Information & Ethics Committee on Feb. 14th, 2011

[Chantal Bernier]

That concern is more and more being voiced under the phrase “the ethics of analytics”.

We realize that we have more possibilities for not only amassing personal information but for generating, through aggregation of the information received, an even more specific profile of a person. We realize that we truly need to have an ethical framework for that.

As we discover new models for, say, energy rates and the consumption of electricity, we realize that some people would actually like to have billing tailored to their personal consumption, which could reveal very personal information about usage. Combined with other information, it could actually draw a rather intrusive profile of an individual. That needs to be addressed, absolutely, through a proper ethical and legal framework to ensure that the principles of privacy we still adhere to—even in this new context of information technology—are respected.

[Patricia Davidson, Conservative, Sarnia—Lambton, ON]

We had the Information Commissioner before us a couple of times last year, or maybe more than that. I know she was here in April and again in November. She talked about open government initiatives and about the initiatives that have been taking place in the United States, in the United Kingdom, and in Australia. These are areas we are still going to hear from. We haven't at this point.

Are you aware of any specific issues those countries have experienced? If you are aware of specific issues, are there recommendations we can take into account as we're working our way through this report so that we do not fall under the same issues they had?

[Chantal Bernier]

I don't have any study in mind specifically. We would be happy to get back to you on that. Of course, each country has different contexts. There was a specific case that you will remember, the debacle of the British MPs' expenses. What was interesting was that the decision on what should be revealed was predicated upon the system in Great Britain for the management of the accounts or the supervision of the expenses.

I think we need to make some adaptation from country to country. Right now I do not have in mind the results of some studies, but if that is something the committee is interested in, we could get back to you on it.

[Patricia Davidson, Conservative, Sarnia—Lambton, ON]

Yes, I think we would be interested in that.

During your opening remarks, you also cautioned us a bit about Internet postings of tribunal decisions. Could you elaborate a bit more on that, please?

[Chantal Bernier]

Certainly.

I would actually take you back first to a more general statement by the Chief Justice of the Supreme Court of Canada, Beverley McLachlin, who herself has brought forward the idea that we have now such a context of dissemination of information that the open court principle has taken a meaning that goes way beyond what was originally envisaged.

The open court principle is there to shed light on the court--not necessarily on the parties--and to hold the court accountable. The situation now is that the posting of decisions on the Internet actually sheds light on the parties, and when you bring this specifically to the context of federal administrative tribunals—tribunals that deal with very personal and sensitive information, such as disability, grievances, and discrimination—the posting on the Internet may actually bring a cost to privacy that goes far beyond the public interest served.

I would like to share with you a simple anecdote. I received an email this summer from someone asking me for advice, saying that they had just found out that their grievance to a federal administrative tribunal was about to proceed and that they understood the decision could be posted on the Internet. Since they had alleged issues of discrimination, they were afraid that if indeed the decision were posted, it could hurt their career in the future. The person asked if I thought they should drop the grievance in light of that. Of course my answer was that I could not take that decision for them, but they should assume that it would be posted on the Internet, because we have not been successful yet in bringing the kind of discretion we would like to see.

The reason I share this anecdote with you is that I thought it brought to the fore the possibility of inhibiting access to administrative justice for the reason that the Internet has brought a differential impact on privacy that goes beyond what was originally anticipated or envisaged.

[The Chair, Liberal, Shawn Murphy]

Thank you very much, Ms. Davidson.

That concludes the time allocated to the first round. We are going to ask you, Madame Bernier, if you have any closing comments you want to leave with the committee. Then we're going to suspend for a minute to invite the second panel. We'll start with Mr. Easter when we return.

Madame Bernier, do you have any closing comments you want to leave with the committee?

[Chantal Bernier]

I believe I will simply summarize by again saying that I applaud this committee for addressing this issue. It is highly topical, and I believe we need to be seized of the privacy considerations toward open government to ensure that as we further the goals of transparency, we still maintain the deep value we afford to privacy.

Merci beaucoup, monsieur le président.

[The Chair, Liberal, Shawn Murphy]

I want to confirm something. I believe there was an undertaking given to Ms. Davidson that if there are any international materials, you're going to provide them to the committee.

