- Greetings,

I am exploring a research idea which will make use of a proprietary set

of data on a large number of individual commercial buildings. Due to

restrictions imposed by the organization that collects this data,

information will not be distributed if it allows for the identification

of the individual properties. Thus one can obtain data on aggregations

of the data, but the individual data points themselves. The exception

is that data can be released if it is sufficiently "masked", thus

ensuring the confidentiality of the specific data points.

I have developed what I consider an interesting hypothesis I wish to

test using this data; one that requires the use of spatial techniques

and hence some sort of specific geographic identifier (the raw data

contains lat-long coordinates). Given this, is there some sort of

mathematical translation one can perform on conventional lat-long

coordinates that will disguise the true location of the underlying

data, yet preserve the spatial characteristics of the data?

Thanks,

-Mark - Mark:

This is a common concern with health data. The reference below has a

good discussion of the issues and survey of possible methods. Much

depends on what operations you wish to perform on the masked data.

-- Dave Stinchcomb

Armstrong MP, Rushton G, Zimmerman DL. 1999. Geographically Masking

Health Data to Preserve Confidentiality. Statistics in Medicine 18:497-525.

=============

Mark Coleman wrote:

> Greetings,

>

> I am exploring a research idea which will make use of a proprietary

> set of data on a large number of individual commercial buildings. Due

> to restrictions imposed by the organization that collects this data,

> information will not be distributed if it allows for the

> identification of the individual properties. Thus one can obtain data

> on aggregations of the data, but the individual data points

> themselves. The exception is that data can be released if it is

> sufficiently "masked", thus ensuring the confidentiality of the

> specific data points.

>

> I have developed what I consider an interesting hypothesis I wish to

> test using this data; one that requires the use of spatial techniques

> and hence some sort of specific geographic identifier (the raw data

> contains lat-long coordinates). Given this, is there some sort of

> mathematical translation one can perform on conventional lat-long

> coordinates that will disguise the true location of the underlying

> data, yet preserve the spatial characteristics of the data?

>

> Thanks,

>

> -Mark

>

> - On Mon, 10 Jan 2005, David Stinchcomb wrote:

> Mark:

There was an interesting paper at the AAG last year by Michael Leitner:

>

> This is a common concern with health data. The reference below has a

> good discussion of the issues and survey of possible methods. Much

> depends on what operations you wish to perform on the masked data.

Leitner M. and A. Curtis (2004) Cartographic Guidelines for Geographically

Masking the Location of Confidential Point Data. Cartographic

Perspectives, No. 49, Fall 2004 (in press).

Roger Bivand

>

--

> -- Dave Stinchcomb

>

> Armstrong MP, Rushton G, Zimmerman DL. 1999. Geographically Masking

> Health Data to Preserve Confidentiality. Statistics in Medicine 18:497-525.

>

> =============

> Mark Coleman wrote:

>

> > Greetings,

> >

> > I am exploring a research idea which will make use of a proprietary

> > set of data on a large number of individual commercial buildings. Due

> > to restrictions imposed by the organization that collects this data,

> > information will not be distributed if it allows for the

> > identification of the individual properties. Thus one can obtain data

> > on aggregations of the data, but the individual data points

> > themselves. The exception is that data can be released if it is

> > sufficiently "masked", thus ensuring the confidentiality of the

> > specific data points.

> >

> > I have developed what I consider an interesting hypothesis I wish to

> > test using this data; one that requires the use of spatial techniques

> > and hence some sort of specific geographic identifier (the raw data

> > contains lat-long coordinates). Given this, is there some sort of

> > mathematical translation one can perform on conventional lat-long

> > coordinates that will disguise the true location of the underlying

> > data, yet preserve the spatial characteristics of the data?

> >

> > Thanks,

> >

> > -Mark

> >

> >

>

>

>

>

Roger Bivand

Economic Geography Section, Department of Economics, Norwegian School of

Economics and Business Administration, Breiviksveien 40, N-5045 Bergen,

Norway. voice: +47 55 95 93 55; fax +47 55 95 93 93

e-mail: Roger.Bivand@...