Regression, Scrapers, and Visualization
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
DEPRECATED
Data
Patterns
Prediction
Regression
Screener
Sentiment
.gitignore
AEIS-feb20.png
CODEOWNERS
FB-feb20.png
LICENSE
README.md
figure_1-1.png
figure_1-2.png
figure_1-3.png
figure_1-4.png
figure_1.png
requirements.txt

README.md

Stock Analysis

This repository contains python scripts that I am devleoping to perform analysis on stock prices and visualization of stock prices and other data such as volume.

Some of the goals I want to achieve with this project include:

  • Get the data I need from Yahoo Finance or other API. Able to specify what I need and the time range.
  • Different regression implementations on the close price data. (Linear, SVM, etc.) Possibly try to fit a polynomial function which follows the data.
  • Predicting Stock price for the next day.

Trends

TrendLine.py Results

trendy

trendy

trendy

trendy

trendy

Regression

Using my code for linear regression and Nvidia's (NVDA) stock prices of each day. I got a slope of 0.1850399032986727 and a y intercept of 24.54867003005582. The 50.08 number is the price predicted for the next day based on the linear formula it calculated.

[0.1850399032986727, 24.54867003005582]
50.0841766853

This is insanely accurate. I decided to compare this result with result's I would get with a widely known API, famous for machine learning and other regression tools in python. The API is called sci-kit-learn.

[ 0.1850399]

Using the linear regression offered from the API We get almost the same result, except with less accuracy.

My method outputs a number with greater significant figures.

###Screenshots

Using my regression model:

Linear Regression performed on NVDA Stock dat from January 2016

Using Sci-kit-learn API:

Linear Regression performed on NVDA Stock dat from January 2016

This barely makes a difference to the naked eye. The two graphs are very similar.

New Screenshots

AEIS

AEIS February 20

FB

Facebook February 20

Version

1.0.0 - Released Stock Scraper

1.0.1 - Minor bug fixes with duplicate entries in the CSV File

Todo

  • Use Machine Learning algorithms to predict stock close price for the next day
  • Add data visualization with technical indicators such as moving average, volume, STOCH.
  • Display technical analysis based on stock prices.
  • Add Ratio Analysis & compare ratio with competitors' ratios. (Allow users to define competitors' ratios)
  • Add Stock screener, to screen through every stock and see which ones are best buys.

License

MIT

Free Software, Hell Yeah!