July 30, 2004 update: Changed Pareto by state to only include US states, added ‘Others’ as a region to group the smallest entries, with color value of gray. Here is the code:
'''
This script produces:
- choropleth map of Rural investment by US county
- Pareto showing investment amounts grouped by state or territory
- Pareto with investment amounts grouped by region.
hover info:
colorpleth map: county name, state abbreviation, investment amount
pareto by state: state name, investment amount, cumulative %
pareto by region: region name, investment amount, cumulative %
pareto by state filtered to only include 50 US states
pareto by region merged smallest invested regions to group called others
'''
import polars as pl # dataframe library
import us # library with USA state info(name, abbr, timezone, etc.)
import plotly.express as px
from plotly.subplots import make_subplots
#------------------------------------------------------------------------------#
# Functions #
#------------------------------------------------------------------------------#
def add_annotation(fig, annotation, x, y, align, xanchor, yanchor):
fig.add_annotation(
text=annotation,
showarrow=False,
xref = 'paper', x=x, yref = 'paper', y=y,
align= align, xanchor=xanchor, yanchor=yanchor,
font = {'size': 14, 'color': 'darkslategray'}
)
return fig
def update_layout(
fig, my_title, my_height, my_width, # mandatory parameters
my_xtitle='', my_ytitle='', my_legend_title='', # optional parameters
my_showlegend=False
):
fig.update_layout(
title = my_title,
xaxis_title=my_xtitle,
yaxis_title=my_ytitle,
legend_title=my_legend_title,
height=my_height, width=my_width,
margin={"r":50, "t":50, "l":50, "b":50},
autosize=False,
showlegend=my_showlegend
)
return fig
#------------------------------------------------------------------------------#
# Map time zones to states & territoriess, assign region names #
#------------------------------------------------------------------------------#
region_dict = {}
for key in us.states.mapping('name', 'time_zones'):
value = (
us.states.mapping('name', 'time_zones')[key][0]
.replace('America/New_York', 'East')
.replace('America/Chicago', 'Central')
.replace('America/Denver', 'Mountain')
.replace('America/Los_Angeles', 'Pacific')
.replace('America/Anchorage', 'Alaska')
.replace('Pacific/Honolulu', 'Hawaii')
.replace('America/Phoenix', 'Pacific')
.replace('America/Boise', 'Central') #for N. Dak, not Idaho
.replace('America/Puerto_Rico', 'Puerto Rico')
)
region_dict[key] = value
#------------------------------------------------------------------------------#
# Load csv file, process rows and columns #
#------------------------------------------------------------------------------#
file_path = (
'https://raw.githubusercontent.com/plotly/Figure-Friday/' +
'main/2024/week-30/rural-investments.csv'
)
investment_data = (
pl.read_csv(file_path, ignore_errors = True)
.rename({'State Name': 'state_name'})
.with_columns(
pl.col('County FIPS').str.replace("'", "").str.zfill(5),
pl.col('Investment Dollars').str.replace_all(',', '').cast(pl.Float64),
state_abbr = pl.col('state_name')
.replace(us.states.mapping('name', 'abbr')),
region = pl.col('state_name').replace(region_dict)
)
.select(
'state_name', 'state_abbr', 'County',
'County FIPS', 'region','Investment Dollars'
)
)
investment_data = (
investment_data
.with_columns(
region = pl.when(pl.col('region')
.is_in(
['Puerto Rico', 'Pacific', 'Central', 'Hawaii', 'Mountain', 'Alaska', 'East']))
.then('region')
.otherwise(pl.lit('Others'))
)
)
#------------------------------------------------------------------------------#
# Create choropleth map #
#------------------------------------------------------------------------------#
grouped_by_county = (
investment_data.group_by('state_abbr', 'County', 'County FIPS').sum()
)
geojson_path = 'https://raw.githubusercontent.com/plotly/datasets/master'
geojson_file = 'geojson-counties-fips.json'
fig = px.choropleth(
grouped_by_county,
geojson=f'{geojson_path}/{geojson_file}',
locations='County FIPS',
color='Investment Dollars',
color_continuous_scale="Viridis",
scope="usa",
title='USA Rural Investment by County - 2024',
custom_data=['County', 'state_abbr','Investment Dollars']
)
# hover info indented to keep mouse point icon from blocking hover display
fig.update_traces(
