pythonpython-3.xbokehprojectioncartodb

Change projection of tile provider in bokeh in EPSG:3857 ("web mercator") to my source's in EPSG:4326

I am able to use Bokeh to plot glyphs from a geopandas dataframe over a Google Map using the gmap() function.

from bokeh.io import output_notebook, show, output_file
from bokeh.plotting import figure
from bokeh.models import GeoJSONDataSource, LinearColorMapper, ColorBar
from bokeh.palettes import brewer#Input GeoJSON source that contains features for plotting.

import json

from bokeh.models import ColumnDataSource, GMapOptions
from bokeh.plotting import gmap

def make_dataset(df, candidate):
    #df_copy = df.copy()
    df_copy = get_df(candidate)
    merged_json = json.loads(df_copy.to_json())#Convert to String like object.
    json_data = json.dumps(merged_json)
    geosource = GeoJSONDataSource(geojson = json_data)
    return geosource

def make_plot(candidate):
    src = make_dataset(df,candidate)
    #Input GeoJSON source that contains features for plotting.    
    p = figure(title = 'Results of candidate X', plot_height = 600 , plot_width = 950, toolbar_location = None)

    map_options = GMapOptions(lat=42, lng=44, map_type="roadmap", zoom=7)

    p = gmap("my-key", map_options, title="Austin")
    p.xgrid.grid_line_color = None
    p.ygrid.grid_line_color = None#Add patch renderer to figure. 
    p.patches('xs','ys', source = src,fill_color = {'field' :'results', 'transform' : color_mapper},
              line_color = 'black', line_width = 0.25, fill_alpha = 1)#Specify figure layout.
    p.add_layout(color_bar, 'below')#Display figure inline in Jupyter Notebook.
    output_notebook()#Display figure.
    return p

It gives me:

enter image description here

However when I plot using Carto as a provider as explained here there is an error in the axes:

    tile_provider = get_provider(Vendors.CARTODBPOSITRON)

    # range bounds supplied in web mercator coordinates
    p = figure(x_range=(-2000000, 6000000), y_range=(-1000000, 7000000))#, x_axis_type="mercator", y_axis_type="mercator")
    p.add_tile(tile_provider)
    p.xgrid.grid_line_color = None
    p.ygrid.grid_line_color = None#Add patch renderer to figure. 
    p.patches('xs','ys', source = src,fill_color = {'field' :'results', 'transform' : color_mapper},
              line_color = 'black', line_width = 0.25, fill_alpha = 1)#Specify figure layout.
    p.add_layout(color_bar, 'below')#Display figure inline in Jupyter Notebook.
    output_notebook()#Display figure.
    return p

So it is located wrong in the map, where one can see the red circle:

introducir la descripción de la imagen aquí

Looks like the map is in EPSG:3857 ("web mercator") while my source is probably in EPSG:4326. How can I do to plot it correctly?

Here is the first few lines of my data:

    id  parent_id common_id common_name  has_children  shape_type_id  \
64  70140      69935         3        63-3         False              4   
65  70141      69935         2        63-2         False              4   
66  70142      69935         5        63-5         False              4   
67  70143      69935         6        63-6         False              4   
68  70144      69935         8        63-8         False              4   

   shape_type_name    value color  title_location results  \
64        Precinct  No Data  None  Precinct: 63-3   65.16   
65        Precinct  No Data  None  Precinct: 63-2   57.11   
66        Precinct  No Data  None  Precinct: 63-5   54.33   
67        Precinct  No Data  None  Precinct: 63-6   59.15   
68        Precinct  No Data  None  Precinct: 63-8   61.86   

                                             turnout  \
64  {'pct': 46.38, 'count': 686.0, 'eligible': 1479}   
65   {'pct': 49.62, 'count': 394.0, 'eligible': 794}   
66  {'pct': 58.26, 'count': 624.0, 'eligible': 1071}   
67   {'pct': 57.54, 'count': 492.0, 'eligible': 855}   
68   {'pct': 50.75, 'count': 506.0, 'eligible': 997}   

                                             geometry  
64  POLYGON ((42.18180 42.18530, 42.18135 42.18593...  
65  POLYGON ((42.20938 42.20621, 42.21156 42.20706...  
66  POLYGON ((42.08429 42.20468, 42.08489 42.20464...  
67  POLYGON ((42.16270 42.16510, 42.16661 42.16577...  
68  POLYGON ((42.16270 42.16510, 42.16315 42.16640...
Solution

  • You have to reproject your data from EPSG:4326 to EPSG:3857

    Here's a solution with some GeoJSON data:

    
# requirements
# !pip install pandas numpy bokeh geopandas

import pandas as pd
import numpy as np


def lon_to_web_mercator(lon):
    k = 6378137
    return lon * (k * np.pi / 180.0)


def lat_to_web_mercator(lat):
    k = 6378137
    return np.log(np.tan((90 + lat) * np.pi / 360.0)) * k


def wgs84_to_web_mercator(df, lon="lon", lat="lat"):
    """Converts decimal longitude/latitude to Web Mercator format"""
    k = 6378137
    df["x"] = df[lon] * (k * np.pi / 180.0)
    df["y"] = np.log(np.tan((90 + df[lat]) * np.pi / 360.0)) * k
    return df


BerlinWGS84 = [13.08835, 13.76116, 52.33826, 52.67551]

Berlin = x_range, y_range = ((lon_to_web_mercator(BerlinWGS84[0]), lon_to_web_mercator(BerlinWGS84[1])),
                             (lat_to_web_mercator(BerlinWGS84[2]), lat_to_web_mercator(BerlinWGS84[3])))


# plot it
from bokeh.plotting import figure, show, output_notebook
from bokeh.tile_providers import get_provider, Vendors
output_notebook()

tile_provider = get_provider(Vendors.CARTODBPOSITRON)


# range bounds sgupplied in web mercator coordinates
p = figure(x_range=x_range, y_range=y_range,
           x_axis_type="mercator", y_axis_type="mercator")
p.add_tile(tile_provider)

show(p)


# geopandas

import geopandas as gpd
import requests


def remoteGeoJSONToGDF(url, display=False):
    # source: https://medium.com/@maptastik/remote-geojson-to-geodataframe-19c3c1282a64
    """Import remote GeoJSON to a GeoDataFrame
    Keyword arguments:
    url -- URL to GeoJSON resource on web
    display -- Displays geometries upon loading (default: False)
    """
    r = requests.get(url)
    data = r.json()
    gdf = gpd.GeoDataFrame.from_features(data['features'])
    if display:
        gdf.plot()
    return gdf


url = 'https://gist.githubusercontent.com/sabman/96730f5949576e7793a3f79eb390f90c/raw/7ffcf34239175cafcc9a63382e6beacd0cab9fa9/BerlinFeatures.geojson'
gdf = remoteGeoJSONToGDF(url)

gdf.plot()

# make sure initial projection is defined
gdf.crs = {'init': 'epsg:4326'}
gdf_webmerc = gdf.copy()
# reproject
gdf_webmerc = gdf['geometry'].to_crs(epsg=3857)
gdf_webmerc.plot()

from bokeh.models import GeoJSONDataSource
geo_source = GeoJSONDataSource(geojson=gdf_webmerc.to_json())

# let's plot and look
p.circle(x='x', y='y', size=15, alpha=0.7, source=geo_source)
show(p)

    enter image description here