Plotting Maps using Bokeh [Scatter Maps, Connection Maps & Choropleth Maps]¶
Bokeh has been the go-to library for many python data scientists for visualization purposes. Bokeh provides good support for handling and visualizing geospatial data. We have already covered the basics of bokeh in other tutorials and will be covering about plotting interactive maps using bokeh in this tutorial. Bokeh provides an easy interface to access various map tiles from tile providers which provide different types of maps which can be used as a base plot on which other scatter plot, connection lines can be plotted. We'll be primarily plotting three different maps using bokeh in this tutorial.
- Scatter Maps
- Connection Maps
- Choropleth Maps
We'll be using different datasets and aggregate their data in various ways to plot different types of map charts mentioned above. We'll start by loading important libraries.
import pandas as pd
import numpy as np
from bokeh.plotting import figure
from bokeh.tile_providers import get_provider, WIKIMEDIA, CARTODBPOSITRON, STAMEN_TERRAIN, STAMEN_TONER, ESRI_IMAGERY, OSM
from bokeh.io import output_notebook, show
import warnings
warnings.filterwarnings("ignore")
pd.set_option("max_columns", 30)
output_notebook()
![Plotting Maps using Bokeh [Scatter Maps, Connection Maps & Choropleth Maps]](https://storage.googleapis.com/coderzcolumn/static/tutorials/data_science/bokeh_2_1_1_loading.jpg)
- Flights in Brazil - It has information about various flights from and to Brazil.
- US States Population - It has information about US states population in 2018.
- Starbucks store locations worldwide - It has information about Starbucks store locations across the world.
We suggest that you download the datasets and follow along with us using the same datasets to better understand the material and get an in-depth idea about the whole process. We'll first load the datasets as pandas dataframe and aggregate it in various ways to create different chord diagrams.
Please make a note that original brazil flights dataset is quite big with size of 600+ MB and nearly 2.5 Mn rows. If you have low RAM then you can use nrows attribute of read_csv() method to load only first few thousand entries to follow along with tutorial without getting stuck. We have loaded only 100k entries for making things simple for explanation purpose.
starbucks = pd.read_csv("datasets/starbucks_store_locations.csv")
starbucks.head()
brazil_flights = pd.read_csv("brazil_flights_data.csv", nrows=100000, encoding="latin")
brazil_flights = brazil_flights.rename(columns={"Cidade.Origem":"City_Orig", "Cidade.Destino":"City_Dest",
"Estado.Origem":"State_Orig", "Estado.Destino":"State_Dest",
"Pais.Origem":"Country_Orig", "Pais.Destino":"Country_Dest",
"Aeroporto.Origem":"Airport_Orig", "Aeroporto.Destino":"Airport_Dest"})
brazil_flights.head()
us_state_population = pd.read_csv("datasets/State Populations.csv")
us_state_population.columns = ["State", "Population"]
us_state_population.head()
The process of plotting a map using a bokeh consists of a few steps.
- We need to load tile from the list of tile providers available from
bokeh.tile_providers. - We need to create a figure with x and y-axis type as Mercator in order to load the chart as a geo chart.
- We need to add loaded tile to figure.
- We can then call other bokeh charting methods like a circle(), square(), line(), etc to plot other charts (glyphs) above maps.
Please make a note that bokeh only accepts web Mercator projection data hence we need to convert all our latitude and longitude projection data to web Mercator system. We'll be using pyproj library in order to map data from latitude/longitude projection to we Mercator projection.
Sample Map¶
Below we have created the first simple map which highlights a particular area of the world using WIKIMEDIA projection. It’s a tile provided by Wikipedia. We have simply followed the above steps without adding new glyphs on the chart. Please make a note that we have provided x and y range of charts as web Mercator projection data.
wikimedia = get_provider(WIKIMEDIA)
fig = figure(plot_width=800, plot_height=700,
x_range=(6000000, 8000000), y_range=(-1000000, 7000000),
x_axis_type="mercator", y_axis_type="mercator")
fig.add_tile(wikimedia)
show(fig)
![Plotting Maps using Bokeh [Scatter Maps, Connection Maps & Choropleth Maps]](https://storage.googleapis.com/coderzcolumn/static/tutorials/data_science/bokeh_maps_1.jpg)
World Map¶
Below we have plotted the world map using the CARTODBPOSITRON tile. We have first found of web Mercator projection for a world based on latitudes and longitudes of the world. We are then using that web Mercator projected data to set a range of x and y-axis.
