import pandas as pd
import numpy as np
from sklearn import datasets, linear_model

def airport_code_to_df(code):
    return pd.to_numeric(airport_dict[code], errors='coerce')

df = pd.read_csv("flightdata.csv")
df = df.dropna(axis=0,how='any')

airports = df["departure_airport"].unique()
airport_dict = dict(zip(airports, range(len(airports))))
airports = df["departure_airport"].replace(airport_dict)
airports = pd.to_numeric(airports, errors='coerce')

data = pd.DataFrame({ "scheduled" : df["departure_schedule"],
                      "airport"   : airports })

target = df["departure_delay"]

reg = linear_model.ElasticNet()
reg.fit(data[:-20], target[:-20])
print(reg.predict(data[-20:]))