-documentation-examples/blob/main/gedi-subset/notebooks/lonboard_exploration.ipynb I think is too much for R |
I'm trying to think of data that would be useful to add to a MAAP-centric graphic. Maybe an ecoregion type of dataset - this source has a zipped shapefile: https://www.worldwildlife.org/publications/terrestrial-ecoregions-of-the-world Or maybe National Forest boundaries from USFS: https://data.fs.usda.gov/geodata/edw/datasets.php?dsetCategory=boundaries In my forest science days I was always making maps of field plot locations which usually had a point at the center and a radius drawn around them. There are not very many examples of Geoparquet out in the wild - maybe something like this? https://planetarycomputer.microsoft.com/dataset/noaa-climate-normals-tabular |
|
Up for review! @wildintellect @smk0033 |
|
View / edit / reply to this conversation on ReviewNB wildintellect commented on 2025-08-01T18:18:56Z I think we also want a static plot example using plot, and tmap |
|
View / edit / reply to this conversation on ReviewNB wildintellect commented on 2025-08-01T18:18:57Z Need to put the sample data into a Shared bucket so the path works for everyone. |
|
View / edit / reply to this conversation on ReviewNB wildintellect commented on 2025-08-01T18:18:57Z I think we need to find a simpler way to read the data into sf. the geoarrow package has a read_geoparquet_sf() function but I'm not sure if that's still the best practice
We should check if adding |
|
thank you @wildintellect. Its up for another review! |
|
View / edit / reply to this conversation on ReviewNB wildintellect commented on 2025-08-04T23:40:00Z Circle Radius is way too big, try 50 or 100, might be good to also include agbd and agbd_se in the popup
|
|
View / edit / reply to this conversation on ReviewNB wildintellect commented on 2025-08-04T23:40:00Z Line #4. df_sf <- st_as_sf(df, crs = 4326) Don't repeat the conversion, use the same from above, conversion to SF should happen in the Load section. Load and convert to SF |
|
View / edit / reply to this conversation on ReviewNB wildintellect commented on 2025-08-04T23:40:02Z Remember AGBD is the single most important value in the data (would be nice to apply color ramp to it even) |
|
View / edit / reply to this conversation on ReviewNB wildintellect commented on 2025-08-04T23:40:03Z World isn't loaded/imported before this. Need to add Also remember to move all Library calls to the top. |
|
Made the changes. Up for review @wildintellect |
|
View / edit / reply to this conversation on ReviewNB wildintellect commented on 2025-08-05T23:56:22Z already did this 2 above |
|
View / edit / reply to this conversation on ReviewNB wildintellect commented on 2025-08-05T23:56:23Z I'm confused, if you already have a sf dataframe what does this section do? Sorry I didn't catch this before, but you shouldn't need to make another dataframe once you have an sf object it works with leaflet directly. https://rstudio.github.io/leaflet/articles/json.html
center <- st_bbox(df_sf) %>%
st_as_sfc %>%
st_centroid %>%
st_coordinates
center
# Create the leaflet map centered on the mean coordinates
leaflet_map <- leaflet(df_sf) %>%
addProviderTiles("OpenStreetMap") %>%
addCircles(
radius = 400,
popup = ~paste("Shot Number:", nrow(df_sf))
) %>%
setView(
lat = center[2],
lng = center[1],
zoom = 12
)
# Show the map
leaflet_map
|
|
View / edit / reply to this conversation on ReviewNB wildintellect commented on 2025-08-05T23:56:23Z also clean this up to use the df_sf object |
|
View / edit / reply to this conversation on ReviewNB wildintellect commented on 2025-08-05T23:56:24Z get rid of redundant display_df duplicating data like this could lead to memory issue with very large data |
|
View / edit / reply to this conversation on ReviewNB wildintellect commented on 2025-08-05T23:56:25Z also df_sf |
|
@wildintellect up for review ! |
|
View / edit / reply to this conversation on ReviewNB wildintellect commented on 2025-08-11T23:01:51Z Line #6. center <- c(mean(c(bbox["xmin"], bbox["xmax"])), mean(c(bbox["ymin"], bbox["ymax"]))) Line 5 and 6 can be simplified to 1 line , like other st_centroid usage in the notebook. |
|
View / edit / reply to this conversation on ReviewNB wildintellect commented on 2025-08-11T23:01:52Z Line #11. center <- st_coordinates(st_centroid(st_union(df_sf))) Already calculated previously, just pick unique variable names to avoid overwriting |
|
View / edit / reply to this conversation on ReviewNB wildintellect commented on 2025-08-11T23:01:53Z If you're going to plot world or even regional like the next map, you might as well just plot the centroid, it will plot much faster.
