Market Analyzer

Stock Market Forecasting Application

Scott B. Norin - Coursera/JHU Data Scicence Specialization
Developing Data Products - 22 February 2015

Market Analyzer Application

    "A fool and his money are soon parted" 
      Variation Thomas Tusser in Five Hundreth Pointes of Good Husbandrie, 1573
  • Utilize Advanced Methods Available in R/Rstudio for Forecasting.
  • Implements Shiny Apps development website / used in conjuction with RStudio.
  • All methods shown are for informational purposes only and are not recommended for actual investing.

Instructions :

  1. Go to link [https://rigelfive.shinyapps.io/MarketAnalyzer]
  2. Input a stock symbol in the text field.
  3. The shiny based application acquires stock market data from Yahoo via the quantmod package in R.
  4. Three plots are generated with historical data, a forecast, and a cyclical analysis.

Historical Chart

plot of chunk unnamed-chunk-1

  • This stock chart is generated using the quantmod package and has a variety of features.
  • OHLC Candlestick charts are shown with a volume graph on the bottom.
  • Chart defaults to the S&P 500 exchange traded fund (ETF) with the symbol: SPY

Market Forecasting

plot of chunk unnamed-chunk-2

  • Forecast written by Rob J Hyndman and George Athanasapoulos prenented by J Leek in Practical Machine Learning - Data Scienece Specialization Coursera
  • Additional methods presented at link https://www.otexts.org/fpp
  • Pursue integration of forecast into historical chart data.

Predict Dynamic Market Behavior

plot of chunk unnamed-chunk-3

  • Loess method (STL) is more robust versus 'decompose' function (chops start/end).
  • Cyclical trading enables opportunities to be able to swing into/out of the market.
  • Caution: Cyclical dynamics may change over time --> invalidates strategy/forecast.