[Open-data-census] Open Data Census Canada

Sat Oct 5 18:50:43 BST 2013

I echo your sentiments, concurrently, best not to mislead with numbers and
paint an erroneous picture. We ask that our politicians and journalists not
do this, and best we too stick to some higher standards. Again, as stated
in earlier streams, the responses to the questions need to be qualified to
reflect the Canadian data reality. Many other jurisdictions have also
discussed these issues and have stated that they too are having difficulty.

The oknf folks, understand this and are continuously improving the tool.
 It is only by continuously communicating will we improve the method, and
better reflect local realities.

In addition, others in Canada who are statisticians and indicator
developers agree with the difficulty of a federated system and the nature
of these indicators. Also, we as open data advocates also need to develop
interprovincial and territorial collaborations, as our open data map has
some blank spots. And the Feds are not the sole arbiters of knowledge in
Canada.

We are all making progress, but I cannot in good conscious report without
qualifications, and readers of the results should be able to compare apples
and apples, and if that is not possible then they need some footnotes to
the results. That is what I and other editors are working on. We need to
trust the numbers and also, be able to advance an issue.

cheers
T

On Saturday, October 5, 2013, David Eaves wrote:

> I just want to echo James comments, in part because from advocacy
> perspective it will be much easier if we restrict to just the federal
> government. Part of this stems - I believe - from some confusion about the
> purpose of the Open Data Census.
>
> The goal - as far as I've been informed - of the open data census is not
> to create a comprehensive list of open data sets around the world for data
> users. The goal is to create a mechanism by which people can pressure on
> various governments by either shaming their performance on the list or by
> pointing to performance of other governments.
>
> If we conflate Federal and provincial and municipal data we will
> essentially make the census useless as such a tool. Each jurisdiction will
> be able to claim that the "bad performance" on the census is a result of
> the failure of either another level of government (e.g. a high or lower
> level of gov), or by the poor performance of a peer. In addition, the
> census methodology does not support, or make clear, about what to do when
> say BC has the data open but Ontario does not. Such a census will be easily
> ignored by the Feds as well as the province.
>
> I'd rather have something we can explicit beat the feds up with and,
> because there are big "N/A" blocks that will be clear that it is not their
> fault (so we won't get into distracting methodology conversations) - but
> these N/A will still will make us look like poorer performers against some
> other countries - this will make it easier to enlist the feds support to
> convene and engage the provinces and, if we are lucky, cities, to make
> these data sets open.
>
> So, in sum, the point of the census is not to figure out what a "pure"
> methodology should be. It is to think strategically about how it maximizes
> our impact as open data advocates by facilitating an influence strategy.
>
>
>
>
>
>
>
> On 2013-10-04, at 5:54 PM, Tracey P. Lauriault <tlauriau at gmail.com> wrote:
>
> I am not James, I am annotating and qualifying the answers.  However, when
> it comes to transport we will have to say no, when it comes to stats we
> have to say some, when it comes to elections we have to say yes but qualify
> that we do not know about ....
>
> And so on.
>
>
> On Fri, Oct 4, 2013 at 10:19 PM, James McKinney <james at opennorth.ca>wrote:
>
> The results will only be false if we are unclear about what the results
> are about. If we make it clear that the results are about the Canadian
> federal government, then the results at least have a chance of being
> correct.
>
> If we try to make it about more than the Canadian federal government, then
> I think we should just give up now, because the answer to each question
> would have to be "Yes" AND "No" since there is not one answer that is true
> for all jurisdictions in Canada.
>
> I think the solution, as I've said before, is to (eventually) add the
> subnational levels to the Census. You cannot smash all of Canada's
> jurisdictions together and produce a coherent answer to any of the Census'
> questions. You need one answer per question per jurisdiction.
>
> It is not sound methodology to try to smash all the jurisdictions
> together. My proposal is to be clear about what we're measuring (the
> Canadian federal government) and to then measure it alone.
>
> James
>
> On 2013-10-04, at 4:48 PM, Tracey P. Lauriault wrote:
>
> I sorta agree with you James, but then the results are false, for example,
> health data, wel that aint the feds, demograhic data are the feds and the
> provinces, transport is only the municipalties, pollution is multi
> jurisdictional as well, so we can, but then that is not our country is it.
>
> Furthermore, one of the arguments I have about the OGP is this national
> approach and the lack of subnational.  In canada it is really the
> subnational doing the heavy lifting, and they do not get a seat at the
> table.
>
> There really needs to be some serious thought, the OECD, WB, UN and
> others, EU, work really hard at the creation of indicators that make sense
> across jurisdictions, and that is the hard work that is not being done
> here.  My issue, is, if we are to do evidence based planning and decision
> making with data, then the methodology we apply has to be sound, otherwise,
> we are telling an erroneous story and misleading policy makers.
>
> I hope open data is also about sound methodology and not just about apis
> or open formats, it is in my mind about data for deliberation and evening
> out the analytical and debate playing field, and we as 'citizen' social
> scientists, will need to be a bit more rigorous, as with time, what we say
> will not stand.
>
> Anyway, I will be editing this weekend, I have not really heard from my
> co-editors much, i wanted to collaborate with them on this as it would make
> more sense if we could discuss things a bit more.
>
> I hope you all can re-look at what I edit one more time before we consider
> it final.
>
> Cheers
> t
>
>
> On Fri, Oct 4, 2013 at 7:26 PM, James McKinney <james at opennorth.ca> wrote:
>
> I think we should just be filling this in with respect to the federal
> level. So, for example, the answers to the election results questions
> should be with respect to federal elections only - each question should not
> attempt to answer for all elections at all levels of government in Canada.
> If the OKF is interested in the subnational level, we should have a
> separate entry in the Census for each province and territory (and each city
> if we really want to go all out). I've answered below in keeping with this
> understanding.
>
> I can't edit most of these because they are "pending review." Who reviews?
> Can I be a reviewer?
>
>  I filled in the Census for National map, pollution, postal code and Elect
>
>

-- 
Tracey P. Lauriault
http://traceyplauriault.wordpress.com/2013/07/23/moving-to-ireland/
https://gcrc.carleton.ca/confluence/display/GCRCWEB/Lauriault
http://datalibre.ca/
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.okfn.org/pipermail/open-data-census/attachments/20131005/5d8536f8/attachment.htm>