1. Google Search Trend and Stock Price Tracker

The following Jupyter notebook supports use of the pyrends API for querying changes in Google search trends and the Polygon.io API for querying changes in a given public stock price. The notebook will generate a simple graph to show how these changes may corrolate over a selected time period. When running the notebook locally, be sure to add your Polygon.io API key to the stocks.py file (may be found in the src/ directory).

1.1. Step 1: Import Libraries

Import pytrends, pandas, datetime, matplotlib and seaborn goodies, and homemade classes Stocks and Google.

from pytrends.request import TrendReq
import pandas as pd
import datetime as dt
from datetime import timedelta

from pytrends_support_hvk import google_trends
from stock_api_call.stocks import Stocks

import matplotlib.pyplot as plt
import matplotlib.dates as mdates
from pandas.plotting import register_matplotlib_converters
import seaborn as sns
register_matplotlib_converters()

1.2. Step 2: Set Variables

Declare Google search terms, days for analysis (max = 250), and stock ticker.

kw_list = "vaccine"
days = 250
stock_ticker = "JNJ"

1.3. Step 3: Run APIs

Query Google Trends and Closing Stock Prices with selected variables. Tweak results (add index to Google dataframe and change dates to datetime objects in Stocks dataframe).

google_trends = google_trends.Google (kw_list, days)
stock_trends = Stocks (stock_ticker, days)
gtf = google_trends.trends()
gtf = gtf.reset_index()
stf = stock_trends.stocks()
stf['Dates'] = pd.to_datetime(stf['Dates'])

1.4. Step 4: Merge results into one Trends table

Save as .csv for playing around with

trends = pd.merge(left=gtf, 
                    right=stf, 
                    how='inner',
                    left_on='date', 
                    right_on='Dates')
trends.to_csv("trends.csv", index = False)
trends

	date	vaccine	isPartial	Dates	Closing Price
0	2022-05-24	66	False	2022-05-24	181.40
1	2022-05-25	51	False	2022-05-25	179.62
2	2022-05-26	49	False	2022-05-26	179.46
3	2022-05-27	48	False	2022-05-27	181.09
4	2022-05-31	53	False	2022-05-31	179.53
...	...	...	...	...	...
164	2023-01-19	41	False	2023-01-19	169.53
165	2023-01-20	41	False	2023-01-20	168.74
166	2023-01-23	41	False	2023-01-23	168.31
167	2023-01-24	41	False	2023-01-24	168.31
168	2023-01-25	39	False	2023-01-25	169.51

169 rows × 5 columns

1.5. Step 5: Create a Pretty Graph

Using Seaborn

with sns.axes_style('darkgrid'):
    plt.figure(figsize=(8,4), dpi=200)
    plt.xticks(fontsize=10, rotation = 45)
    plt.title(f'Web Search Index for "{kw_list}" and Corresponding {stock_ticker} Price', fontsize=14)

    ax1 = plt.gca()
    ax2 = plt.twinx()
   
    sns.lineplot(x= trends['date'], y=trends[kw_list], data=trends, palette="tab10", linewidth=2.5, ax=ax1)
    sns.lineplot(x = trends['date'], y = trends['Closing Price'], data=trends, color = "orange", linewidth=2.5, ax=ax2)
    
    ax1.set_ylabel('Google Search Index', fontsize=12, color='blue')
    ax2.set_ylabel('Stock Price', fontsize=12, color='orange')

plt.show()

/tmp/ipykernel_7020/779364875.py:9: UserWarning: Ignoring `palette` because no `hue` variable has been assigned.
  sns.lineplot(x= trends['date'], y=trends[kw_list], data=trends, palette="tab10", linewidth=2.5, ax=ax1)