from pyproj import Proj, transform
inProj = Proj(init='epsg:3857')
outProj = Proj(init='epsg:4326')
world_lon1, world_lat1 = transform(outProj,inProj,-180,-85)
world_lon2, world_lat2 = transform(outProj,inProj,180,85)
cartodb = get_provider(CARTODBPOSITRON)
fig = figure(plot_width=800, plot_height=700,
x_range=(world_lon1, world_lon2),
y_range=(world_lat1, world_lat2),
x_axis_type="mercator", y_axis_type="mercator")
fig.add_tile(cartodb)
show(fig)
![Plotting Maps using Bokeh [Scatter Maps, Connection Maps & Choropleth Maps]](https://storage.googleapis.com/coderzcolumn/static/tutorials/data_science/bokeh_maps_2.jpg)
Starbucks US Store Locations Scatter Map¶
Below we are creating our first scatter map using bokeh. We are the first filtering rows of Starbucks locations and keeping only rows where the country is the US. We are then converting store locations data present in latitude and longitude to web Mercator projection data. We are setting that mapped data into the pandas dataframe. We have then used that data for plotting circles on top of the US Map. We have used WIKIMEDIA projection this time.
We also have used tooltip which highlights latitude, longitude, state, and name of the store when someone hovers over any point in the scatter map.
%%time
starbucks_us = starbucks[starbucks.Country == "US"]
lons, lats = [], []
for lon, lat in list(zip(starbucks_us["Longitude"], starbucks_us["Latitude"])):
x, y = transform(outProj,inProj,lon,lat)
lons.append(x)
lats.append(y)
starbucks_us["MercatorX"] = lons
starbucks_us["MercatorY"] = lats
starbucks_us = starbucks_us.rename(columns={"Store Name":"Name", "State/Province":"State"})
starbucks_us.head()
wikimedia = get_provider(WIKIMEDIA)
us_lon1, us_lat1 = transform(outProj,inProj,-140,10)
us_lon2, us_lat2 = transform(outProj,inProj,-50,55)
p = figure(plot_width=900, plot_height=700,
x_range=(us_lon1, us_lon2), y_range=(us_lat1, us_lat2),
x_axis_type="mercator", y_axis_type="mercator",
tooltips=[
("State", "@State"), ("Name", "@Name"), ("(Long, Lat)", "(@Longitude, @Latitude)")
],
title="Starbucks Store Locations US")
p.add_tile(wikimedia)
p.circle(x="MercatorX", y="MercatorY",
size=2,
fill_color="dodgerblue", line_color="dodgerblue",
fill_alpha=0.3,
source=starbucks_us)
show(p)
![Plotting Maps using Bokeh [Scatter Maps, Connection Maps & Choropleth Maps]](https://storage.googleapis.com/coderzcolumn/static/tutorials/data_science/bokeh_maps_3.jpg)
Starbucks Stores Count Per State Scatter Map¶
We are now creating another scatter map which highlights Starbucks store count per various state of the world. We are the first aggregating data of Starbucks location by grouping it according to the state of each country and getting count per state. We are then taking an average of latitude and longitude of stores per state in order to get latitude and longitude of state. We are then converting latitude and longitude data to web Mercator projection data and adding it to the dataframe.
