COVID-19 and Election Data Analysis

Scrape and Wrangle:

Code Editor:

#IMPORTS
import pandas as pd 
import numpy as np 
from scipy import stats
#Pandas view settings
pd.options.display.max_columns = None
pd.options.display.max_rows = None
pd.set_option('max_colwidth', None)
pd.set_option('max_rows', None)
pd.set_option('display.expand_frame_repr', False)
#Plots
import matplotlib.pyplot as plt
import plotly.io as pio
import plotly.figure_factory as ff
from plotly.subplots import make_subplots
pio.renderers.default = "browser"

def make_color_scale():
    '''
    Function to make a color scale for a plot. 
    '''
    r = 255 ; g = 255 ; b = 255
    colorscale = []
    for i in range(20):
        color = 'rgb({}, {}, {})'.format(r, g, b)
        r -= 5
        g -= 5
        b -= 0
        colorscale.append(color)
    return colorscale
#make_color_scale()

def reverse(lst):
    '''
    Function to reverse a list. Intended for colorscale lists 
    if they need to be reversed. 
    '''
    return [ele for ele in reversed(lst)]
#Reverse(SOMELIST)

def get_data():
    '''
    This function scrapes the data. 
    '''
    
    '''
    Retrieve data from MIT Election Data and Science Lab
    Source: https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/VOQCHQ
    This one I had from https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/VOQCHQ
    '''
    df_pres_elect_by_county_2020_raw = pd.read_csv(r'Path\To\File\df_pres_elect_by_county_2000_2020_raw.csv')
    
    '''
    Covid data from Center for Systems Science and Engineering at Johns Hopkins University.
    Source: https://github.com/CSSEGISandData/COVID-19
    These are cumulative sums over time. 
    '''
    tm_series_cov_con_cases_raw = pd.read_csv(r'https://github.com/CSSEGISandData/COVID-19/raw/master/csse_covid_19_data/csse_covid_19_time_series/time_series_covid19_confirmed_US.csv')
    tm_series_cov_con_deaths_raw  = pd.read_csv(r'https://github.com/CSSEGISandData/COVID-19/raw/master/csse_covid_19_data/csse_covid_19_time_series/time_series_covid19_deaths_US.csv')
    ''' 
    Vaccine data from the US Center for Disease Control and Prevention (CDC)
    Source: https://data.cdc.gov/Vaccinations/COVID-19-Vaccinations-in-the-United-States-County/8xkx-amqh/data
    '''
    df_vac_data_raw = pd.read_csv(r'Path\To\File\df_vac_data_raw.csv')
    return df_pres_elect_by_county_2020_raw, tm_series_cov_con_cases_raw, tm_series_cov_con_deaths_raw, df_vac_data_raw
df_pres_elect_by_county_2020_raw, tm_series_cov_con_cases_raw, tm_series_cov_con_deaths_raw, df_vac_data_raw = get_data()

def clean_covid_df(df):
    '''
    This function makes the format of the state and county data in the vaccine 
    data uniform with the other dataframes.
    '''
    df['FIPS'] = df['FIPS'].astype(str).replace('\.0', '', regex=True)
    #by this point fips is a str col
    df['FIPS'] = df['FIPS'].str.zfill(5)
    return df 
df_cov_cases = clean_covid_df(tm_series_cov_con_cases_raw)
df_cov_deaths = clean_covid_df(tm_series_cov_con_deaths_raw) 

def clean_cdc_vac(df):
    df = df.loc[df['Date']=='12/30/2021']
    return df
df_vaccine = clean_cdc_vac(df_vac_data_raw)

