'''
* @What Data Mining Final Project : Using Tweeter sentiment analysis to predict change in stock price.
* @Author Sabastian Mugazambi & Simon Orlovsky
* @Date 05/27/2017
* @Purpose : Analyses a tweet and predicts the effect of the tweet on the mentioned company stock price.
'''
List of Files
_____________
1. main.py - main file to run
2. tweet_dumper.py - gets tweets to be used / already done for you
3. stock_prices.py - gets stock prices
4. prediction_knn.py - predicts the price change using knn
5. get-pip.py - Used for installing pip
csv files are used for data import.
Dependencies
_____________
There are a couple of libraries that our code relies on and it is essential to have these libraries installed before running the program.
- Tweepy :
1. install pip using the command line
$ sudo python get-pip.py
2. install Tweepy
$ sudo pip install --ignore-installed tweepy
- Quandl :
1. install Quandl
$ sudo pip install --ignore-installed quandl
Make sure that these dependencies are installed for successfully running the
main program.
Running the main
_________________
- In the command line terminal navigate to the directory where the files exists
Run main.py as follows:
$ python main.py <k> <Company Ticker> <Twitter UserID> <Tweet Text> <Date>
Parameters Explained
____________________
<k> : The number of nearest neighbors to used predict price change
<Company Ticker> : Official stock exchange ticker symbol
<Twitter UserID> : The official UserID of the individual tweeting
<Tweet Text> : The tweet text with the company mentioned in quotes
<Date> : Date of the tweet in the format YYYY-DD-MM with no quotes
Example
____________________
$ python main.py 5 F realDonaldTrump 'Totally biased @NBCNews went out of its way to say that the big announcement from Ford, G.M., Lockheed & others that jobs are coming back...' 2017-01-18
Output Explained
_________________
➜ $ main.py 5 F realDonaldTrump 'Totally biased @NBCNews went out of its way to say that the big announcement from Ford, G.M., Lockheed & others that jobs are coming back...' 2017-01-18
------------------------
<Displays the tweet entirely>
The tweet is:
' Totally biased @NBCNews went out of its way to say that the big announcement from Ford, G.M., Lockheed & others that jobs are coming back... '
------------------------
<Displays the predicted sentiment of the tweet understood as the effect it has
on stock price >
Predicted Sentiment Classification : negative
------------------------
<Displays the predicted percentage change in price>
Predicted Percentage Change in Price : 0.00819809018105
------------------------
<Displays the predicted price EOD of the company in question and the actual EOD of that day>
Predicted Price Due to Tweet: 12.5066220828
Actuall EOD Price: 12.41
------------------------
Have fun!