We have then used the final dataframe to plot a scatter chart on world maps. We have also used tooltip which highlights state name, store count per state, latitude, and longitude of state information when someone hovers over any point in the chart.
starbucks_cnt_per_state = starbucks.groupby(by="State/Province").count()[["Store Number"]]\
.rename(columns={"Store Number":"Count"})
starbucks_avg_lat_lon = starbucks.groupby(by="State/Province").mean()[["Longitude", "Latitude"]]
starbucks_cnt_per_state = starbucks_cnt_per_state.join(starbucks_avg_lat_lon)
starbucks_cnt_per_state = starbucks_cnt_per_state.reset_index().rename(columns={"State/Province":"State"})
starbucks_cnt_per_state.head()
lons, lats = [], []
for lon, lat in list(zip(starbucks_cnt_per_state["Longitude"], starbucks_cnt_per_state["Latitude"])):
x, y = transform(outProj,inProj,lon,lat)
lons.append(x)
lats.append(y)
starbucks_cnt_per_state["MercatorX"] = lons
starbucks_cnt_per_state["MercatorY"] = lats
p = figure(plot_width=800, plot_height=600,
x_range=(world_lon1, world_lon2), y_range=(world_lat1, world_lat2),
x_axis_type="mercator", y_axis_type="mercator",
tooltips=[
("State", "@State"), ("Count", "@Count"), ("(Long, Lat)", "(@Longitude, @Latitude)")
],
title="Starbucks Store Count Per State")
p.add_tile(cartodb)
p.circle(x="MercatorX", y="MercatorY",
size=4,
fill_color="tomato", line_color="tomato",
fill_alpha=0.3,
source=starbucks_cnt_per_state,
)
show(p)
![Plotting Maps using Bokeh [Scatter Maps, Connection Maps & Choropleth Maps]](https://storage.googleapis.com/coderzcolumn/static/tutorials/data_science/bokeh_maps_4.jpg)
Flights From Brazil To Other Countries Connection Map¶
The second type of chart that we'll be plotting using bokeh is a connection map. We'll follow the same steps as mentioned earlier but will use the Brazil flight dataset this time for explanation purposes.
The connection map that we'll plot using the Brazil flights dataset will show flights from brazil to all other countries along with their count when hovered over the endpoint of the edge. We'll first filter the brazil flight dataset to keep only rows where the source country is brazil. We'll then aggregate data to keep all combinations of flights from Brazil to other countries to get a count of flights from brazil to all other countries. We'll then convert source and destination latitude/longitude data to web Mercator projection and add it to the dataframe for later use.
We have used STAMEN_TONER and ESRI_IMAGERY tiles for this chart. We have used line and circle glyphs of bokeh to plot a line between the source and destination of flight and highlight source and destinations. We have also used tooltip which highlights the source country, the destination country, and a number of flights to that country.
flights_from_brazil = brazil_flights[brazil_flights.Country_Orig == 'Brasil']
flights_from_brazil_cnt = flights_from_brazil.groupby(by=["Country_Orig", "Country_Dest"]).count()[["Voos"]].rename(columns={"Voos":"Count"})
flights_from_brazil_lat_lon = flights_from_brazil.groupby(by=["Country_Orig", "Country_Dest"]).mean()[["LongDest","LatDest","LongOrig","LatOrig"]]
flights_from_brazil_cnt = flights_from_brazil_cnt.join(flights_from_brazil_lat_lon).reset_index()
flights_from_brazil_cnt.head()
cols = ["LongDest","LatDest","LongOrig","LatOrig"]
lines_x, lines_y = [],[]
lons, lats = [], []
for lon_dest, lat_dest, lon_orig, lat_orig in flights_from_brazil_cnt[cols].values:
lon_orig, lat_orig = transform(outProj,inProj,lon_orig,lat_orig)
lon_dest, lat_dest = transform(outProj,inProj,lon_dest,lat_dest)
lons.append(lon_dest)
lats.append(lat_dest)
lines_x.append([lon_orig, lon_dest])
lines_y.append([lat_orig, lat_dest])
flights_from_brazil_cnt["MercatorX"] = lons
flights_from_brazil_cnt["MercatorY"] = lats
flights_from_brazil_cnt.head()
stamen_toner = get_provider(STAMEN_TONER)
world_lon1, world_lat1 = transform(outProj,inProj,-150,-75)
world_lon2, world_lat2 = transform(outProj,inProj,140,75)
p = figure(plot_width=800, plot_height=700,
x_range=(world_lon1, world_lon2), y_range=(world_lat1, world_lat2),
x_axis_type="mercator", y_axis_type="mercator",
tooltips=[
("Source", "@Country_Orig"), ("Destination", "@Country_Dest"), ("Count", "@Count")
],
title="Starbucks Store Count Per State")
p.add_tile(stamen_toner)
p.multi_line(lines_x, lines_y, color="red")
p.circle(x="MercatorX", y="MercatorY",
