High-level geospatial data visualization library for Python.

Here: https://github.com/ResidentMario/geoplot/blob/master/geoplot/geoplot.py#L6

geoplot seems to be outdated w/r/t geopandas. I'm not sure where the norm_cmap function got moved, but it doesn't seem to be there anymore...

I just cloned geoplot and it's >100MB!

I bet a lot of this comes from the data module...could you host that data online somewhere (e.g. figshare?) and have users download it similarly to the geopandas.datasets module?

Trying to follow the KDEplot + webmap tutorial but the webmap renders upside down and no combination of parameter specifications or tweaking will get it to render in the correct orientation.

Code copied from here: https://residentmario.github.io/geoplot/gallery/plot_boston_airbnb_kde.html

Resulting plot from plt.show():

Any insights on how to fix / flip the webmap?

The geoplot release seems to have broken the geopandas examples (the voronoi one). I am getting the following error on our readthedocs build:

Unexpected failing examples:
/home/docs/checkouts/readthedocs.org/user_builds/geopandas/checkouts/latest/examples/plotting_with_geoplot.py failed leaving traceback:
Traceback (most recent call last):
  File "/home/docs/checkouts/readthedocs.org/user_builds/geopandas/checkouts/latest/examples/plotting_with_geoplot.py", line 80, in <module>
    linewidth=0)
  File "/home/docs/checkouts/readthedocs.org/user_builds/geopandas/conda/latest/lib/python3.6/site-packages/geoplot/geoplot.py", line 2133, in voronoi
    geoms = _build_voronoi_polygons(df)
  File "/home/docs/checkouts/readthedocs.org/user_builds/geopandas/conda/latest/lib/python3.6/site-packages/geoplot/geoplot.py", line 2687, in _build_voronoi_polygons
    ls = np.vstack([np.asarray(infinite_segments), np.asarray(finite_segments)])
  File "/home/docs/checkouts/readthedocs.org/user_builds/geopandas/conda/latest/lib/python3.6/site-packages/numpy/core/shape_base.py", line 234, in vstack
    return _nx.concatenate([atleast_2d(_m) for _m in tup], 0)
ValueError: all the input arrays must have same number of dimensions

I've consolidated the example datasets under a geopandas-like geoplot.datasets domain (here).

This has the beneficial effect that it makes all of the examples in the geoplot documentation, especially the ones in the gallery, immediately reproducible for the user. However, the drawback is that I also have to distribute the example data with the library. After hewing and hawing every which way, I've gotten that down to a ~10 MB examples.zip file.

@choldgraf I would like your feedback on this idea. Is a 10 MB add-on like this an acceptable load for such a library? Or should I maybe provide an nltk-like downloader instead?

Hi there. Here is the alteration that seems to pass your tests and fixes the legends. In particular

• circle style for _paint_carto_legend is now consistent with the plot itself (edge colors, face colors)
• by default Sankey lines have round end filling
• Sankey now renders a new type of legend, _paint_sankey_legend that draws lines, also style-consistent.
• Legends now add legend artist to the plot on the contrary to calling ax.Legend.

A few questions/warnings

• I don't mean this PR to be accepted (at this time), but rather to be a place for you to express your concerns/opinions

1. I didn't get why you plot markers with 20*(dscale(x))**(1/2) in _paint_carto_legend. As I "fixed" markersize in the N1, it might now affect your example images.

2. while adding artists fixed an issue with double legend for me, it might affect plotting multiple subplots. To my shape, I didn't check that yet. Still, I think there should be a way to have multiple legends on multiple subplots.

3. As I plot network, I'd like to have one scaling factor both for nodes (points) and edges (sankey). Currently the only way to do that is to calculate upper bound from one using data from another.

4. Generally, do you think if it makes sense to add yet another type of plot - [geo]network? and handle everything with nodes and legends there?

I tried to install geoplot in various different ways but can never get rid of errors.

