Michael Smith, Quan Gan, and Robert Kohn, "Modeling Dependence using 
Skew t Copulas: Bayesian Inference and Applications", Journal of Applied 
Econometrics, Vol. 27, No. 3, 2012, pp. 500-522.

1) Electricity Spot Price Data (Section 4)

The price data in this example were sourced from the Australian Energy
Market Operator (AEMO), which currently publishes their price data online at
http://www.aemo.com.au.  Averages were calculated manually, as were day type
indicators. The data are stored in both ASCII and *.csv formats in the file
"sgk-electricity-prices.zip". Unix/Linux users should use "unzip -a".

The data are arranged in files by state (NSW, VIC, QLD, TAS & SA) and
contain an observation on average daily log-prices, the date and day type
dummies corresponding to public holidays and Monday to Saturday (Sunday is
the base case). The dummies are labelled D1, D2,... etc. with the days
listed in the files "Dummies for XXX.txt" where XXX is the state.

The data run from 1 Jan 2007 to 31 Jan 2010, giving a total of 1127
observations.

2) Website Exposure Data (Section 5)

The data were collected by ComScore Networks and are proprietary. They are
made available only by subscription through the Wharton Research Data
Service (WRDS), and details on how to purchase a subscription can currently
be found online at http://wrds.wharton.upenn.edu/.

The dataset itself was extracted from the "Sessions" data in the 2007
"comScore Web Behaviour (panellist-level)" Database. We restricted attention
here to all machines that had recorded internet activity at our chosen
websites on 1 May 2007, resulting in 50,217 observations, each corresponding
to a particular machine out of the more than 90,000 machines in the panel.

Each observation has 15 variables, each of which corresponds to the 15
websites listed in Table 4. The variables are the total number of pages
viewed during 1 May 2007 for that website. They are created by adding
together the "page_viewed" variable for all visits during that day. For most
margins this will usually be exactly zero.

For example, if during 1 May 2007 the following visit history is recorded
for a specific machine:

a)	myspace.com with page_viewed = 3
b)	myspace.com with page_viewed = 23
c)	google.com with page_viewed = 3
d)	myspcae.com with page_viewed = 12

Then the observation for that machine has a count of 38 for myspace.com, 3
for google.com and 0 for the other 13 websites.