Accessing nytimes election data for procrastination purposes

nytimes_election_data.py

import requests
import pandas as pd


def fetch_state_results(state: str) -> dict:
    r = requests.get(f"https://static01.nyt.com/elections-assets/2020/data/api/2020-11-03/race-page/{state}/president.json")
    return r.json()["data"]


def _process_records(input: "list[dict]") -> pd.DataFrame:
    records = []
    for el in input:
        f_el = {}
        for k, v in el.items():
            if isinstance(v, dict):
                f_el.update({f"{k}.{subk}": subv for subk, subv in v.items()})
            else:
                f_el[k] = v
        records.append(f_el)
    return pd.DataFrame.from_records(records)


class StateResults(object):
    def __init__(self, state):
        self.state = state
        self._data = fetch_state_results(state)
        self._created = pd.Timestamp.now()

    def __repr__(self):
        return f"< StateResults for '{self.state}', created at {self._created} >"

    def _get_races(self):
        races = self._data["races"]
        assert len(races) == 1, f"Length of 'races' list not 1, was {len(races)}. Bad assumption."
        return races[0]

    def get_updated(self):
        return StateResults(self._state)

    def county(self):
        return _process_records(self._get_races()["counties"])

    def timeseries(self):
        df = _process_records(self._get_races()["timeseries"])
        df["timestamp"] = df["timestamp"].apply(pd.Timestamp)
        return df