I first tried to create a conda venv from scratch and install: conda install geoplot -c conda-forge I open a Jupyter Notebook, go through your quick start and get stuck at the very beginning when importing geopandas.

Thinking that this was an issur vith my venv, I decided to completely clone your github repo ResidentMario/geoplot and creata a brand new environment from your environment.yml. I did that, and run my Jupyter Notebook from that geoplot-dev venv: I still cannot import geopandas without error.

That is strange, I have no clue on how to get set-up and use geoplot at that point. Would be glad if you can help.

ImportError Traceback (most recent call last) in ----> 1 import geopandas as gpd

~/opt/anaconda3/envs/geoplot-dev/lib/python3.8/site-packages/geopandas/init.py in 3 from geopandas.array import _points_from_xy as points_from_xy # noqa 4 ----> 5 from geopandas.io.file import read_file # noqa 6 from geopandas.io.sql import read_postgis # noqa 7 from geopandas.tools import sjoin # noqa

~/opt/anaconda3/envs/geoplot-dev/lib/python3.8/site-packages/geopandas/io/file.py in 4 import pandas as pd 5 ----> 6 import fiona 7 from shapely.geometry import mapping 8 from shapely.geometry.base import BaseGeometry

~/opt/anaconda3/envs/geoplot-dev/lib/python3.8/site-packages/fiona/init.py in 82 os.environ["PATH"] = os.environ["PATH"] + ";" + libdir 83 ---> 84 from fiona.collection import BytesCollection, Collection 85 from fiona.drvsupport import supported_drivers 86 from fiona.env import ensure_env_with_credentials, Env

~/opt/anaconda3/envs/geoplot-dev/lib/python3.8/site-packages/fiona/collection.py in 7 8 from fiona import compat, vfs ----> 9 from fiona.ogrext import Iterator, ItemsIterator, KeysIterator 10 from fiona.ogrext import Session, WritingSession 11 from fiona.ogrext import buffer_to_virtual_file, remove_virtual_file, GEOMETRY_TYPES

ImportError: dlopen(/Users/Armellini1/opt/anaconda3/envs/geoplot-dev/lib/python3.8/site-packages/fiona/ogrext.cpython-38-darwin.so, 2): Library not loaded: @rpath/libnetcdf.15.dylib Referenced from: /Users/Armellini1/opt/anaconda3/envs/geoplot-dev/lib/libgdal.26.dylib Reason: image not found

When I try to import geoplot, I get the following error: ModuleNotFoundError: No module named 'rasterio.io'. I have installed the rasterio package, but I see that it is missing the .io and .vrt modules, both of which are needed to import geoplot. I thought about raising this issue on the rasterio repository but figured I would start here. I'm using Anaconda on OSX.

Here's the full error message:

import geoplot Traceback (most recent call last):

File "", line 1, in import geoplot

File "/Users/pregiste/opt/anaconda3/lib/python3.7/site-packages/geoplot/init.py", line 1, in from .geoplot import (

File "/Users/pregiste/opt/anaconda3/lib/python3.7/site-packages/geoplot/geoplot.py", line 17, in import contextily as ctx

File "/Users/pregiste/opt/anaconda3/lib/python3.7/site-packages/contextily/init.py", line 7, in from .place import Place, plot_map

File "/Users/pregiste/opt/anaconda3/lib/python3.7/site-packages/contextily/place.py", line 6, in from .tile import howmany, bounds2raster, bounds2img, _sm2ll, _calculate_zoom

File "/Users/pregiste/opt/anaconda3/lib/python3.7/site-packages/contextily/tile.py", line 20, in from rasterio.io import MemoryFile

ModuleNotFoundError: No module named 'rasterio.io'

I'm looking to create static category bins so my coloring doesn't change between plots. I'm not sure how to achieve this sort of behavior currently and was looking for assistance.

