[@OKau] Public sector data management report
Steven De Costa
steven.decosta at linkdigital.com.au
Thu Dec 3 02:06:15 UTC 2015
Govt tends to like the term data matching in some use cases. In one agency
dataset you can create a profile of a person based on statistical
algorithms which abstract away personal data. If you use similar algorithms
in another dataset you can then generate a correlation and have a strong
suspicion that two individuals are the same person.
When requesting data on a person from another agency, jurisdiction or
country such profile matching is used to build a case for the request. It
is most often used in matters relating to immigration and security. In
private sector the technique is used between industries like banking,
retail, travel, etc to share customer 'insight' data without breaching
privacy regulation. It all sounds a bit grey hat though.
In the way the report talks about linking dataset's I'd describe it more in
terms of a distributed relational database architecture, or as it mentioned
once, linked datasets.
Someone with a formal degree in this stuff would likely have a better
Data literacy is an important area to focus on in all walks of life. As
Linked Data has begun to enter the mainstream conversation around data
management systems the conversational use of those two words in other
contexts will create a fair bit of confusion. It is hard to capitalize
letters when speaking to an audience...
Very cool to see the report though. Lots of good thinking and leadership
coming out of the public sector right now.
On Thursday, December 3, 2015, Rosie Williams <BudgetAus at hotmail.com> wrote:
> I had to watch a video on YouTube the other day as I had a feeling Linked
> Data meant more than using 2 or more data sets together. What is the term
> to refer to using two or more data-sets with a shared field? This is an
> area that needs some work from government to ensure that there are specific
> fields like ABN included where-ever possible as the best means to match
> data across data-sets. Matching on name or address is pretty useless,
> especially wrt the quality level of those fields.
> Rosie Williams BA (Sociology)
> NoFibs.com.au <http://nofibs.com.au> - Open Data Reporter | OpenAus
> <https://openaus.net.au> - Founder and Developer
> *From:* okfn-au <okfn-au-bounces at lists.okfn.org
> behalf of Steven De Costa <steven.decosta at linkdigital.com.au
> *Sent:* Thursday, 3 December 2015 11:51 AM
> *To:* Open Knowledge discussion list for Australia.
> *Subject:* Re: [@OKau] Public sector data management report
> Interesting that the report defines Linked data as:
> Data created from matching and integration of two or more datasets. This
> may occur through either an explicit match on unique identifiers, or
> through a combination of information that gives a high confidence match
> between the datasets
> Or, is is a shame that 'linked data' is so easily used conversationally
> around data when it has a pretty specific meaning for those interested in
> it with regards to web 3.0 technologies.
> I guess when we are really talking about Linked Data we need to capitalise
> both words :)
> Ref: https://en.wikipedia.org/wiki/Linked_data
> *STEVEN DE COSTA *|
> *EXECUTIVE DIRECTOR *www.linkdigital.com.au
> On 3 December 2015 at 10:49, Rosie Williams <BudgetAus at hotmail.com
>> Rosie Williams BA (Sociology)
>> NoFibs.com.au <http://nofibs.com.au> - Open Data Reporter | OpenAus
>> <https://openaus.net.au> - Founder and Developer
>> okfn-au mailing list
>> okfn-au at lists.okfn.org
>> Unsubscribe: https://lists.okfn.org/mailman/options/okfn-au
*STEVEN DE COSTA *|
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the okfn-au