Dennis Fok and Richard Paap, "Modeling Category-level Purchase Timing 
with Brand-level Marketing Variables", Journal of Applied Econometrics, 
Vol. 24, No. 3, 2009, pp. 469-489.

The data we use are based on the so-called ERIM database, which is collected
by A.C.Nielsen. The data span the years 1986 to 1988, and the particular
subset we use concerns purchases of detergent by households in Sioux Falls
(South Dakota, USA). For our purposes, the data are aggregated to the brand
level.

We select households making at least 4 purchases and buying only brands that
are available throughout the complete sample. The first selection is not yet
applied to the data files we make available.

There are four data files, all of which are zipped in the file
fp-data.zip. All file are ASCII files in DOS format. Unix/Linux users
should use "unzip -a".

Description of data files

**purch.txt (5059 records)
Contains household level purchase histories
Variables:
id	  	household identifier
date		date of purchase
vol		purchased volume in ounces
brand		brand number

Brand coding
1=Cheer
2=Oxidol
3=Surf
4=Tide
5=Wisk
6=Rest

**hhchar.txt (516 records)
Contains household characteristics
Variables:
hh id		household identifier
hh income	household income category
hhsize		number of household members
insample	0/1 variable indicating whether this household was used for
		the in sample data
outofsample	0/1 variable indicating whether this household was used for
		the out of sample data

**prices.txt (7541 records)
Store level price and promotion data
Variables:
sid	  	store identifier
week		week number (week=floor(15/14 + (date-9490)/7)
brand		brand number
price		price of brand (aggregated across SKUs)
display		display variable (aggregated across SKUs)

**weights.txt (5128 records)
Importance of a store for a household
Variables:
id		household identifier
sid		store identifier
weight		importance weight