My voroni call looks like this, and the data is changing between invocations

gplt.voronoi(
		    data, 
		    projection=ep.proj, clip=clipper,
		    hue='lmp', cmap='Reds', 
		    k=10, scheme='fisher_jenks',
		    ax=ax, alpha=0.9,zorder=7,
		    edgecolor='white', legend=True,extent=clipper.total_bounds
		    )

ex. plot 1

ex. plot 2

I would just like to create the bins beforehand, or maybe preprocess the data in some way so that the legend and colors are consistent between plots so I can use them in an animation.

Thanks!

I refer on Choropleth subplots

to subplot spatial map with data.

individually it works pretty well.

ax = gplt.polyplot(AB_base )


def power_scale(minval, maxval):

    def scalar(val):
        val = val + abs(minval) + 1
        return (val/100)**3/150
    return scalar


gplt.kdeplot(elevation,  ax=ax,linewidth=0,
             legend = True,
             shade_lowest=False,
             cbar = True,
              clip=AB_base.geometry, cmap='summer_r',
                  shade=True, alpha = 0.6)

gplt.pointplot(geo_station,  ax=ax,
               scale= 'elev(m)',  k = None,
                           limits =(1,30) , 
                          scale_func= power_scale,                             
              hue=geo_station['elev(m)'].astype(float), cmap='viridis_r',
                               alpha = 0.8,
               legend=True, legend_var='hue', 
              )

plt.title("~~~")
plt.show()

but when I subplot more variables like this:

def plot_to_ax(state, ax):
   
    
    gplt.polyplot(AB_base, ax = ax)

    gplt.kdeplot(elevation, linewidth=0.0, ax = ax,
                 shade_lowest=False,
                 clip=AB_base.geometry, cmap='summer_r',
                 shade=True, alpha = 0.6)

    gplt.pointplot(geo_station, k = None, ax=ax,
                   scale= state, limits =(1,30),                           
                   hue= state, 
                   cmap='viridis_r',alpha = 0.8,
                   legend=True, legend_var='hue'
                   )
    

# Finally, plot the data.
f, axarr = plt.subplots(2, 2, figsize=(5, 5))

plt.subplots_adjust(top=0.95)

plot_state_to_ax('ANUSPLIN_output', axarr[0][0])
axarr[0][0].set_title('ANUSPLIN (n=6679268)')

plot_state_to_ax('CaPA_output', axarr[0][1])
axarr[0][1].set_title('CaPA (n=854647)')

plot_state_to_ax('NARR_output', axarr[1][0])
axarr[1][0].set_title('NARR(n=215065)')

plot_state_to_ax('TPS_output', axarr[1][1])
axarr[1][1].set_title('TPS (n=126661)')

all plot in one result, I think the method should be good - as reference is from the official tutorial. tried many ways and still get stuck in this problem for 3 days...

any advice or solutions?

thanks

Hi!

I am struggling a bit with the KDEplot and clipping the data on my geometry. When I perform the code on the whole of Germany it works fine, and I get this map:

While when I try to perform it on the states of Germany, it gives weird results, which are different for each state. See here a few examples:

This is the code:

def sentimentKDE (data, state, title):
    proj = gcrs.Mercator()

    fig = plt.figure(figsize=(15, 10))
    ax1 = plt.subplot(121, projection=proj)
    ax2 = plt.subplot(122, projection=proj) 

    gplt.kdeplot(data[data['label']=='Positive'], shade=True, cmap='Reds',clip=geom_germany[geom_germany['GEN'] == state].geometry,ax=ax1,)
    gplt.polyplot(geom_germany[geom_germany['GEN'] == state].geometry,ax=ax1,  facecolor='white')
    ax1.set_title("Positive tweets", fontsize=20)

    gplt.kdeplot(data[data['label']=='Negative'], shade=True, cmap='Reds',clip=geom_germany[geom_germany['GEN'] == state].geometry,ax=ax2)
    gplt.polyplot(geom_germany[geom_germany['GEN'] == state].geometry,ax=ax2, facecolor='white',)
    ax2.set_title("Negative tweets", fontsize=20)

    plt.savefig(title)

Do you have any idea what is going wrong? Would love the help.