size=10, alpha=0.8, color="lime",
source=flights_from_brazil_cnt)
show(p)
![Plotting Maps using Bokeh [Scatter Maps, Connection Maps & Choropleth Maps]](https://storage.googleapis.com/coderzcolumn/static/tutorials/data_science/bokeh_maps_5.jpg)
esri = get_provider(ESRI_IMAGERY)
world_lon1, world_lat1 = transform(outProj,inProj,-150,-75)
world_lon2, world_lat2 = transform(outProj,inProj,140,75)
p = figure(plot_width=800, plot_height=700,
x_range=(world_lon1, world_lon2), y_range=(world_lat1, world_lat2),
x_axis_type="mercator", y_axis_type="mercator",
tooltips=[
("Source", "@Country_Orig"), ("Destination", "@Country_Dest"), ("Count", "@Count")
],
title="Starbucks Store Count Per State")
p.add_tile(esri)
p.multi_line(lines_x, lines_y, color="lime")
p.circle(x="MercatorX", y="MercatorY",
size=10, alpha=0.9, color="tomato",
source=flights_from_brazil_cnt)
show(p)
![Plotting Maps using Bokeh [Scatter Maps, Connection Maps & Choropleth Maps]](https://storage.googleapis.com/coderzcolumn/static/tutorials/data_science/bokeh_maps_6.jpg)
United States Starbucks Store Count Per State Choropleth Map¶
The third chart type that we'll introduce using bokeh is a choropleth map. The process of plotting choropleth maps using bokeh is different from previous chart types. It does not need any tile providers or latitude/longitude information.
The first choropleth map that we'll plot using bokeh is Starbucks store count per US states. It'll color code states of the US according to the count of Starbucks stores in that state.
We'll be first loading US states boundary data which is available in bokeh as a part of bokeh.sampledata.us_states. We'll load it as a pandas dataframe so that we have each state’s boundary latitude and longitude data. We'll be using this list of latitude and longitudes to create a polygon consisting of the US state by using the patches() glyph of bokeh. We have also merged Starbucks store counts per state data with this boundary data so that we have all data available in a single dataframe.
We have then created single dictionary consisting of this data which will be used as a source of a choropleth map. We have used bokeh LogColorMapper which maps the count of store per state to a particular color in the selected color palette.
from bokeh.models import LogColorMapper
from bokeh.palettes import Oranges256 as oranges
from bokeh.sampledata.us_states import data as us_states
us_states_df = pd.DataFrame(us_states).T
us_states_df = us_states_df[~us_states_df["name"].isin(['Alaska', "Hawaii"])]
us_states_df["lons"] = us_states_df.lons.values.tolist()
us_states_df["lats"] = us_states_df.lats.values.tolist()
us_states_df = us_states_df.reset_index()
us_states_df = us_states_df.merge(starbucks_cnt_per_state[["State", "Count"]], how="left", left_on="index", right_on="State")
us_states_df.head()
us_states_datasource = {}
us_states_datasource["lons"] = us_states_df.lons.values.tolist()
us_states_datasource["lats"] = us_states_df.lats.values.tolist()
us_states_datasource["name"] = us_states_df.name.values.tolist()
us_states_datasource["StateCodes"] = us_states_df.index.values.tolist()
us_states_datasource["StarbuckStores"] = us_states_df.Count.values.tolist()
fig = figure(plot_width=900, plot_height=600,
title="United States Starbucks Store Count Per State Choropleth Map",
x_axis_location=None, y_axis_location=None,
tooltips=[
("Name", "@name"), ("StarbuckStores", "@StarbuckStores"), ("(Long, Lat)", "($x, $y)")
])
fig.grid.grid_line_color = None
fig.patches("lons", "lats", source=us_states_datasource,
fill_color={'field': 'StarbuckStores', 'transform': LogColorMapper(palette=oranges[::-1])},
fill_alpha=0.7, line_color="white", line_width=0.5)
show(fig)
![Plotting Maps using Bokeh [Scatter Maps, Connection Maps & Choropleth Maps]](https://storage.googleapis.com/coderzcolumn/static/tutorials/data_science/bokeh_maps_7.jpg)
The United States Population Per State [2018] Choropleth Map¶
The second choropleth map that we have created using bokeh is the US state population choropleth map. We have merged the US state population dataframe loaded at the beginning with the dataframe of the previous step and also added population count to the data source dictionary created in the previous step. We have used the same data source dictionary which we have used in the previous step with this data added to plot choropleth map.
