brockmanmatt

fig, axs = plt.subplots(4, 5, sharex=True)

fig.set_size_inches(16,12)

x = y = 0

for issue in time_series:
  train_l = len(time_series)-5

  selected_series = time_series[[col for col in time_series.columns if (col.find(issue[:issue.find("_")]) > -1)]]

brockmanmatt

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt

## This assumes a jupyter notebook file to install gdelt
## if not using notebook, remove and run on the command line for your OS
!pip install gdelt
## (the ! gives command line capabilities to notebook)

import gdelt

brockmanmatt

fig, axs = plt.subplots(4, 5, sharex=True)

fig.set_size_inches(16,12)

x = y = 0

for issue in time_series:

  if not issue.find(".com") > -1:
    continue

brockmanmatt

mape_df = pd.DataFrame()

fig, axs = plt.subplots(4, 5, sharex=True)

fig.set_size_inches(16,12)

x = y = 0

for issue in time_series:

brockmanmatt

fig, axs = plt.subplots(4, 5, sharex=True)

fig.set_size_inches(16,12)

x = y = 0

for issue in time_series:
  if not issue.find(".com") > -1:
    continue
  train_l = len(time_series)-5

brockmanmatt

mySources = ["cnn.com", "washingtonpost.com", "nytimes.com", "foxnews.com"]

#unnest the entries with multiple sources in them
df = df.set_index(df.columns.drop('SOURCES',1).tolist()).SOURCES.str.split(';', expand=True).stack().reset_index().rename(columns={0:'SOURCES'}).loc[:, df.columns]

df.DATE = df.DATE.apply(lambda x: str(x)) #convert date
df.DATE = pd.to_datetime(df.DATE)
df.fillna("", inplace=True)
df.set_index("DATE", drop=True, inplace=True)

brockmanmatt

from statsmodels.tsa.statespace.sarimax import SARIMAX

def gen_SARIMA_result(p,d,q, df, issue, test_length): #takes pdq, data, issue to use

  s_model = SARIMAX(endog = df[issue][:-test_length],
                    exog = df[[x for x in df.columns if x != issue]][:-test_length],
                    order=(p,d,q), seasonal_order=(1,0,1,7)).fit()

  f_ru = df[[issue]].copy()[1:] #haven't bothered to change this, but it's the results

brockmanmatt

fig, axs = plt.subplots(1,3, figsize=(12,8))

issue = time_series.columns[0]
test_length = 10
selected_series = time_series[[col for col in time_series.columns if (col.find(issue[issue.find("_"):]) > -1)]].shift()[1:].drop(columns=issue).add_suffix("_l1")
pub_series = time_series[[col for col in time_series.columns if (col.find(issue[:issue.find("_")]) > -1)]].drop(columns=issue).shift()[1:].add_suffix("_l1")
selected_series = selected_series.join(pub_series).join(time_series[issue])
x = 0

for p in [1,5,10]:

brockmanmatt

model_dict = {}

for issue in time_series:

  model_dict[issue]=[]
  test_length = 5

  selected_series = time_series[[col for col in time_series.columns if (col.find(issue[issue.find("_"):]) > -1)]].shift()[1:].drop(columns=issue).add_suffix("_l1")
  pub_series = time_series[[col for col in time_series.columns if (col.find(issue[:issue.find("_")]) > -1)]].drop(columns=issue).shift()[1:].add_suffix("_l1")
  selected_series = selected_series.join(pub_series).join(time_series[issue])

brockmanmatt

fig, axs = plt.subplots(4, 5, figsize=(20,12))
x = y = 0

for issue in time_series:

  pd.DataFrame(model_dict[issue]).plot(x="aic", y="test_mape", style=".", title="{}".format(issue), ax=axs[x,y])
  pd.DataFrame(model_dict[issue]).plot(x="aic", y="train_mape", style=".", title="{}".format(issue), ax=axs[x,y])

  if x == 0 and y==0:
    handles, labels = axs[0,0].get_legend_handles_labels()