Introduction

In mid-2021, a phenomenon occurred where failing stocks such as Gamestop, AMC, and Bed Bath and Beyond saw a dramatic demand that pushed the price of these stocks upwards. The financial markets sought an explanation, finding one from one of the most obscure places: Reddit. On the social media platform Reddit, people were discussing on the subreddit “WallStreetBets” which stocks they should hedge their bets on to make money. Using positive phrases such as “[stock prices will] rise to the moon!”, “guaranteed winner” and “buy every chance I get” propelled people to invest in these stocks, which explained the drastic fluctuation in stock prices. Furthermore, our curiosity spanned analyzing stock data from the blue-chip technology stocks Meta (formerly Facebook), Amazon, Apple, Netflix, and Google.  Inspired by the WallStreetBets phenomenon and yearning to analyze MAANG stock data, we aimed to analyze Reddit Sentiment Data to capitalize on Redditor’s sentiments for profit.

 

Approach

-Our project focused on using the subreddit with the most popular pages related to the MAANG company.
-Our methodology was to get the most amount of information relevant to what people (everyday Redditors who consume the technology the MAANG companies provide) thought of the company compared to the information relevant to MAANG stock data.

-We also used Yahoo Finance and focused on collecting data from MAANG stocks to implement a complete stock analysis.

-Once we found the data relevant to the project, we scraped the data using the YFinance API and the Python Reddit API Wrapper (PRAW).

-Included in our analysis was Wordclouds to differentiate and visualize sentiments towards the MAANG companies and the Stock Price X Vader Sentiment Model to help predict stock price changes with sentiment data.

​

Data

• Yahoo Finance

  • Meta Historical Stock Information ( YFinance API, 138 KB, 227 Records, Date Range 01-01-2022 to 11-27-2022 ) 

  • Apple Historical Stock Information ( YFinance API, 138 KB, 227 Records, Date Range 01-01-2022 to 11-27-2022 ) 

  • Amazon Stock Information ( YFinance API, 125 KB, 227 Records, Date Range 01-01-2022 to 11-27-2022 ) 

  • Netflix Historical Stock Information (YFinance API, 131 KB, 227 Records,  Date Range 01-01-2022 to 11-27-2022 ) 

  • Google Historical Stock Information (YFinance API, 132KB, 227 Records, Date Range 01-01-2022 to 11-27-2022 ) 

• Reddit Data Sources

  • Metastock Subreddit(PRAW API, 124 KB, 400 Records, Date Range 30-04-2021 to 02-09-2022)

  • metaverse Subreddit(PRAW API, 258 KB, 852 Records, Date Range 01-01-2022 to 11-27-2022)

  • Facebook Subreddit(PRAW API, 258 KB, 852 Records, Date Range 01-01-2022 to 11-27-2022)

  • Apple Subreddit (PRAW API, 258 KB, 852 Records, Date Range 01-01-2022 to 11-27-2022)

  • IOS Subreddit (PRAW API, 262 KB, 864 Records, Date Range 01-01-2022 to 11-27-2022)

  • iPhone Subreddit (PRAW API, 285 KB, 938 Records, Date Range 01-01-2022 to 11-27-2022)

  •  Amazonprime Subreddit (PRAW API, 300 KB, 985 Records, Date Range 30-04-2021 to 02-09-2022) 

  •  Amzn Subreddit(PRAW API, 90 KB , 75 Records, Date Range 08-09-2011 to 19-11-2022)     

  •  Amazonreviews Subreddit(PRAW API, 210 KB , 884 Records, Date Range 30-04-2021 to 09-10-2022)

  • Netflix Subreddit(PRAW API, 300 KB , 400 Records, Date Range 30-04-2021 to 09-10-2022)

  • NetflixBestof Subreddit(PRAW API, 550 KB , 500 Records, Date Range 30-04-2021 to 09-10-2022)

  • NetflixViaVpn Subreddit(PRAW API, 124 KB , 400 Records, Date Range 30-04-2021 to 09-10-2022)

