
from pytrends.request import TrendReq
import datetime as dt
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
from sklearn.linear_model import LinearRegression
from sklearn.metrics import mean_squared_error, r2_score

%matplotlib inline


pytrends = TrendReq(hl='en-US', tz=360)  # trend request
kw_list = ["\"substance abuser\" + \"substance abusers\" + \"substance abuse\"", "\"substance user\" + \"substance users\" + \"substance use\"" ]  # keyword list
pytrends.build_payload(kw_list, cat=0, timeframe='2004-01-01 2021-12-31', geo='US', gprop='')  # build pytrends data
trendinfo = pd.DataFrame(pytrends.interest_over_time())  # interest over time, dataframe format


# SHOWS ALL ROWS
#pd.set_option('display.max_rows',trendinfo.shape[0]+1) 

# SHOWS HEAD (first 5 rows in dataframe)
trendinfo.head()


# define variables for plotting
indices = trendinfo.index  # in datetime format

x = np.array(trendinfo['"substance abuser" + "substance abusers" + "substance abuse"']).reshape((-1, 1)) # substance use array
y = np.array(trendinfo['"substance user" + "substance users" + "substance use"']).reshape((-1, 1)) # substance abuse array

fig, ax = plt.subplots()
ax.scatter(indices, y, color='g', s=1, label='\"substance use\" OR \"substance user\" OR \"substance users\"')
ax.scatter(indices, x, color='r', s=1, label='\"substance abuse\" OR \"substance abuser\" OR \"substance abusers"')

plt.legend(fontsize=8.9, bbox_to_anchor=(0, 1), loc='upper left', borderaxespad=0.)
plt.ylabel('Search Interest')
plt.xlabel('Year')
plt.show()
fig.savefig("substance_use___substance_abuse.png", dpi=300)

	"substance abuser" + "substance abusers" + "substance abuse"	"substance user" + "substance users" + "substance use"	isPartial
date
2004-01-01	76	4	False
2004-02-01	87	0	False
2004-03-01	100	7	False
2004-04-01	70	4	False
2004-05-01	88	2	False

"Substance use" / "Substance abuse"¶