Tweet Street is a data analysis project to identify if any correlation exists between tweet sentiments and stock prices. Our target stocks are: Amazon, Apple, Macy's, Netflix, Nike, Visa, and Walmart.
Here are the following visualizations in order:
- Compound sentiments
- Average stock prices (daily, two-day, weekly)
- Moving averages of stock prices (weekly) vs Tweet sentiments
- Stock closing prices vs Tweet sentiments
- Percentage comparisons of target stocks
Jefferson-Henrique GitHub Project (Historic Tweets), Tweepy (Daily Tweets)
Yahoo Finance(Historic Stocks), IEX API(Daily Stocks)
import pandas as pd
import numpy as np
import time
import matplotlib.pyplot as plt
import matplotlib
from datetime import datetime
import matplotlib.cm as cm
plt.style.use('bmh')handles = ["@Apple", "@Visa", "@amazon", "@Walmart",'@netflix','@Nike','@Macys']
dataframes = {}for handle in handles:
print(f"data/sentiments_{handle}.csv")
df = pd.read_csv(f"data/sentiments_{handle}.csv")
df = df.drop("Unnamed: 0",axis=1)
df = df.reset_index()
df = df.rename(columns={"index":"id"})
print(df.columns)
df = df.drop(df[df['compound'] == "compound"].index)
df = df.dropna(how='any')
df["compound"] = pd.to_numeric(df["compound"] )
df["neg"] = pd.to_numeric(df["neg"] )
df["neu"] = pd.to_numeric(df["neu"] )
df["pos"] = pd.to_numeric(df["pos"] )
df["date"] = [datetime.strptime(x,"%a %b %d %H:%M:%S %z %Y").date() for x in df["tweet_date"]]
df = df[["compound","date"]]
date_group = df.groupby(["date"])
compound_date = pd.DataFrame(date_group.mean())
compound_date = compound_date.reset_index()
dataframes.update({handle:compound_date})
data/sentiments_@Apple.csv
Index(['id', 'compound', 'entity', 'neg', 'neu', 'pos', 'tweet_date',
'tweet_id'],
dtype='object')
data/sentiments_@Visa.csv
Index(['id', 'compound', 'entity', 'neg', 'neu', 'pos', 'tweet_date',
'tweet_id'],
dtype='object')
data/sentiments_@amazon.csv
Index(['id', 'compound', 'entity', 'neg', 'neu', 'pos', 'tweet_date',
'tweet_id'],
dtype='object')
data/sentiments_@Walmart.csv
Index(['id', 'compound', 'entity', 'neg', 'neu', 'pos', 'tweet_date',
'tweet_id'],
dtype='object')
data/sentiments_@netflix.csv
Index(['id', 'compound', 'entity', 'neg', 'neu', 'pos', 'tweet_date',
'tweet_id'],
dtype='object')
data/sentiments_@Nike.csv
Index(['id', 'compound', 'entity', 'neg', 'neu', 'pos', 'tweet_date',
'tweet_id'],
dtype='object')
data/sentiments_@Macys.csv
Index(['id', 'compound', 'entity', 'neg', 'neu', 'pos', 'tweet_date',
'tweet_id'],
dtype='object')
for key, df in dataframes.items():
df.plot(kind="line",x="date",y="compound",color="blue",label="compound",marker="o")
plt.axhline(0.00)
plt.xlabel("date")
plt.ylabel("sentiment")
plt.title(f"Compound Sentiments {key}")
plt.xticks(rotation=60)
plt.legend()
plt.tight_layout()
plt.savefig(f"figures/Compound_Sentiments_{key}.png")
def plot_mvg_stocks(base,df,window_size):
df.plot(kind="line",x="Date",y="Adj Close",color="green",label="adj closing price",marker="o")
plt.xlabel("Date")
plt.ylabel("Adj Close")
plt.title(f"{base}_{window_size}_{ticker}")
plt.xticks(rotation=60)
plt.legend()
plt.tight_layout()
plt.savefig(f"figures/{base}_{window_size}_{ticker}.png")
tickers = ['AAPL', 'V', 'AMZN', 'WMT', 'NFLX', 'NKE', 'M']
for ticker in tickers:
df = pd.read_csv(f"data\{ticker}.csv")
df = df[["Date","Adj Close"]]
df_2 = df.rolling(2).mean()
df_2 = df_2.rename(columns={"Adj Close":"closing_avg_2"})
df_5 = df.rolling(5).mean()
df_5 = df_5.rename(columns={"Adj Close":"closing_avg_5"})
df = df.rename(columns={"Adj Close":"closing_avg_daily"})
df = df.merge(df_2,on="Date")
df = df.merge(df_5,on="Date")
df = df.dropna(how="any")
matplotlib.rcParams.update({'font.size': 16})
plt.rcParams["font.size"] = 18
plt.rcParams["font.weight"] = "bold"
