In [1]:
# import the classes
from chorogrid import Colorbin, Chorogrid
# read the docs
help(Chorogrid)
We'll load some sample data that reports the number of people living in the same home as one year ago per state in the United States, and use Colorbin to associate these numbers to colors. See Tutorial 1 for information on how Colorbin works.
In [2]:
mycolors = ['#b35806', '#f1a340', '#fee0b6', '#d8daeb', '#998ec3', '#542788']
import pandas as pd
df = pd.read_csv('chorogrid/sample_data/sample_state_data.csv')
mybin = Colorbin(df['Percent_living_in_same_home_as_one_year_ago'], mycolors, proportional=True, decimals=None)
mybin.set_decimals(1)
mybin.recalc(fenceposts=True)
mybin.calc_complements(0.5, '#e0e0e0', '#101010')
Here's the tail of the dataframe.
In [3]:
df.tail()
Out[3]:
Now here's a look at all the objects we'll use to make further maps, lists that are either the length of the number of observations (51) or the number of colors (6).
In [4]:
states = list(df.state)
colors_by_state = mybin.colors_out
font_colors_by_state = mybin.complements
legend_colors = mybin.colors_in
legend_labels = mybin.labels
for lst in ['states', 'colors_by_state', 'font_colors_by_state', 'legend_colors', 'legend_labels']:
obj = eval(lst)
print("{:>20}: len {:2}: {}...".format(lst, len(obj), obj[:3]))
Here is the database within Chorogrid for U.S. states, containing identifying information and instructions for hex, square, multihex and map choropleths.
In [5]:
_ = pd.read_csv('chorogrid/databases/usa_states.csv')
print(_.iloc[0])
And here's a help file with descriptions of all the columns in the cell above.
In [6]:
with open('chorogrid/databases/usa_states_column_descriptions.txt') as f:
print(f.read())
In [7]:
cg = Chorogrid('chorogrid/databases/usa_states.csv', states, colors_by_state)
cg.set_title('% Living at same address as one year ago', font_dict={'font-size': 19})
cg.set_legend(legend_colors, legend_labels, title='% of population')
cg.draw_squares(spacing_dict={'margin_right': 150}) # otherwise legend will be cut off
# another strategy would be to pass a legend_offset to spacing_dict
cg.done(show=True)
Note that it's very difficult to see the text in the darkest-colored states. Luckily we've create a list of font colors based on Colorbin's complement method. Let's rerun the last two lines of the cell above, but with font colors specified.
In [8]:
cg.draw_squares(spacing_dict={'margin_right': 150}, font_colors=font_colors_by_state)
cg.done(show=True)
Here's an alternate layout of squares.
In [9]:
cg.draw_squares(x_column='altsquare_x', y_column='altsquare_y', spacing_dict={'margin_right': 150},
font_colors=font_colors_by_state)
cg.done(show=True)
And here's a hex layout.
In [10]:
cg.draw_hex(spacing_dict={'margin_right': 150}, font_colors=font_colors_by_state)
cg.done(show=True)
And an alternate hex layout
In [11]:
cg.draw_hex(x_column='althex_x', y_column='althex_y', spacing_dict={'margin_right': 150},
font_colors=font_colors_by_state)
cg.done(show=True)
And a traditional choropleth map.
In [12]:
cg = Chorogrid('chorogrid/databases/usa_states.csv', states, colors_by_state)
cg.draw_map(spacing_dict={'legend_offset': [-150,-25]})
cg.done(show=True)
# To Do: Add state names. The required data is there in the database, but hasn't been implemented in code.
And a fancy one where states are represented by four hexes each.
In [13]:
cg.draw_multihex()
cg.done(show=True)
There are currently three databases in Chorogrid: USA by state, as seen above, USA by county, which we'll look at ehre, and finally Europe by country, which will be shown below. FIPS (Federal Information Processing Codes) standards are used to identify counties.
In [14]:
df = pd.read_csv('chorogrid/sample_data/sample_county_data.csv', encoding='latin-1')
Let's have a look at the data. There are 3143 counties (including the 0-indexed one): counties are redefined every few years, this is the most recent count as of June 2015. If you have older data or if the borders change after that date, the mapping between data and map may not be perfect.
In [15]:
df.tail()
Out[15]:
Use colorbin to divide the quantities into six bins
In [16]:
mybin = Colorbin(df['Median_value_of_owner-occupied_housing_units_2009-2013'], mycolors, proportional=False, decimals=None)
mybin.fenceposts
Out[16]:
Reset the fenceposts in order to have whole numbers
In [17]:
mybin.fenceposts = [0, 50000, 100000, 150000, 250000, 500000, 1000000]
mybin.recalc(False)
mybin.labels
Out[17]:
Here are the columns for the counties database. Note that it only has maps, not squares or hexes or multihexes (for now, anyway).
In [18]:
with open('chorogrid/databases/usa_counties_column_descriptions.txt', 'r', encoding='utf-8') as f:
print(f.read())
And here's the map.
In [19]:
cg = Chorogrid('chorogrid/databases/usa_counties.csv', df.fips, mybin.colors_out, 'fips_integer')
cg.set_title('Median value of owner-occupied housing units, 2009-2013', font_dict={'font-size': 19})
cg.set_legend(mybin.colors_in, mybin.labels, title='US dollars')
cg.draw_map(spacing_dict={'legend_offset':[-300,-200], 'margin_top': 50}) # otherwise legend will be cut off
# another strategy would be to pass a legend_offset to spacing_dict
cg.done(show=True, save_filename='eraseme')
Note that there are no state borders in the above graph; we can add it by adding the statelines.txt file, which has the appropriate path descriptions.
In [20]:
with open('chorogrid/databases/usa_counties_statelines.txt', 'r') as f:
statelines = f.read()
cg.add_svg(statelines)
cg.done(show=True, save_filename='eraseme')
### Known bug: if you run this cell again, the borders will be offset.
### For now, you'll have to restart the kernel to get the state borders in the right place.
And now Europe. Our source data has two- and three-letter county abbreviations; for now, the database only has two-letter abbreviations.
In [21]:
with open('chorogrid/databases/europe_countries_column_descriptions.txt', 'r', encoding='utf-8') as f:
print(f.read())
In [22]:
df = pd.read_csv('chorogrid/sample_data/sample_europe_data.csv', encoding='latin-1')
df.tail()
Out[22]:
And here's the choropleth.
In [23]:
mybin = Colorbin(df['Pct Internet users'], mycolors, proportional=False, decimals=None)
mybin.fenceposts = [0, 40, 60, 70, 80, 90, 100]
mybin.recalc(False)
font_colors_europe = mybin.complements
cg = Chorogrid('chorogrid/databases/europe_countries.csv', df.abbrev2, mybin.colors_out, 'abbrev')
cg.set_title('% Internet users', font_dict={'font-size': 19})
cg.set_legend(mybin.colors_in, mybin.labels, title='% of population')
cg.draw_map(spacing_dict={'legend_offset':[-200,-100], 'margin_top': 50}) # otherwise legend will be cut off
# another strategy would be to pass a legend_offset to spacing_dict
cg.done(show=True)
Note that in this case there was not a perfect mapping of country abbreviations between the source data and the database. Some small 'countries' like the Faroe Islands [FO] (which is part of the Danish Realm but is self-governing) are not in the database, while several countries including Russia [RU] are not in the data.
Finally, here's the hex grid:
In [24]:
cg.draw_hex(spacing_dict={'legend_offset':[0, -100], 'margin_top': 50, 'margin_right': 200})
cg.done(show=True)
# Known issue: font_colors does not work for this Europe hex map.
In [ ]: