Amid consternation over a defunct census and leeriness towards its less reliable replacement, a group of cities has banded together to build a database of their own.
Fifteen years ago, back when “big data” was chiefly the purview of geeks, one of Canada’s oldest NGOs started a Community Data Program.
The initial scheme, spawned by the Canadian Council on Social Development, was simple: Cities and local organizations from across the country teamed up to purchase statistics they needed but didn’t have the resources to get on their own.
“We recognized there was this shared common purpose around the types of data we were looking for … that postal-code level data that really tells you as much as possible about what’s happening in our neighbourhoods,” said Peggy Taillon, the council’s president. They hammered out a data-sharing agreement with Statistics Canada.
It grew from there: Now, their information-gathering has “expanded well beyond Statistics Canada,” to include other groups – government agencies, pollsters, analytics companies – anyone who’d collect localized data on Canadians.
“What started out as just a group data purchase to make it more cost effective has become a planning body” and a created national commune for “data wonks,” Taillon said. In addition to sharing information they collaborate on collecting, analyzing and applying it.
Now those new partnerships are coming in handy, as the Community Data Program finds itself without the country’s most comprehensive source of stats – the long-form census, scrapped in 2010 in favour of a National Household Survey whose voluntary nature means its responses are less representative. Statistics Canada has said that while it was “generally a success” as a voluntary survey, response rates for more specific populations “may be insufficient to provide a valid statistical picture.”
Those behind the Community Data Program had a series of heated discussions last year when trying to decide whether to buy drilled-down data from this year’s National Household Survey: Was it worth it, knowing the numbers are flawed?
“We had a long difficult conversation last year leading up to, ‘What do we do with this?’ and we decided some data is better than no data,” Taillon said. And, she added, they have enough information already that they’ll recognize any wild anomalies. Some disagreed: The group now has 21 communities, compared to 25 last year. “Some municipalities said, ‘We just don’t think we can sign on to this.”
Red Deer, Alberta is still on board.
Franklin Kutuadu, the city’s research and evaluation coordinator, knows they couldn’t get the kind of cross-tabulated info they need – to study seniors’ housing, for example – on their own.
And now, he noted, “we have a variety of sources now to get data from – we may be getting a much better picture than the total dependence on Statistics Canada alone.”
Taillon and her colleagues hold out hope the long-form census will make a comeback once policymakers realize how badly they need it for just about everything.
“I hope, historically, we can look back at this five-year cycle and say, this was Canada’s quote-unquote ‘interesting’ phase.”
In the meantime, they’re mining partnerships with any other number-tracking group they can find.