plt.rcParams["axes.labelweight"] = "bold"
ax = df.plot(kind="line",x="Date",y="closing_avg_2",linewidth=4,figsize=(20,10),alpha=0.80,marker="o",color="royalblue")
df.plot(kind="line",x="Date",y="closing_avg_5",linewidth=7,ax=ax,marker="o",color="navy")
df.plot(kind="line",x="Date",y="closing_avg_daily",linewidth=2,ax=ax,marker="o",alpha=0.80,color="deepskyblue")
plt.xlabel("Date")
plt.ylabel("Adj Close")
plt.title(f"Stock Adjacent Closing Prices {ticker}")
plt.xticks(rotation=45)
plt.legend()
plt.tight_layout()
plt.savefig(f"figures/ppt/Stocks_Closing_{ticker}.png")
plt.show()
def plot_mvg_sentiments(base,df,window_size):
df.plot(kind="line",x="date",y="compound",color="blue",label="compound",marker="o")
plt.axhline(0.00)
plt.xlabel("date")
plt.ylabel("sentiment")
plt.title(f"{base}_{window_size}_{key}")
plt.xticks(rotation=60)
plt.legend()
plt.tight_layout()
plt.savefig(f"figures/{base}_{window_size}_{key}.png")
for key, df in dataframes.items():
base_file = "Compound_Senti_Mvg"
plot_mvg_sentiments(base_file,df.rolling(2).mean(),2)
plot_mvg_sentiments(base_file,df.rolling(5).mean(),5) tweet_Stock_map = {"@Apple":"AAPL","@Visa":"V","@amazon":"AMZN","@Walmart":"WMT",'@netflix':"NFLX","@Nike":"NKE",'@Macys':"M"}
stock_tweets_final = pd.DataFrame()
stock_tweets_m_avg = pd.DataFrame()
for handle,sentiment_df in dataframes.items():
sentiment_df = sentiment_df.rename(columns={"date":"Date"})
stock_df = pd.read_csv(f"data/{tweet_Stock_map[handle]}.csv")
stock_df = stock_df[["Date","Adj Close"]]
stock_df["Date"] = [datetime.strptime(x,"%Y-%m-%d").date() for x in stock_df["Date"]]
stock_df["Company"] = [handle for x in stock_df["Date"]]
stock_tweets_df = sentiment_df.merge(stock_df, on="Date")
stock_tweets_final = stock_tweets_final.append(stock_tweets_df)
stock_tweets_df = stock_tweets_df.sort_values(['Date'], ascending=True)
stock_tweets_df_rolling_2 = stock_tweets_df.rolling(2).mean()
stock_tweets_df_rolling_2 = stock_tweets_df_rolling_2.rename(columns = {"compound": "compound_m_avg 2",
"Adj Close":"stock_closing_m_avg 2"})
stock_tweets_df_rolling_5 = stock_tweets_df.rolling(5).mean()
stock_tweets_df_rolling_5 = stock_tweets_df_rolling_5.rename(columns = {"compound": "compound_m_avg 5",
"Adj Close":"stock_closing_m_avg 5"})
stock_tweets_df_w_2 = stock_tweets_df.merge(stock_tweets_df_rolling_2,on=["Date","Company"])
stock_tweets_df_w_5 = stock_tweets_df_w_2.merge(stock_tweets_df_rolling_5,on=["Date","Company"])
stock_tweets_m_avg = stock_tweets_m_avg.append(stock_tweets_df_w_5)
stock_tweets_df_w_5 = stock_tweets_df_w_5.dropna(how="any")
matplotlib.rcParams.update({'font.size': 16})
plt.rcParams["font.size"] = 18
plt.rcParams["font.weight"] = "bold"
plt.rcParams["axes.labelweight"] = "bold"
ax = stock_tweets_df_w_5.plot(kind="line",x="Date",y="compound_m_avg 5",linewidth=4,figsize=(20,10),alpha=0.80,marker="o",color="royalblue")
ax.set_xlabel('Date (s)')
# Make the y-axis label, ticks and tick labels match the line color.
ax.set_ylabel("Twitter Sentiments", color='b')
ax.set_title(f"Tweets vs Stocks (5-day Averages) - {handle}")
ax.legend(loc='upper left')
ax2 = ax.twinx()
stock_tweets_df_w_5.plot(kind="line",x="Date",y="stock_closing_m_avg 5",linewidth=4,ax=ax2,figsize=(20,10),alpha=0.80,marker="o",color="tomato")
ax2.set_ylabel("Stock closing price", color='g')
ax2.legend(loc='upper right')
plt.savefig(f"figures/ppt/TweetsVsStocks(5-day)_{handle}.png")
plt.show()
stock_tweets_final.to_csv("Summary_Historic_data.csv")
stock_tweets_m_avg.to_csv("Summary_Historic_moving_Averages.csv")
handle_group = stock_tweets_final.groupby("Company")
company_summary = handle_group.mean()
company_summary = company_summary.rename(columns={"compound":"Tweet_sentiment","Adj Close":"Stock_closing_price"})
company_summary