Mary Agnes Welch had a fun story in Saturday’s Freep about popular baby names in Manitoba. Working with data provided by the province, MAW and online editor Wendy Sawatzky produced a customized google map showing the most popular names for each postal code in the province.
MAW followed up on the story with a blog post in which she noted how it only took one phone call to get the data from the province, and said “that’s how open government is supposed to work.”
I appreciate that tracking down this data took considerably less effort than many requests for government information — see the recent “Open Secrets” series for some context — but to me this still isn’t open government.
In a truly open government, information like baby names — and perhaps especially baby names since it’s about the most innocuous issue around — should be available in a public data set for anyone to download and analyze, parse and repurpose as they see fit. The same goes for playground safety reports, traffic ticket volumes, penitentiary pudding budgets or transit ridership numbers.
Forget about helpful bureaucrats who are responsive to phone calls, there shouldn’t be any need for the phone call in the first place.
This data belongs to the public, not just the politicians and the bureaucrats. The default should be to make it publicly available unless there’s a compelling legal reason to do otherwise.
Transit info, spending info, etc… all that should be public. However there is virtually an unlimited variety of information like baby names, and ways to cut and slice that information (by postal code, etc) that I can’t see phone calls going away completely. There is some cost to maintaining datasets online for people to access, and you can’t always predict what someone might want to research. I would make the unimportant stuff like baby names a low priority.
I would agree with Cherenkov about this one.
Guys, I think you’re missing the broader point, which is that most information (regardless of its level of importance) should automatically be made available to the public as a raw dataset.
Obviously the “baby names by postal code” was already sortable in the government database, as I’m certain that someone didn’t do that by hand in the course of 24 hours.
And yes, there is a cost to maintaining pubic data datasets. Just as there is a cost to maintaining private datasets. And a cost for responding to informal and formal (FIPPA) information requests.
In this particular case, there were probably at least three staff involved (comm guy, stats mgr, stats grunt) in facilitating the request. Multiply that by the thousands of internal and external requests for info that the province receives each year and it’s a safe bet that hundreds of positions exist just to serve this closed system.
However, if open data became the default for government then the processes and costs for stripping out truly sensitive information (e.g. names) and making the rest publicly available would eventually be reduced considerably.
That would really involve maintaining two separate banks of data, some of which may only rarely be used. For example baby names would be in one table which also contains confidential information, and postal codes would be in another table containing other information. You would need to create parallel tables, stripped of the confidential information, and post those on line, AND keep them synched with the primary data sets. Then the end user would need to have the software and know-how to join the tables in a data model to pull the information that they need. They may very well end up calling anyhow for help, or end up with results that are incorrect because they screwed up their data model. It will cost more and provide worse service to the public.
Like I said, the important information should be scrubbed and put online in usable form, and the rest can be left to a government data jockey who presumably knows the data and how to pull it.
B.C. is already on this: http://www.data.gov.bc.ca/