Would providing those materials within three weeks be appropriate, Mr. McKay?

[Chantal Bernier]

I will do that with pleasure.

[The Chair, Liberal, Shawn Murphy]

Okay, I appreciate that.

On behalf of all members of the committee, I want to thank both of you for appearing today. Your testimony was very valuable. On behalf of the committee, please pass on our best wishes to Madam Stoddart. Hopefully she will get better soon.

[Chantal Bernier]

Thank you very much.

[The Chair, Liberal, Shawn Murphy]

We're going to suspend for one minute and then we'll start the second panel.

[The Chair, Liberal, Shawn Murphy]

I call the meeting back to order.

This is the second panel of witnesses we have today to deal with the ongoing study in open government.

The committee is very pleased to have, first of all, Madame Tracey Lauriault. Her resumé has been circulated. It's very extensive. She's done a lot of work on this particular issue.

We also have Mr. David Mason, executive director of Visible Government, who is also well informed on this particular issue.

On behalf of all members of the committee, I want to welcome both of you to the committee today. We're going to go until about 5:15, so we have about an hour. If you wish, I invite you to provide the committee with any open comments you have.

We'll start with you, Madame Lauriault.

[Tracey Lauriault, As an Individual]

Good afternoon, everybody, and of course happy Valentine's Day. It's great to be talking about data on such a wonderful day.

I received a great homework assignment from your clerk, and most of that homework is available to you in my submission, which I hope you've received.

I won't be going through everything that's in the submission, however. As an overview, I've introduced to you what civil society groups are about and what they do. I've introduced you to two civil society organizations, namely the Community Data Consortium and the Federation of Canadian Municipalities' quality of life reporting system. I've also introduced in that submission the Geomatics and Cartographic Research Centre, which is an official university research centre that uses quite a lot of data, but also produces data.

Then I talk about why open government is important. I provide some of the issues. I discuss which public data should be made available according to the perspective of community-based research groups, how the federal government can move towards more open government data policies, and ways to consult with users. Finally, I provide some recommendations.

I'm not going to focus on all of this today, but on the community-based organizations, on the research organizations, and on issues and recommendations.

I will begin with community-based organizations. There are thousands of them across Canada, and these organizations are heavily involved with doing work critically important to civil society. Some of their work involves such things as helping the homeless or working on issues of food security, as well as urban planning, education, population health, etc.

The Community Data Consortium is an organization that group-purchases Statistics Canada data on a consortium type of licence. It would do so otherwise, except that as we all know, Statistics Canada data is cost-recovered--and therefore very expensive--and has very exclusive and restrictive sharing licences. Therefore, they've had to form this consortium so that they can share between and among themselves, build a data-sharing type of entity and infrastructure, and develop capacity-building on how to use public data to inform their users.

The Community Data Consortium consists of 17 data consortia from 50 municipalities, cities, and regions across the country. It has 850 members, which includes school boards, police forces, counties, cities, large metropolitan areas, the United Way, social planning councils, and community health and resource centres, just to name a few of the 850. In here we have community-based researchers who use all kinds of public data from multiple government institutions, primarily from Statistics Canada, to do evidence-based decision-making at the local scale. These groups use these data to inform human services plans, poverty reduction strategies, sustainable development and environment, population health, etc., and as I discussed, they do so through this infrastructure called the Community Data Consortium.

What's important to these groups is to have data that are aggregated at the level of the community, so we're talking about neighbourhoods, health districts, city wards, etc. We're talking about a sub-municipal scale. The reason is that when you start looking at trends and patterns at the community scale, you can focus better and better target your efforts to meet the mandate of the variety of the civil society organizations you represent. I've given you a list of those organizations.

The quality of life reporting system, which is produced by the Federation of Canadian Municipalities, includes 24 cities in seven provinces across Canada. It collects data from CMHC, Industry Canada, Citizenship and Immigration Canada, Environment Canada, Elections Canada, the private sector, provinces, and NGOs, as well as collecting administrative data from the 24 cities involved in this project. They produce an indicator system that crosses 10 domains, such as demographic, civic engagement, community infrastructure, education, environment, etc., just to name a few.

