AP Statistics Final Project: Is FRC pay-to-win?

main.py

from itertools import count
import sys
import re
import random
import requests as req_orig
import requests_cache
import statbotics
from code import interact
import geocoder

requests = requests_cache.CachedSession('demo_cache')

apiKey = open(".apikey").read().strip()
sb = statbotics.Statbotics()
def getTBA(path): return requests.get("https://www.thebluealliance.com/api/v3"+path, headers={"X-TBA-Auth-Key":apiKey}).json()

def printLine(*args):
    out = ""
    for arg in args:
        out += str(arg).ljust(30) + "\t"
    print(out.strip())

import csv
from geopy.geocoders import Nominatim
geolocator = Nominatim(user_agent="apstats_cole_wilson")

incomes = {}
with open('income.csv', newline='') as csvfile:
    spamreader = csv.reader(csvfile, delimiter=',', quotechar='"')
    prevcounty = ""
    for row in spamreader:
        state, county, income = row
        if prevcounty != "" and income != "":
            if state not in incomes:
                incomes[state] = {}
            incomes[state.strip()][county.strip()] = float(income)
            # print(county, state)
        prevcounty = county
        # print(geolocator.geocode(f"{county}, {state}").latitude, county, state)

# city_dict = {}                                  #   0         1         2           3             4           5
# with open('cities.csv', newline='') as csvfile: #"city","city_ascii","state_id","state_name","county_fips","county_name","lat","lng","population","density","source","military","incorporated","timezone","ranking","zips","id"
#     spamreader = csv.reader(csvfile, delimiter=',', quotechar='"')
#     for row in spamreader:
#         city_dict[(row[0], row[3])] = row[5]

page = 0
teams = []
while True:
    data = getTBA(f"/teams/2024/{page}/simple")
    if len(data) == 0:
        break
    teams += data
    page += 1
# print(teams)
printLine("Team Number", "Income", "EPA", "County", "State", "Coords", "Name")
teams = list(filter(lambda i:i["country"]=="USA", teams))


print(len(teams))
teams = random.sample(teams, k=281)
i = 0
for team in teams: # state_prov name city team_number
    sys.stderr.write(f"{i}/281\r")
    i+=1
    city = team["city"]
    # print(team)
    state = team["state_prov"]
    location = geolocator.geocode(f"{city}, {state}")
    coords = (location.latitude, location.longitude)
    location = location.address
    county = re.findall(r"(\w[\w\s]*) County", location, re.MULTILINE)
    if len(county) == 0:
        sys.stderr.write(f"county for {location} >> ")
        sys.stderr.flush()
        county = input()
    else:
        county = county[0].replace("Saint", "St.")
    try:
        income = incomes[state.strip()][county.strip()]
    except KeyError:
        sys.stderr.write(f"nothing found for county {county}, {state}. income?? >> ")
        sys.stderr.flush()
        income = float(input())
        # interact(local=locals())
    epa = sb.get_team_year(team["team_number"], year=2024)["epa_end"]
    printLine(team["team_number"], income, epa, county, state, coords, team["nickname"])



    # print(city, state, location.address.split(",")[1])