def create_graph_from_shapefile(shapefile_path='./data/wes_with_districtings/wes_with_districtings.shp'):
log.info(f"Creating the queen adjacency graph")
pa_queen = Graph.from_shapefile(
shapefile_path, adjacency_type='queen', data_columns=None, id_column='wes_id')
log.info(f"Saving the queen adjacency graph")
pa_queen.add_columns_from_shapefile(
shapefile_path, columns, id_column='wes_id')
# Fix nodes whose assignment is different from all their neighbors, and
# whose neighbors all have the same assignment.
# In practice, this affects one node's assignment in two plans. (Node '1')
for plan in plans:
correct_islands(pa_queen.graph, plan)
pa_queen.save('./PA_queen.json')
log.info(f"Creating the rook adjacency graph")
pa_rook = Graph.from_shapefile(
shapefile_path, adjacency_type='rook', data_columns=None, id_column='wes_id')
pa_rook.add_columns_from_shapefile(
shapefile_path, columns, id_column='wes_id')
log.info(f"Saving the queen adjacency graph")
for plan in plans:
correct_islands(pa_rook.graph, plan)
pa_rook.save('./PA_rook.json')
def main(shapefile_path='./wes_unitsPA/wes_units_PA.shp',
graph_path='wes_graph.json'):
pa = Graph.load(graph_path)
shape = geopandas.read_file(shapefile_path)
shape = shape.set_index('wes_id')
for plan in plans:
log.info(f"Plan: {plan}")
try:
plot_plan(shape, plan, f"./plots/{plan}_before.svg")
except Exception:
log.error("There was an error while trying to plot", exc_info=1)
correct_islands(pa.graph, plan)
assignment = {
node: pa.graph.nodes[node][plan]
for node in pa.graph.nodes
}
try:
plot_plan(shape, assignment, f"./plots/{plan}_after.svg")
except Exception:
log.error("There was an error while trying to plot", exc_info=1)
pa.save()
def set_up_plan(plan):
graph = Graph.load('./PA_queen.json').graph
assignment = {node: graph.nodes[node][plan] for node in graph.nodes}
updaters = {
'perimeters': perimeters,
'exterior_boundaries': exterior_boundaries,
'interior_boundaries': interior_boundaries,
'boundary_nodes': boundary_nodes,
'cut_edges': cut_edges,
'areas': Tally('area', alias='areas'),
'polsby_popper': polsby_popper,
'cut_edges_by_part': cut_edges_by_part
}
return Partition(graph, assignment, updaters)
def match_wes_units_to_remedial_plan():
pa = Graph.load('./wes_graph.json')
blocks = pandas.read_csv('./data/block_to_wesid.csv', dtype=str)
plan = pandas.read_csv('./data/remedial/blocks_to_remedial.csv',
names=['GEOID10', 'remedial'],
dtype=str)
blocks = blocks.set_index('GEOID10')
plan = plan.set_index('GEOID10')
blocks['remedial'] = plan['remedial']
log.info('Mapping units to parts')
mapping = map_units_to_parts_via_blocks(blocks,
pa.graph,
unit='wes_id',
part='remedial')
for node in mapping:
if node != '0':
pa.graph.nodes[node]['remedial'] = mapping[node]
pa.save('./wes_graph2.json')
log.info('Getting block populations')
block_pops = geopandas.read_file(
'../graphmaker/graphmaker/blocks/42/tabblock2010_42_pophu.shp')
block_pops['GEOID10'] = block_pops['BLOCKID10'].astype('object')
block_pops = block_pops.set_index('GEOID10')
blocks['population'] = block_pops['POP10'].astype(int)
log.info('Creating report')
report = splitting_report(blocks, 'wes_id', 'remedial')
with open('./reports/splitting_energy_remedial_plan.json', 'w') as f:
json.dump(report, f)
return report
def set_up_chain(plan, total_steps, adjacency_type='queen'):
graph = Graph.load(f"./PA_{adjacency_type}.json").graph
assignment = {node: graph.nodes[node][plan] for node in graph.nodes}
updaters = {
**votes_updaters(elections["2016_Presidential"],
election_name="2016_Presidential"),
**votes_updaters(elections["2016_Senate"], election_name="2016_Senate"), 'population':
Tally('population', alias='population'),
'perimeters':
perimeters,
'exterior_boundaries':
exterior_boundaries,
'interior_boundaries':
interior_boundaries,
'boundary_nodes':
boundary_nodes,
'cut_edges':
cut_edges,
'areas':
Tally('area', alias='areas'),
'polsby_popper':
polsby_popper,
'cut_edges_by_part':
cut_edges_by_part
}
partition = Partition(graph, assignment, updaters)
population_constraint = within_percent_of_ideal_population(partition, 0.01)
compactness_constraint = SelfConfiguringLowerBound(L_minus_1_polsby_popper,
epsilon=0.1)
is_valid = Validator(default_constraints +
[population_constraint, compactness_constraint])
return partition, MarkovChain(propose_random_flip, is_valid, always_accept,
partition, total_steps)