hovertemplate="<br>".join([
' %{customdata[0]} County, %{customdata[1]}',
' $%{customdata[2]:,d}',
])
)
fig = update_layout(fig, '2024 Rural Investment by USA County', 600, 1200)
annotation = '<b>Data Source:</b> US Department of Agriculture<br><br>'
add_annotation(fig, annotation, 0.8, 1.0, 'left', 'right', 'top')
fig.show()
#------------------------------------------------------------------------------#
# Create colormap dictionaries for both pareto charts #
#------------------------------------------------------------------------------#
# keys are regions, values are colors. for the 'other' region, hardcode gray
my_colors = px.colors.qualitative.Dark24
my_regions = sorted(investment_data['region'].unique().to_list())
my_color_dict = dict(zip(my_regions, my_colors))
my_color_dict['Others'] = 'gray'
#------------------------------------------------------------------------------#
# Pareto: data grouped by USA state #
#------------------------------------------------------------------------------#
grouped_by_state = (
investment_data
.select('state_name', 'state_abbr', 'Investment Dollars', 'region')
.group_by('state_name', 'state_abbr', 'region').sum()
.sort('Investment Dollars', descending=True)
.with_columns(
CUM_PCT =
(100 * pl.col('Investment Dollars')/pl.col('Investment Dollars').sum())
.cum_sum()
)
.filter(pl.col('state_name').is_in([str(s) for s in list(us.states.STATES)]))
)
fig = px.bar(
grouped_by_state.sort('Investment Dollars', descending=True),
x = 'state_abbr',
y = 'Investment Dollars',
template='plotly_white',
color='region',
color_discrete_map=my_color_dict,
custom_data=['state_name','Investment Dollars', 'CUM_PCT']
)
# Set y-axes titles
fig.update_yaxes(title_text="<b>primary</b> Investment Dollars", secondary_y=False)
fig.update_yaxes(title_text="<b>secondary</b> Cumulative PCT", secondary_y=True)
# add custom hover information
fig.update_traces(
hovertemplate="<br>".join([
'%{customdata[0]}',
'$%{customdata[1]:,d}',
'Cumulative: %{customdata[2]:.1f}%',
'<extra></extra>'
])
)
fig = update_layout(
fig, 'USA Rural Investment by States & Territories - 2024', 600, 1200,
my_xtitle= 'State - Abbreviated',
my_ytitle= 'Investment (US$)',
my_legend_title='Region',
my_showlegend=True
)
fig.update_xaxes(showgrid=False)
fig.update_yaxes(showgrid=False)
annotation = '<b>Data Source:</b> US Department of Agriculture<br>'
annotation += 'states & territories are color coded by region<br>'
add_annotation(fig, annotation, 0.5, 0.8, 'left', 'right', 'top')
# chart gets reordered after applying colors, next line restores intended order
fig.update_xaxes(categoryorder='total descending')
fig.show()
#------------------------------------------------------------------------------#
# Pareto: data grouped by region #
#------------------------------------------------------------------------------#
grouped_by_region = (
investment_data
.select('region', 'Investment Dollars')
.group_by('region').sum()
.sort('Investment Dollars', descending=True)
.with_columns(
CUM_PCT =
(100 * pl.col('Investment Dollars')/pl.col('Investment Dollars').sum())
.cum_sum()
)
)
fig = px.bar(
grouped_by_region.sort('Investment Dollars', descending=True),
'region',
y = 'Investment Dollars',
template='plotly_white',
color='region',
color_discrete_map=my_color_dict,
custom_data=['region','Investment Dollars', 'CUM_PCT'],
)
# add custom hover information
fig.update_traces(
hovertemplate="<br>".join([
'%{customdata[0]}',
'$%{customdata[1]:,d}',
'Cumulative: %{customdata[2]:.1f}%',
'<extra></extra>'
])
)
fig = update_layout(
fig, 'USA Rural Investment by Region - 2024', 600, 1200,
my_xtitle='Region',
my_ytitle='Investment (US$)',
)
fig.update_xaxes(showgrid=False)
fig.update_yaxes(showgrid=False)
annotation = '<b>Data Source:</b> US Department of Agriculture<br>'
annotation += 'Regions are appoximations of time zones. Exceptions<br>'
annotation += 'made for states with multiple time zones and for<br>'
annotation += 'states that do not observe daylight saving time.<br><br>'
annotation += "<b>'Others'</b> incudes Virgin Islands, Guam, Samoa & Palua"
add_annotation(fig, annotation, 0.9, 0.8, 'left', 'right', 'top')
fig.show()
If you prefer to collaborate with others on Discord, join the Plotly Discord channel .