Maybe this map should be regional, with centroid as a single point, and the next map with TMAP should zoom in to the bbox of the the data. |
|
@wildintellect thanks! up for review |
|
View / edit / reply to this conversation on ReviewNB wildintellect commented on 2025-08-19T17:45:15Z [GEDI Subsetter](https://github.com/MAAP-Project/gedi-subsetter/) dataset Can also link to the MAAP tutorial on using GEDI subsetter https://docs.maap-project.org/en/latest/science/GEDI/GEDI_L4A.html
Parquet -> GeoParquet |
|
View / edit / reply to this conversation on ReviewNB wildintellect commented on 2025-08-19T17:45:16Z two periods in a row |
|
View / edit / reply to this conversation on ReviewNB wildintellect commented on 2025-08-19T17:45:16Z Line #7. if (!identical(sf::st_crs(world), crs)) {
the data variable is World not world |
|
View / edit / reply to this conversation on ReviewNB wildintellect commented on 2025-08-19T17:45:17Z Line #13. dx <- (bb$xmax - bb$xmin) * 0.15 this is actually too zoomed in *250 works so you can actually tell where this is. |
|
View / edit / reply to this conversation on ReviewNB wildintellect commented on 2025-08-19T17:45:18Z It's good that this is zoomed in now, but we need to apply colors to the points. I'm also researching how to add context to the map, but hit a bit a of a problem...
I installed 0.9 with but it still doesn't seem to work.
So I went about trying to find a Raster we could use as background shaded_relief <- rnaturalearth::ne_download(scale=50,
type="HYP_50M_SR",
category="raster")
bb <- st_bbox(df_sf)
pad <- 25
dx <- (bb["xmax"] - bb["xmin"]) * pad
dy <- (bb["ymax"] - bb["ymin"]) * pad
bb_exp <- structure(
c(xmin = bb["xmin"] - dx, ymin = bb["ymin"] - dy,
xmax = bb["xmax"] + dx, ymax = bb["ymax"] + dy),
class = class(bb), crs = st_crs(df_sf)
)
shaded_relief_crop <- terra::crop(shaded_relief, bb_exp)
tmap_mode("plot")
#tm_shape(world, bbox = bb_exp) +
# tm_polygons(col = "grey90", border.col = "grey60") +
tmap::tm_shape(shaded_relief_crop) +
tmap::tm_rgb(r=3, g=2, b=1) +
tm_shape(df_sf) +
tm_dots(size = 0.1) +
tm_layout(
title = "GEDI Footprint Locations (Static Map)",
frame = FALSE,
legend.show = FALSE
)
but this data isn't not high resolution so it also doesn't work well when zoomed in.
So I think this plan can work, we just need to find an easy to get layer to put behind the points, and give the points colors like the above maps.
tmap documentation to reference https://r-tmap.github.io/tmap-book/layers.html |
|
It does seem to be a bug that tm_basemap (based on maptiles) doesn't work in Jupyter notebooks. I was able to use direct functions to cache the tiles and plot as a raster. This will be useful if/when we manage to get updated packages though. 
|
|
@wildintellect up for review! |
250 not 0.25 the map should look like Make that change and we'll call version 1 of this tutorial done. |
|
@HarshiniGirish Looks like you need to resolve a conflict now before we can approve. |
3 participants
This notebook I referenced.
A lot of steps here is data extraction and data analysis. Do I keep everything or just reference the other notebook.
I don't know what else can be done in terms of visualization.