Database Reference
In-Depth Information
these individuals. In a dataspace approach, however, aggregate data can be used,
so the database may contain the total number of suspects questioned by the police
(aggregated from HKS) and the total number of suspects (cases) handled by the
prosecution (from OM-data). Then, a comparison can be made between the two
totals, and the difference between output and input can be calculated. This task
can be performed easily by the public safety monitor (described in Subsection
10.3.2). Thus, for this type of questions, a dataspace is more efficient as the heavy
computational and troublesome task of uniquely linking individuals does not have
to be performed.
10.5 Protecting Privacy When Combining Judicial Data
Tools or information systems that collect, relate, and present safety-related data,
pose a serious privacy threat as the identity of individuals or groups of individuals
may be exposed. For instance, assume that in the public safety monitor (see Sub-
section 10.3.2) the number of sex offenders is presented, and that it is possible to
categorize them by age, gender, and city. If there is only one female sex offender
in a certain city, then the age of this female is exposed. Depending on the addi-
tional information that is shown about her, or the information that can be gathered
from alternative sources, it is likely that her full identity is exposed. If this is in-
deed the case, privacy laws are violated.
In the data integration process several precautions can be taken to respect the
privacy of individuals and to minimize the risk of exposing someone's identity.
First, a data source that contains crime data should only record attributes that are
in line with the Dutch Personal Data Protection Act (PDPA). This act defines a
set of sensitive attributes that should be handled with care, namely data on some-
one's religion or life conviction, ethnic origin, political opinions, health, sexual
orientation, and memberships of (trade) unions. 12 Such sensitive attributes should
not be stored. Second, aggregate data has a clear advantage over microdata as da-
ta on a higher aggregation level does not provide personal information. There-
fore, for privacy reasons, it is recommended to use aggregate data instead of mi-
crodata when possible. Finally, whenever there is a risk of exposing the identity
of an individual to a user of a tool, the result of the user's question or selection
should not (or only in part) be shown. For example, if a user wants to view the
number of sexual offenders per region, and if there are just two offenders in a
certain region, this number should not be presented to the user. After all, in this
case there is a reasonable chance that with additional information, the identity of
the offenders concerned can be deduced. When all three precautions are fol-
lowed, the risk of disclosing personal data and thereby violating the privacy of
individuals is minimized.
The preceding sections focused on ways to combine and integrate data from
various judicial databases. Combined crime data may help in gaining insight into
the criminal law chain and in developing new policies. An even deeper under-
standing of crime and delinquency may be acquired by applying data analysis
12
Sauerwein, L.B. & Linnemann, J.J. (2001).
Search WWH ::




Custom Search