  •  Google Subreddit (PRAW API, 124 KB, 408 Records, Date Range 30-11-2022 to 01-12-2022)

  •   GooglePixel Subreddit (PRAW API, 124 KB, 408 Records, Date Range 30-11-2022 to 01-12-2022)

  •   Googlehome Subreddit (PRAW API, 124 KB, 408 Records, Date Range 30-11-2022 to 01-12-2022)

  •   Chrome Subreddit (PRAW API, 124 KB, 408 Records, Date Range 30-11-2022 to 01-12-2022)

  •  GoogleMaps Subreddit (PRAW API, 124 KB, 408 Records, Date Range 30-11-2022 to 01-12-2022)

  •  Googlephotos Subreddit (PRAW API, 124 KB, 408 Records, Date Range 30-11-2022 to 01-12-2022)

  •  Googlecloud Subreddit (PRAW API, 124 KB, 408 Records, Date Range 30-11-2022 to 01-12-2022)

  •  Degoogle Subreddit (PRAW API, 124 KB, 408 Records, Date Range 30-11-2022 to 01-12-2022)

• YFinance API 

  • Extracts Data from Yahoo Finance. Documentation is here: https://pypi.org/project/yfinance/

• Python Reddit API Wrapper (PRAW)

  • Extract Reddit Post Data. NOTE: to run, need to have a Reddit account. Documentation is here: https://praw.readthedocs.io/en/stable/
     

Experiments:

Google Colab with our Data:  https://colab.research.google.com/drive/1nWm6_bOOACaBu42Me8hBpBJclUjl9lFw?usp=sharing

-Investigating Reddit, the team took the incentive to actively seek the most recently active subreddits related to each MAANG company.  

-The first step was scraping the data using the YFinance API and the Python Reddit API Wrapper (PRAW). 

-Afterward, we used the Natural Language Toolkit (NLTK) library module VADER (Valence Aware Dictionary and sEntiment Reasoner) to analyze our Reddit data.

-The two most relevant data visualizations were i) Word Cloud Visualization to Sentiments and ii) our created VADER X Stock Price Model.

-By normalizing the stock and Vader compound data, we were able to accurately analyze the relationship between sentiments and the stock price, which correlates the stock price with the trend of the sentiment.

Results

I. Apple

WordClouds:  All Words (gray), Negative Words (red), Positive Words (green)

​

​

​

​

​

​

​

 

 

 

Distribution of Words

​

​

​

​

​

​

​

​

​

​

​

​

​

​

Stock Price X VADER Sentiment Model (Created by us)

​

​

​

​

​

​

​

​

​

​

​

​

​

​

III. Amazon

​

​

Discussion

​

-Meta Backlash
  -Yet sentiment stayed almost     Neutral
  -Result of Moderation
Sentiment-Stock Price Inversions
   - Seen with Meta and Google 
   - Buy or Hold, don’t sell.
Apple’s Pro Line
- Premium is rated negatively 
- Negative Sentiment because of increased cost w/ fewer improvements

​

​

Conclusion

-Stock Prices are positively correlated with sentiment.
   -If Stock Price increases, Sentiment is Positice.
   -If Stock Price decreases, Sentiment is Negative.
   -Sentiment and Stock Price Fluctuations are visible with examples like Apple

-Correlation in Company Sentiments
    -Apple and Amazon are correlated because of physical capital
    -Netflix and Google are correlated because of software

-How can investors make money?
    -Invest in strengths of Company Positives
    - Companies can know about the effect of positive sentiments and capitalize by moderating sentiments
    -Be Vigilant of Product Improvements and Invest wisely

 

References

Mac, R. (2022, October 26). Meta's profit slides by more than 50 percent as challenges Mount. The New York Times. Retrieved December 6, 2022, from https://www.nytimes.com/2022/10/26/technology/meta-facebook-q3-earnings.html 

Malde, R. (2020, July 6). A short introduction to Vader. Medium. Retrieved December 6, 2022, from https://towardsdatascience.com/an-short-introduction-to-vader-3f3860208d53

​