us_states_df = us_states_df.merge(us_state_population, how="left", left_on="name", right_on="State")
us_states_df.head()
us_states_datasource["Population"] = us_states_df.Population.values.tolist()
from bokeh.palettes import YlGnBu9 as YlGnBu
fig = figure(plot_width=900, plot_height=600,
title="United States Population Per State [2018] Choropleth Map",
x_axis_location=None, y_axis_location=None,
tooltips=[
("Name", "@name"), ("Population", "@Population"), ("(Long, Lat)", "($x, $y)")
])
fig.grid.grid_line_color = None
fig.patches("lons", "lats", source=us_states_datasource,
fill_color={'field': 'Population', 'transform': LogColorMapper(palette=YlGnBu[::-1])},
fill_alpha=0.7, line_color="white", line_width=0.5)
show(fig)
![Plotting Maps using Bokeh [Scatter Maps, Connection Maps & Choropleth Maps]](https://storage.googleapis.com/coderzcolumn/static/tutorials/data_science/bokeh_maps_8.jpg)
This ends our small tutorial on creating a scatter map, connection map, and choropleth map using bokeh. Please feel free to let us know your views in the comments section.
References¶
- Plotting Static Maps with Geopandas[Working with Geospatial Data]
- How to Convert Static Maps (Geopandas) to Interactive Maps (hvplot)?
- How to Create a Connection Map in Python [Plotly & Geopandas]?
- Cartopy - Basic Maps [Scatter Map, Bubble Map & Connection Map]
- Folium - Interactive Maps
- bqplot - Interactive Plotting Python
- ipyleaflet - Interactive Maps in Python based on leafletjs
- Interactive Choropleth Maps using bqplot
- Choropleth Maps using ipyleaflet
- Interactive Choropleth Maps using Cufflinks
Sunny Solanki
Sunny Solanki
Intro: Software Developer | Youtuber | Bonsai Enthusiast
About: Sunny Solanki holds a bachelor's degree in Information Technology (2006-2010) from L.D. College of Engineering. Post completion of his graduation, he has 8.5+ years of experience (2011-2019) in the IT Industry (TCS). His IT experience involves working on Python & Java Projects with US/Canada banking clients. Since 2020, he’s primarily concentrating on growing CoderzColumn.
His main areas of interest are AI, Machine Learning, Data Visualization, and Concurrent Programming. He has good hands-on with Python and its ecosystem libraries.
Apart from his tech life, he prefers reading biographies and autobiographies. And yes, he spends his leisure time taking care of his plants and a few pre-Bonsai trees.
Contact: sunny.2309@yahoo.in
![YouTube Subscribe]()
Comfortable Learning through Video Tutorials?
If you are more comfortable learning through video tutorials then we would recommend that you subscribe to our YouTube channel.
![Need Help]()
Stuck Somewhere? Need Help with Coding? Have Doubts About the Topic/Code?
Stuck Somewhere? Need Help with Coding? Have Doubts About the Topic/Code?When going through coding examples, it's quite common to have doubts and errors.
If you have doubts about some code examples or are stuck somewhere when trying our code, send us an email at coderzcolumn07@gmail.com. We'll help you or point you in the direction where you can find a solution to your problem.
You can even send us a mail if you are trying something new and need guidance regarding coding. We'll try to respond as soon as possible.
![Share Views]()
Want to Share Your Views? Have Any Suggestions?
Want to Share Your Views? Have Any Suggestions?If you want to
- provide some suggestions on topic
- share your views
- include some details in tutorial
- suggest some new topics on which we should create tutorials/blogs
bokeh, maps
Subscribe to Our YouTube Channel