''' 
Isolate the election data we need from 2020.
''' 
def iso_elect(df):
    df = df.loc[df['year']==2020]
    df = df[['state', 'county_name', 'county_fips', 
             'candidate', 'candidatevotes', 'totalvotes']]
    
    df = df.loc[(df['candidate']=='DONALD J TRUMP')|
                (df['candidate']=='JOSEPH R BIDEN JR')]
    
    df.rename(columns={'county_fips':'FIPS'}, inplace=True)
    df['FIPS'] = df['FIPS'].astype(str).replace('\.0', '', regex=True)
    df['FIPS'] = df['FIPS'].str.zfill(5)
    #added a percentage of total votes column 
    df['pct_votes'] = (df['candidatevotes'] / df['totalvotes'])*100
    return df
df_pres = iso_elect(df_pres_elect_by_county_2020_raw)

Analysis:

'''
What is most important here is the swing states. Although first let us look at 
all the states as a whole. 
'''
def merge_pol_vac(pol_df, vac_df, cases_df, deaths_df):
    '''
    First I will merge the political data with vac_data by percent of 18 to 64 
    and 65plus, also the current cases and deaths to date. 
    '''
    vac_df = vac_df[['FIPS', 'Series_Complete_Pop_Pct']]
    cases_df = cases_df[['FIPS', '12/29/21']] #pulls totals over time from 12/29/21
    deaths_df = deaths_df[['FIPS', 'Population', '12/29/21']]
    
    pol_df = pol_df.merge(vac_df, on='FIPS')
    pol_df = pol_df.merge(cases_df, on='FIPS', suffixes=('_cases', '_deaths'))
    pol_df = pol_df.merge(deaths_df, on='FIPS', suffixes=('_cases', '_deaths'))
    pol_df['pct_votes'] = pol_df.groupby(['FIPS', 'county_name', 'candidate'])['pct_votes'].transform('sum')
    pol_df['tot_state_votes_for_can'] = pol_df.groupby(['state', 'candidate'])['candidatevotes'].transform('sum')
    pol_df['tot_deaths_by_state'] = pol_df.groupby(['state' ])['12/29/21_deaths'].transform('sum')
    return pol_df 
merged_pol_vac = merge_pol_vac(df_pres, df_vaccine, df_cov_cases, df_cov_deaths)

'''
What I need now is cases and deaths per capita. So Here I will add that column. 
'''
def add_per_cap(df):
    df['per_capita_deaths'] = df['12/29/21_deaths']/df['Population']*100000
    df['per_capita_cases'] = df['12/29/21_cases']/df['Population']*100000
    return df
merged_pol_vac = add_per_cap(merged_pol_vac)

'''
Now let us isolate the swing states. 
'''
def get_swing_states(df): 
    swing_states = ['ARIZONA', 'COLORADO', 
                    'FLORIDA', 'GEORGIA', 
                    'MICHIGAN', 'NEVADA', 
                    'NEW HAMPSHIRE', 'NORTH CAROLINA', 
                    'OHIO', 'PENNSYLVANIA', 'WISCONSIN']
    df = df.loc[df['state'].isin(swing_states)]
    df = df[['state', 'county_name', 'FIPS', 
             'candidate', 'pct_votes', 
             'Series_Complete_Pop_Pct', 
             '12/29/21_cases', 'Population', 
             '12/29/21_deaths', 'per_capita_deaths',
             'per_capita_cases', 'tot_state_votes_for_can', 
             '12/29/21_deaths', 'tot_deaths_by_state']]
             
    df.drop_duplicates(inplace=True)
    
    return df 
df_sw = get_swing_states(merged_pol_vac)

'''
So now we pull the swing states tot_state_votes_for_can and tot_deaths_by_state;
and compare by how much each candidate won or lost by. 
'''
def get_victory_by(df):
    df = df[['state','candidate','tot_state_votes_for_can', 'tot_deaths_by_state']]
    df.drop_duplicates(inplace=True)
    temp1 = df.iloc[0::2, 2]
    temp2 = df.iloc[1::2, 2]
    
    d_wins = temp2.values - temp1.values
    b_wins = temp1.values - temp2.values
    
    result = [None]*(len(temp1)+len(temp2))
    result[::2] = b_wins
    result[1::2] = d_wins
    df.insert(4, 'win_loss_diff', result)
    return df
df_sw_vic = get_victory_by(df_sw)