They also have a great tool call the “municipal data collection tool”. They have an official in each city who scurries through their respective municipal institutions to find data related to homelessness, to housing, to recreational facilities, to the quality of the public parks and swimming pools, and so on. They find out much it costs to take a bus in your city and what the issues related to social assistance are. They find out if people can afford those things.

They also have a data visualization tool they're going to be releasing in the summer. As you are probably all aware, they produce a number of really important thematic reports nationally that also have local flavours in the 24 cities. While we can have national platforms on housing and homelessness, immigration, and social infrastructure, we know that there are particularities in each city that differ. Calgary is not Vancouver, and it is not Halifax. However, there are some national strategies that these reports inform.They use data as evidence to inform a variety of those issues.

There's also the Geomatics and Cartographic Research Centre. It's an official research centre at Carleton University. The focus of that research centre is primarily open source interoperability, cybercartography, archiving, and preservation. It also produces atlases--multimedia, multi-sensory, dynamic, and engaging types of atlases--on a variety of issues, such as indigenous knowledge, aboriginal peoples in the north and treaties, the risk of homelessness, and a variety of other issues.

This research centre gets its resources primarily from public funds. They therefore believe that the outcomes of their research belong to the public. Therefore, their atlases and the committees within which they conduct their work focus on using open data whenever possible. If they produce data, they share those data in open formats and under open licences. They use and develop open-source technologies, they develop interoperable technologies, and they distribute the technologies they produce and the products they use under open BSD types of licensing.

These three organizations--the Community Data Consortium, the Federation of Canadian Municipalities quality of life reporting system, and the Geomatics and Cartographic Research Centre--represent about 1,000 data researchers across the country who all use, manipulate, study, and analyze public data at a variety of scales in their communities.

[The Chair, Liberal, Shawn Murphy]

Madame Lauriault, you're at seven and a half minutes, so could you perhaps conclude in a couple of minutes?

However, the main reason I'm interrupting is to ask if you could slow down a little bit for the sake of the interpreters.

[Tracey Lauriault]

Yes, I keep forgetting the poor ladies back there who are translating.

Excusez-moi, mesdames. Je parle un peu trop vite. Je vais faire de mon mieux.

These research organizations work in universities and communities, and they could all benefit from a more open government and an open data policy. Why? Because they could focus on their research and not focus on trying to find money to pay for public data. They could focus on actually using the data, as opposed to spending 70% of their time—and I really mean it, because I spend a lot of time doing this—trying to find those data in the myriad government institutions we have at all levels of government in Canada.

They could also benefit from that policy by not having to negotiate with public administrators on licensing. It's very difficult to negotiate access to data to do social research and policy research. No one is asking for private data; everyone's asking for aggregated data according to whatever geography they use when they do their area of analysis. If you ever try to negotiate access, in the current context of risk-averseness in the public service, to a public data set that may or may not make a particular minister look good, you will not get access to those data, because there is no overarching policy that guides how public officials should make decisions on the data sets they're using.

We have a number of issues, such as lack of public data standards in terms of formats. In particular, for the community groups I work with, data aggregation is important. The federal government does not have a mandate, it believes, to serve communities, yet that's where we live. We all live in a neighbourhood. We all live in a ward. We all live in a city or a county, and so on. We would ask that data be aggregated according to commonly recognized geographies--that is, according to these different communities as well as to Statistics Canada-recognized geographies and to the geographies of federal electoral districts or health districts.

There are regressive cost recovery policies. We often joke that you have to mortgage the house to study homelessness in Canada. That is deplorable. These are our public data. We have paid for these public data already through taxation. Please fund Statistics Canada in a way that it does not have to sell its public data. Don't make them give their data away free tomorrow, but then not properly fund them to do so. Increase their budget to cover the costs they would no longer recover, and let those public data be available to citizens so that they can do this great work these community groups are doing.

I already mentioned the issue of restrictive and non-interoperable data licensing. I already mentioned to you the lack of data access policies and the absence of data discovery mechanisms, which means that there is no portal and there is no catalogue. You're talking to all these federal departments and crown corporations and agencies. You have to make cold calls, and each time you make a cold call, you talk to at least 20 or 30 people before you find the data set, and then you have to negotiate. Please make those data easier to find; as well, organize them and wrap them in good descriptions with good metadata.

