Commit 87b7ac85 authored by Jonathan Mattingly's avatar Jonathan Mattingly
Browse files

initial set up

parents

Atlas.py

0 → 100644
+84 −0
Original line number Diff line number Diff line 
import json

import gzip

import os





class Atlas:

    description = ""

    date = ""

    atlasParamType = ""

    mapParamType = ""

    atlasParam = dict()

    fp = None



    def __repr__(self):

        return "Atlas()"

    

    def __str__(self):

        return "atlas\ndescription: " + self.description + "\ndate: " + \

               self.date + "\natlasParam: " + str(self.atlasParam)





class Map:

    name = ""

    weight = 1

    data = dict()

    districting = dict()

    

    def __repr__(self):

        return "Map()"

    

    def __str__(self):

        return "map\nname: " + self.name + "\nweight: " + str(self.weight) + \

               "\ndistricting: " + str(self.districting)





def openAtlas(fileName):

    s_name = os.path.splitext(fileName)

    if s_name[1] == '.gz':

        fp = gzip.open(fileName, "r")

    else:

        fp = open(fileName, "r")

    

    atlas = Atlas()

    atlas.fp = fp

    line = fp.readline()  # drop first line

    line = fp.readline() 

    atlasHeader = json.loads(line)

    atlas.description = atlasHeader['description']

    atlas.date = atlasHeader["date"]

    atlas.atlasParamType = atlasHeader["atlasParamType"]

    atlas.mapParamType = atlasHeader["mapParamType"]

    line = fp.readline() 

    atlas.atlasParam = json.loads(line)



    return atlas





def closeAtlas(atlas):

    atlas.fp.close()





def nextMap(atlas):

    line = atlas.fp.readline()

    if not line:

        return None

    exp = json.loads(line)



    map = Map()

    map.name = exp["name"]

    map.weight = exp["weight"]

    map.districting = {}

    for d in exp["districting"]:

        for k, v in d.items():

            map.districting[k] = v

    map.data = exp["data"]

    

    return map

example_atlas.py

0 → 100644
+18 −0
Original line number Diff line number Diff line 
import Atlas



    

atlas=Atlas.openAtlas("./test.jsonl.gz")

print(atlas)

map=[]

print("\n")

while map!=None:

    map=Atlas.nextMap(atlas)

    print(map)

    print("\n")

example_atlas_multilevel.py

0 → 100644
+64 −0
Original line number Diff line number Diff line 
import Atlas

import helper_functions as hf

import json

import matplotlib.pyplot as plt

import pandas as pd

import seaborn as sns



pctDataF = open("pct21_20cen_wMCD.json")

pctData = json.load(pctDataF)

dataElection = pctData['nodes']



atlas = Atlas.openAtlas("./truncated_nc_multiscale.jsonl")

print(atlas)

map = []

print("\n")

electionName = "G16_USS"  # This is the name of the eleection we will consider.



pctToDistVotes = {}

for node in pctData["nodes"]:

    pctToDistVotes[node["id"]] = {}



while map is not None:  # This loops through all of the map in the atlas

    try:

        map = Atlas.nextMap(atlas)  # Get the next map in the atlas 

        print(map.name)

        # The maps are multi scale in the sense that if a county is kept whole

        # the following fuction makes a map from precicts to districts out of 

        # the multiscale assignement 

        node_to_dist = hf.get_node_to_district(map.districting, 

                                               pctData["nodes"]) 

        # The next fuction  sums up the election for this districting (defined 

        # by the map)

        distVoteR, distVoteD, distVoteT = hf.sumElection(electionName, 

                                                         node_to_dist, 

                                                         dataElection)

        for id in pctToDistVotes.keys():

            votes = {}

            votes["Dem"] = distVoteD[node_to_dist[id]]

            votes["Rep"] = distVoteD[node_to_dist[id]]

            votes["Total"] = distVoteT[node_to_dist[id]]

            pctToDistVotes[id][map.name] = votes

    

    except Exception:

        break





# now do somthing with data

id = 100  # Choose precinct

print(dataElection[id]["county"], dataElection[id]["prec_id"])



# or

county = "BEAUFORT"

prec_id = "BLCK"

id = [ii for ii in range(len(dataElection)) 

      if dataElection[ii]["county"] == county 

      and dataElection[ii]["prec_id"] == prec_id][0]

print(id)



df = pd.DataFrame(pctToDistVotes[id]).T

df["Dem %"] = df["Dem"]/df["Total"]

print(df)



sns.displot(data=df, x="Dem %", bins=10, stat="density")

plt.show()

example_atlas_multilevel_v2.py

0 → 100644
+73 −0
Original line number Diff line number Diff line 
import Atlas

import helper_functions as hf

import json

import matplotlib.pyplot as plt

import pandas as pd

import seaborn as sns









pctDataF = open("pct21_20cen_wMCD.json")

pctData = json.load(pctDataF)

dataElection = pctData['nodes']



atlas = Atlas.openAtlas("./truncated_nc_multiscale.jsonl")

print(atlas)

map = []

print("\n")

electionName = "G16_USS"  # This is the name of the eleection we will consider.



