For beginners who are completely new to the field of data modeling and time series analysis, this should serve as a good introduction to get started with Time series analysis.

Time series is a type of data that is collected and measured over time. Examples of time series data could be:

1. The GDP values of a country represented over a decade or

2. The change in room temperature ever hour for the last 24 hours or

3. The monthly rainfall for the year 2013 or

It could be anything that is represented for any unit of time be it second, minutes, hours or years together.

So why is Time series modeling given so much special attention? There is a lot of research and analyses performed around time series that it forms a whole domain and people specialize in it.

That I think is primarily because it has to do with the discipline called ‘forecasting’.

Businesses in every industry out there has to forecast its sales, predict the demand for the next year or two before it can even start planning on the resources and investments.

Don’t get me wrong here. Forecasting does not apply exclusively to product sales or stock prices. It could be literally anything, be it predicting the weather, phone call traffic in a call center,

global population growth in the 21st century. Get the picture?

So how do we go about forecasting them? Some may think, irrespective of the complex procedures you may apply there is always an element of astrology to every forecast-able event.

If you were one of them, Well that is a common way to think about it. In most cases, we may not be able reach 100% point accuracy, but we will be much informed and well equipped to mitigate the risk.

Now R has a wide variety of facilities available for handling time series data and make predictions. In fact, most latest research get published in R first before being implemented elsewhere.

Here is some operations you will need to perform time series modelling.

1. Creating Time Series Data

2. Accessing time series data for public data bases.

3. Packages relating to Time Series Modeling

4. Seasonal Decomposition

5. Exponential Models

6. Holt Winters Models

7. ARIMA Forecasting

8. Automated Forecasting

For a deeper theoretical understanding, if at all you plan to invest in a book, this is the one you should go for: Forecasting: principles and practice by Rob Hyndman. This is the gold standard and the best book for a complete and clear understanding of forecasting principles.

If you are looking for more of an application oriented book, Forecasting: Methods and Applications is the one.