NYT election data analyzer

# -*- coding: utf-8 -*-

""" HOW TO USE SCRIPT:
    Download data from nyt like below:
        https://static01.nyt.com/elections-assets/2020/data/api/2020-11-03/race-page/alabama/president.json
    Change "alabama" to the state you want to download. Use hyphens for states w space in their name like "new-york".
    Place all the .json files in the same directory as this script.
    If you want to run only specific state(s), place the .json files ONLY for those state(s) in the script's directory.
    Run the script. If you are not using an IDE like SPYDER you might have to add a line for plt.show().

    Version 2020/11/12-12:00
"""

import json
import matplotlib.pyplot as plt
from datetime import datetime
import os

import locale
locale.setlocale(locale.LC_ALL, '')


def findfraud():
    stateFiles = [x for x in os.listdir(".") if x.endswith(".json")]

    switchTimes = [] #-number of times data got switched

    for state in stateFiles:
        with open(state, encoding="utf8") as f:
            x = json.load(f)

        series = x["data"]["races"][0]["timeseries"]

        trumpTotal = 0
        bidenTotal = 0

        trumpLoss = 0
        bidenLoss = 0

        trumpWin = 0
        bidenWin = 0

        shiftedToBiden = 0
        shiftedToTrump = 0
        shiftedFromBiden = 0
        shiftedFromTrump = 0

        shiftsToBiden = 0
        shiftsToTrump = 0

        for i in range(1, len(series)):

            curr =  series[i]
            prev = series[i-1]

            votes = curr["votes"]
            prevVotes = prev["votes"]

            trumpShare = curr["vote_shares"]["trumpd"]
            trumpVotes = votes * trumpShare
            prevTrumpShare = prev["vote_shares"]["trumpd"]
            prevTrumpVotes = prevVotes * prevTrumpShare

            bidenShare = curr["vote_shares"]["bidenj"]
            bidenVotes = votes * bidenShare
            prevBidenShare = prev["vote_shares"]["bidenj"]
            prevBidenVotes = prevVotes * prevBidenShare

            trumpDelta = trumpVotes - prevTrumpVotes
            bidenDelta = bidenVotes - prevBidenVotes

            #print("INDEX " + str(i))

            if trumpVotes < prevTrumpVotes:
                trumpLoss += abs(trumpDelta)
                trumpTotal += trumpDelta
                #print ("TRUMP LOSS: " + str(trumpDelta))

            if bidenVotes < prevBidenVotes:
                bidenLoss += abs(bidenDelta)
                bidenTotal += bidenDelta
                #print ("BIDEN LOSS: " + str(bidenDelta))

            if trumpVotes > prevTrumpVotes:
                trumpWin += abs(trumpDelta)
                trumpTotal += trumpDelta
                #print ("TRUMP WIN: " + str(trumpDelta))

            if bidenVotes > prevBidenVotes:
                bidenWin += abs(bidenDelta)
                bidenTotal += bidenDelta
                #print ("BIDEN WIN: " + str(bidenDelta))

            #if trumpVotes == prevTrumpVotes:
                #print ("TRUMP NEUTRAL: " + str(trumpDelta))

            #if bidenVotes == prevBidenVotes:
                #print ("BIDEN NEUTRAL: " + str(bidenDelta))


            if trumpVotes < prevTrumpVotes and bidenVotes > prevBidenVotes:
                shiftsToBiden+=1
                shiftedToBiden += bidenDelta
                shiftedFromTrump += -trumpDelta
                switchTimes.append(extract_timestamp(curr['timestamp']))

            if trumpVotes > prevTrumpVotes and bidenVotes < prevBidenVotes:
                shiftsToTrump+=1
                shiftedToTrump += trumpDelta
                shiftedFromBiden += -bidenDelta
                switchTimes.append(extract_timestamp(curr['timestamp']))
        """
        #-PRINT STATEMENTS FROM ORIGINAL AUTHOR
        print("CALCULATION COMPLETE")
        print("-----")
        print("Total sum of votes gained/lost for both candidates")
        print("TRUMP TOTAL        " + "{:>12,d}".format(int(trumpTotal)))
        print("BIDEN TOTAL        " + "{:>12,d}".format(int(bidenTotal)))
        print("-----")
        print("No. times 'A' gained votes while 'B' lost votes")
        print("SHIFTS TO TRUMP    " + "{:>12,d}".format(int(shiftsToTrump)))
        print("SHIFTS TO BIDEN    " + "{:>12,d}".format(int(shiftsToBiden)))
        print("-----")
        print("No. votes that has been shifted")
        print("SHIFTED TO TRUMP   " + "{:>12,d}".format(int(shiftedToTrump)))
        print("SHIFTED FROM TRUMP " + "{:>12,d}".format(int(shiftedFromTrump)))
        print("SHIFTED TO BIDEN   " + "{:>12,d}".format(int(shiftedToBiden)))
        print("SHIFTED FROM BIDEN " + "{:>12,d}".format(int(shiftedFromBiden)))
        print("-----")
        print("Total votes gained and lost")
        print("TRUMP WIN          " + "{:>12,d}".format(int(trumpWin)))
        print("TRUMP LOSS         " + "{:>12,d}".format(int(trumpLoss)))
        print("BIDEN WIN          " + "{:>12,d}".format(int(bidenWin)))
        print("BIDEN LOSS         " + "{:>12,d}".format(int(bidenLoss)))
        """

    """HERE IS HOW TO SWITCH BETWEEN SINGLE DAY AND MANY DAYS"""

    #-Use the following line to get Nov 4th only:
    # switchTimes = Nov4Only(switchTimes)

    #-Update titles and axes labels below for what you are plotting:
    fig, ax = plt.subplots(figsize=(10,6))
    ax.tick_params(axis='x', which='major', labelsize=6.5)
    ax.set_ylabel('Number of times votes \'switched\'')
    ax.set_xlabel('Date')
    ax.set_title('Votes \'Switched\' - All 50 States + D.C. - 11/3/2020->Onward\nAnalysis of data obtained from NYT')

    #-Use the following line for entire dataset from 11/3 onward
    ax.hist(switchTimes, bins=183)

    #-Use the following 3 lines for 24 hour time period only; e.g.: Nov 4th only
    # ax.set_xticks(range(0, 86400, 3600)) #-sets each tick on the x axis to be 1 hour (the numbers are in seconds)
    # plt.xlim(0, 86400) #-sets limit to full day in case there is no data at beginning or end
    # ax.hist(switchTimes, bins=range(0, 90000, 3600), rwidth=0.9) #-Each bin to correspond to each hour


def extract_timestamp(timeStr):
    """ Extracts timestamp from datum """
    #-Extract date and time
    return datetime.strptime(timeStr, "%Y-%m-%dT%H:%M:%SZ")

    #-Extract time of day only
    #return datetime.strptime(timeStr, "%Y-%m-%dT%H:%M:%SZ").time()

def Nov4Only(switchTimes):
    """ use only nov 4th data """

    newList = []

    for switch in switchTimes:
        if (switch >= datetime(2020, 11, 4, 0, 0) and switch < datetime(2020, 11, 5, 0, 0)):
            newList.append(switch.time()) #-add only if on Nov 4th, and remove date

    return newList


if __name__ == "__main__":
    findfraud()