Thank you in advance!

It appears that an expected warning is no longer raised:

_______________________ TestClip.test_clip_init_default ________________________

self = <mixin_tests.TestClip testMethod=test_clip_init_default>

    def test_clip_init_default(self):
        clipmixin = self.create_clipmixin()
    
        # UserWarning because we have a narrow clip
>       with pytest.warns(UserWarning):
E       Failed: DID NOT WARN. No warnings of type (<class 'UserWarning'>,) were emitted. The list of emitted warnings is: [].

tests/mixin_tests.py:593: Failed

I'm trying to create a plot using data that is not in EPSG:4326. This is necessary since I need to make computations which require a cartesian coordinate system. This MWE returns a blank plot:

import geopandas as gpd
import geoplot as gplt
import matplotlib.pyplot as plt

# read the following dataset using geopandas method: gplt.datasets.get_path('ny_census')
ny = gpd.read_file(gplt.datasets.get_path('ny_census'))

# transform the x,y data into that of a projected CRS
ny_proj = ny.to_crs(32016)

# create a variable for population density
ny_proj['population_density'] = ny_proj['POP2000'] / ny_proj.area

# create a choropleth map of population density
gplt.choropleth(ny_proj, hue = 'POP2000')
plt.show()
# it's blank

This works, however, using the native plot() method:

ny_proj.plot()
plt.show()

I know that I could feed gplt.choropleth() an unprojected geodataframe and then specify the projection argument as something that is projected, but this feels a bit cumbersome and will result in counterintuitive results. E.g., I'd have to attach a variable back to the original ny geodataframe, and if there were units attached to this variable (e.g., meters or feet), these would not match the linear unit of the CRS.

Am I missing something in my geoplot MWE or was it only intended to work with EPSG:4326?

Using the kdeplot, with only one dataset, it appears that isolines are overlapping instead of merging, producing two local maxima instead of a consistent isoline around all features.

Apologies if this turns out to be a cartopy issue (I'm not deeply familiar with cartopy), but after digging around, I can't find anyone else running into it.

I'm hitting two odd edge cases where setting projection=gcrs.OSGB() creates a blank plot, but AlbersEqualArea() plots fine.

1. A GeoDataFrame read from the following (Great Britain and Northern Ireland borders), dropping Northern Ireland with .dropna([11]) plots when extent is set but not when it's omitted. AlbersEqualArea() plots fine without extent.

gplt.polyplot(
    uk_borders_partial,
    zorder=1,
    projection=gcrs.OSGB(approx=False),
    extent=uk_borders_partial.total_bounds,
    edgecolor='gray',
    figsize=(12,12)
    )

2. Setting the previous to ax= and tacking on the following KDEPlot creates a blank plot if clip=uk_borders_partial (Northern Ireland is dropped). Using the full GeoDataFrame or any subset of the full GeoDataFrame that does not drop Northern Ireland, e.g. uk_borders_full[5:12], plots exactly as expected. The code, with Ireland dropped, also plots as expected with AlbersEqualArea(). Just not with OSGB().

gplt.kdeplot(
    lichen_2011_11,
    clip=uk_borders_partial,
    projection=gcrs.OSGB(approx=False),
    cmap='Reds',
    shade=True,
    ax=ax,
    extent=uk_borders_partial.total_bounds,
    )

Would the latter be an issue with how clip gets passed to seaborn kdeplot?

Thank you for your help, and for this beautiful library.

Hi,

it would be really helpful to have a style argument (Grouping variable that will produce points with different markers) similar to what done in seaborn (example)

I experimented with a multiprocessing implementation of the quadtree algorithm in the following notebook. The multiprocessing component isn't all that useful, but the changes also include logic for better, faster splitting rules. If I take this to completion it will represent a very significant speedup in this algorithm's runtime.