Roman Liesenfeld, Jean-Francois Richard, and Jan Vogler, "Likelihood
Based Inference and Prediction in Spatio-Temporal Panel Count Models
for Urban Crimes", Journal of Applied Econometrics, Vol. 32, No. 3,
2017, pp. 600-620.

There are four ASCII (text) data files. Columns are always
tab-delimited. 

Throughout this text, N=138 denotes the number of census tracts
103--9822, T=72 the number of months spanning the time period
01/2008--12/2013, and K=15 the number of socioeconomic regressors. 


The file 'part1.txt' contains the number of Part I offenses stored as a 
N x (T + 1) matrix. The columns 1:T+1 contain the following variables:

Column		Name
1		Census Tract ID
2		number of Part I offenses in 01/2008 
3		number of Part I offenses in 02/2008
.		.
.		.
.		.
73		number of Part I offenses in 12/2013


The file 'part2.txt' contains the number of Part II offenses stored as a 
N x (T + 1) matrix. The columns 1:T+1 contain the following variables:

Column		Name
1		Census Tract ID
2		number of Part II offenses in 01/2008 
3		number of Part II offenses in 02/2008
.		.
.		.
.		.
73		number of Part II offenses in 12/2013


The file 'socio.txt' contains the socio-economic variables stored as a 
N x (K + 1) matrix. The columns 1:K+1 contain the following variables:

Column		Name
1		Census Tract ID
2		total population
3		population density per square mile
4		median income
5		dropout rate age 16-19
6		civilian unemployment rate
7		poverty rate
8		percentage of total population under 18
9		group quarter proportion
10		percentage of total population that is African-American
11		percentage of population (age 25 and over) with less than
                a high school degree
12		percentage of population (age 25 and over) with a bachelor
                degree or higher
13		rental housing units as a percentage of occupied housing units
14		percentage of households having been in the same house for
                more than one year
15		percentage of female headed household
16		housing units vacancy rate

The file 'W.txt' contains the spatial weight matrix W. The first row
and column contain the census tract ID so that there are (N+1) x (N+1)
entries in that file.

All the data are also stored in the MATLAB file 'lrv-data.mat'.

The MATLAB file 'Likeli_Spatial_Panel.m' contains the EIS code for
evaluating the likelihood.

All files except lrv-data.mat are zipped in the file lrv-files.zip.
Since these files are in DOS format, Unix/Linux users should use
"unzip -a".

Please refer to the paper and the online appendix for further details
regarding the variables and the EIS code.