Shiny app: Forecast of GOOG stock

Andreas Rimbe
June 21, 2015

Introduction

This is a simple shiny app that forecast the opening stock price of Google's stock (GOOG).

Screenshot of app

The app can be found here: Forecast of GOOG stock

What does the app do?

The app plots a timeseries of Google's opening stock price for a selected date range.

Parameters that can be adjusted:

Date range - The date range to forecast the stock

Moving average order - a slider to the order of the moving average (using SMA from TTR package)

Number of years to forecast - a slider to set the number of years to forecast. The forecast is created by a simple exponential smoothing predictive model using HoltWinters. The darker blue indicates the 80% prediction interval and the lighter blue the 95% prediction interval.

Moving Average Trend

The plot shows a moving average, which is a smoothing method to estimate the trend. The SMA() function in the TTR package is used to calculate the moving average.

Moving average is used to analyze data points by creating a series of averages of different subsets of the full dataset.

Exponential Smoothing Predictive Model

An exponential smoothing predictive model is used to make short-term forecasts of the time series.

The predictive model is fit using the HoltWinters function.

We then use the forecast.HoltWinters to make forecasts for future time points. The h parameter is used to set the nuber of years in the future we want to to forecast.