# As a replacement, we can use Census tracts instead.
# The same technique can be used to make a graph of counties (e.g. for Iowa)
# or any other shapefile you want.
# First we'll download the tracts:
# The Tiger class lets us access Census shapefile URLs as if they were
# just python objects.
kentucky_tracts = Tiger(2012).tract.ky
# or, equivalently (fips)
kentucky_tracts = Tiger(2012).tract['21']
# or, equivalently (full name)
kentucky_tracts = Tiger(2012).tract.kentucky
# To download from the url, call download() and pass the directory
# you want to save the shapefiles in. (Make sure the directory exists.)
kentucky_tracts.download(target='./kentucky/')
# The shapefile will be named 'tl_2012_21_tract.shp
kentucky_queen = Graph.from_shapefile('./kentucky/tl_2012_21_tract.shp',
adjacency_type='queen')
# ...This might take a while...
# Once it's done, you can view some statistics about the graph like this:
print(graph_report(kentucky_queen.graph))
# And then save the graph wherever you want:
kentucky_queen.save('./kentucky/queen.json')
from graphmaker.graph import Graph
from graphmaker.match import match
# Load the state adjacency graph from wherever you downloaded it to
my_state = Graph.load(
'../graphmaker/graphs/vtd-adjacency-graphs/vtd-adjacency-graphs/12/queen.json'
)
# Match the VTDs to 'State Legislature Upper Chamber' Districts
match(my_state, 'VTD', 'SLDU')
# Print all the nodes, to see that the 'SLDU' attribute has been added
print(my_state.graph.nodes(data=True))
# Save the graph back to the same file
my_state.save()
from graphmaker.resources import BlockAssignmentFile
from graphmaker.graph import Graph
from graphmaker.integrate import integrate
# You can also add columns from a Pandas dataframe, which lets you do whatever joining,
# transformation, or merging that you need to do before adding it to the graph.
# A good way to do this is with our `integrate` function to take Census block-level
# statistics (like population) and add them to the VTD graph.
# First we need to download Block Assignment Files. These match census blocks to bigger
# subdivisions (like VTDs) so that we know how to add up the block-level data.
# We'll use Florida:
blocks = BlockAssignmentFile('12').download(target='./florida/blocks/')
# This actually downloads a whole set of files, with matchings from blocks to multiple
# different types of subdivisions, including Upper and Lower State Legislatures and
# elementary school districts.
# Now let's load our graph:
florida = Graph.load('./florida/queen.json')
df = integrate('./florida/blocks/BlockAssign_ST12_FL_VTD', ['POP10'], 'VTD')
florida.add_columns_from_df(df, columns=['POP10'])
linewidth=0.5,
edgecolor='0.5',
column='assignment',
categorical=True,
cmap='tab20')
ax.set_axis_off()
plt.axis('equal')
plt.savefig(filepath)
if __name__ == '__main__':
shapefile_path = './data/wes_unitsPA/wes_units_PA.shp'
shape = geopandas.read_file(shapefile_path)
pa = Graph.load('./wes_graph2.json')
for node in pa.graph.nodes:
if 'remedial' not in pa.graph.nodes[node]:
print(node)
print(
len([
node for node in pa.graph.nodes
if 'remedial' not in pa.graph.nodes[node]
]))
assignment = {
node: pa.graph.nodes[node]["remedial"]
for node in pa.graph.nodes
}
from graphmaker.graph import Graph
# We'll continue from the Kentucky example:
# Load your graph from wherever you saved it:
kentucky_queen = Graph.load('./kentucky/queen.json')
# You can add columns from a shapefile:
kentucky_queen.add_columns_from_shapefile(
'./kentucky/tl_2012_21_tract.json',
columns=['ALAND', 'AWATER', 'COUNTYFP', 'STATEFP'])
# Or from a CSV (using a made up example):
kentucky_queen.add_columns_from_csv('./votes.csv',
columns=['D_VOTES_2020', 'R_VOTES_2020'],
id_column='TRACT')
# By writing id_column='VTD', we note that the column with IDs that match the nodes of the
# graph (the tract GEOIDs) is 'TRACT'. Usually this won't be necessary, because the program
# will automatically recognize common names like 'GEOID' or 'ID'.