8.3 Data Preparation

This section describes how to prepare data so that it has a time-series component and appropriate dummy variables. The scripts explained in the video are attached here:

Deriving data from dates

Data comes in a variety of forms and must often be manipulated to prepare it so it can be used for time series analysis.

Sometimes data is in daily form and it is necessary to create the daily time index from complete dates (mm/dd/yyyy). Dummy variables for months and days of the week also need to be extracted from the dates. This video show how to prepare this data in JMP.

Sometimes data is monthly data and the dates are listed as the first or last day of the month, e.g., 5/1/2012 or 5/31/2012. In these cases, the day of the month is not important. It is just a placeholder so that a date can be provided for the month. But the month and the year must be extracted to produce a monthly time index. This video shows how to accomplish this in JMP.

Deriving Data from non-dates

Sometimes the data we receive for time series does not come as traditional dates. For example, we might have only years and quarters and the outcome variable. The video below gives examples of how to prepare this type of data for time series analysis.