SamuelLawrence876 / Twitter-Sentiment

A Stock Market predictor based on twitter data

A Stock Market sentiment analyzer based on twitter data using nlp

-- Project Status: [Complete]

Project phases:

• [curate tweets]
• [Perform data cleaning]
• Analyze data
• submit findings
• EDA analysis
• compare stock price movement to buy/sell positions

Project Intro

In an attempt to understand the voice of investors, this project seeks to understand the contextual language used on specific stocks. (Tesla in this example)

Methods Used

• Natural Processing Language
• Data lemmatization/stemming
• Data Tokenization
• Data Visualization
• Predictive Modeling

Technologies

• Python
• GetOldTweets3
• Pandas, jupyter
• Numpy
• Matplotlib
• Nltk
• Wordcloud
• Text Blob
• yfinance

Project Objectives

As people tweet about stocks on a daily scale, some things we hoped to discover included:

• What is the overall sentiment of a particular stock?

• Is the overall sentiment correlated to the stock price in anyway?

• What positions do people on average towards the stock?

  things to note:

• The tweets were curated using GetOldTweets3. Twitter's API wasn't used as we found it to be very limited in its capabilities as a free user. This project does open up the question to what the sentiment is like on a grand scale.
• This analysis was dont one 8000 tweets

Key findings

• People overall are very bullish about tesla's stock. However, there is a level of skepticism about how far the stock can climb given the current valuation.
• Most common words included: Call, Split, nice, wow, crazy

Contributing Members

Team Leads (Contacts) : [Samuel Lawrence]: http://samuel-lawrence.co.uk/