pctToDistVotes = {}

for node in pctData["nodes"]:

    pctToDistVotes[node["id"]] = {}



while map is not None:  # This loops through all of the map in the atlas

    try:

        map = Atlas.nextMap(atlas)  # Get the next map in the atlas 

        print(map.name)

        # The maps are multi scale in the sense that if a county is kept whole

        # the following fuction makes a map from precicts to districts out of 

        # the multiscale assignement 

        node_to_dist = hf.get_node_to_district(map.districting, 

                                               pctData["nodes"]) 

        # The next fuction  sums up the election for this districting (defined 

        # by the map)

        distVoteR, distVoteD, distVoteT = hf.sumElection(electionName, 

                                                         node_to_dist, 

                                                         dataElection)

        for id in pctToDistVotes.keys():

            votes = {}

            votes["Dem"] = distVoteD[node_to_dist[id]]

            votes["Rep"] = distVoteD[node_to_dist[id]]

            votes["Total"] = distVoteT[node_to_dist[id]]

            pctToDistVotes[id][map.name] = votes

    

    except Exception:

        break



# This secton adds the uniform swings by the amounts listed in swings for the lections

# listed in ellections to the dataEllections

swings=[.47+i/200 for i in range(13)]

electons=["G16_AG","G12_PR","G08_USS"]

hf.addUniformSwings(swings,electons,dataElection)





# now do somthing with data

id = 100  # Choose precinct

print(dataElection[id]["county"], dataElection[id]["prec_id"])



# or

county = "BEAUFORT"

prec_id = "BLCK"

id = [ii for ii in range(len(dataElection)) 

      if dataElection[ii]["county"] == county 

      and dataElection[ii]["prec_id"] == prec_id][0]

print(id)



df = pd.DataFrame(pctToDistVotes[id]).T

df["Dem %"] = df["Dem"]/df["Total"]

print(df)



sns.displot(data=df, x="Dem %", bins=10, stat="density")

plt.show()

helper_functions.py

0 → 100644
+57 −0
Original line number Diff line number Diff line 
def get_node_to_district(districting, nodes):

    # This function reconstucts for a multiscale districting description

    # the mapping from precinct IDs to Districts for a given map 

    node_id_to_district = {}

    for node in nodes:

        county = node["county"]

        cnty_key = '[\"'+county+'\"]'

        if cnty_key in districting:

            node_id_to_district[node["id"]] = districting[cnty_key]

            continue

        pct = node["prec_id"]

        cnty_pct_key = '[\"'+county+'\", \"' + pct + '\"]'

        if cnty_pct_key in districting:

            node_id_to_district[node["id"]] = districting[cnty_pct_key]

        else:

            print(node, county, pct, cnty_key, cnty_pct_key)

    return node_id_to_district





def sumElection(electionName, node_to_dist,  data):

    # This fuction 

    distVoteR = {}

    distVoteD = {}

    distVoteT = {}

    for id in node_to_dist.keys():

        d = node_to_dist[id]

        if d in distVoteT.keys():

            distVoteR[d] += data[id][electionName+"_R"]

            distVoteD[d] += data[id][electionName+"_D"]

            distVoteT[d] += data[id][electionName+"_T"]

        else:

            distVoteR[d] = data[id][electionName+"_R"]

            distVoteD[d] = data[id][electionName+"_D"]

            distVoteT[d] = data[id][electionName+"_T"]

    return distVoteR, distVoteD, distVoteT



def stateWideVotes(electionName,dataElection):

    votes={}

    votes["Total"]=sum([ p[electionName+'_T'] for  p in dataElection])

    votes["Rep"]=sum([ p[electionName+'_R'] for  p in dataElection])

    votes["Dem"]=sum([ p[electionName+'_D'] for  p in dataElection])

    return votes





def addUniformSwings(targetDemFractions,electionNames,dataElection):

    for electionName in electionNames:

        votes=stateWideVotes(electionName,dataElection)

        stateWideVoteFraction=votes["Dem"]/votes["Total"]

        for targetDemFraction  in targetDemFractions:

            for p in dataElection:

                p[electionName+'_USF'+str(targetDemFraction)+'_D']=p[electionName+'_D']*targetDemFraction/stateWideVoteFraction

                p[electionName+'_USF'+str(targetDemFraction)+'_R']=p[electionName+'_R']*(1.0-targetDemFraction)/(1.0-stateWideVoteFraction)

                p[electionName+'_USF'+str(targetDemFraction)+'_T']=p[electionName+'_T']





def listElections(dataElection, prefix="G", exluded={'id','prec_id','pop2020cen','MCD','area','border_length'},idx=0):

    return [ e[0:-2] for e in dataElection[idx].keys() if e not in {'id','prec_id','pop2020cen','MCD','area','border_length'} and e[-1] == 'T']