Also, mandate that anybody who receives public research funding in Canada must have a data management strategy. It's deplorable that when Canadian research is done, researchers aren't mandated or financially supported to share their data. This is very simple. CIHR has started to do this; as well, the International Polar Year is an excellent example of one of the first research funding projects that has done that in Canada.

In addition, we're not archiving and preserving our data. Please support the creation of a data archive for Canada. It would just make sense. These are our heritage resources; let's keep them and maintain them for the long term.

Of course, there is a lack of research funding generally on issues related to research around data.

Finally, I'll conclude with some basic recommendations. You can go through the submission in more detail later. I provide you with names of organizations you can consult with and things you can do, but immediately appoint an entity called the chief data officer in each agency and each department. That individual's role and responsibility would be to conduct an inventory in the agency of what those data resources are, who produces them, and how they produce them. The officer would wrap them in all those good open-access and metadata types of principles and data management principles. Then he or she would create a portal so that when researchers and these civil society organizations in Canada do their work, they can call one person, not 50 people, for one data set.

If you think of the Federation of Canadian Municipalities project, that's 200 variables. I spend all my time talking with wonderful public officials on the phone, but I would prefer to do the analysis and write the reports, because that's what helps us Canadians at the end of the day.

I would also suggest developing a catalogue. I would look at the GeoGratis model. I would look at the GeoBase model. I would look at how the geospatial data infrastructure was created, so I would go to Natural Resources Canada, which is an excellent example of how we can consider building an open data infrastructure for Canada. Then I would put all of the best minds of the country together and have them collaboratively work on addressing this issue. I don't think it is only the responsibility of government to do this. I think there are many organizations--in research, the provinces, the territories, all the federal departments, the community, and the private sector--that should help you with this project.

Finally, you should consider more creative and common types of licensing for all of the Government of Canada's data, whether it be administrative data, maps, the census, and so on. New Zealand has done it. England has done it, and so has Australia—all Westminster countries that have crown copyright. We should also be able to do it.

Thank you very much.

[The Chair, Liberal, Shawn Murphy]

Thank you, Madame Lauriault. You've certainly given us a lot to consider. Thank you very much.

We're now going to hear from Mr. David Mason, executive director of Visible Government.

Mr. Mason, the floor is yours.

[David Mason, Executive Director, Visible Government]

Thank you.

I hope you have received our crowdsourced briefing document, where we covered an array of topics around open data, open government, and more involved citizens. We covered topics such as the usefulness of open data inside government to enable connections and better enable relationships with vendors.

[The Chair, Liberal, Shawn Murphy]

You're making the same mistake that Madame Lauriault made. You're going a little too fast.

It is Valentine's Day. We have to be good and kind to the people behind us, so we'll just slow it down a bit.

[David Mason]

Okay.

In science communities, having open data helps create wider standards for more data sharing and enables a culture of scientist-citizen. In education, notable institutions enable free access to the world's best information.

We talked about poisonous data and systems that assume individuals would never get access to their own health care record, as well as inspiring science from GCPedia and our geomatics community.

Others have spoken about how open data can make access to information more efficient and useful. Business is exploring more open and social modes. Consumer-serving openness is a competitive advantage. We talked about creating a culture of innovation and problem-solving built on the fact that so many Canadians are online and how what we're building can create consistent, reusable knowledge systems for everyone, through which a 14-year-old or 80-year-old can access the same data and networks as a researcher, organize it according to their perspective, and connect with others. We talked about how people can stop using their computers as typewriters and instead create reusable data, about how many more people can be deeply involved in democratic processes, and about how this can be used to build up trust in government.

I want to talk about a specific open data project. Today, if I go to a health clinic, I may be told I can't be seen that day. If I search many completely different sources of health clinic information, I might get a better idea of the best clinic to visit at that moment. Modern Internet-based software can provide easy solutions to these kinds of problems. In an afternoon, I scraped the locations of hospital emergency departments across Montreal. I put them on a map and included the user's current position and the closest hospital, and added scraped information about capacity and resource usage. Even this effort would be useful for someone trying to form an informed opinion and take more responsibility for their own health. It could help many people waste less of their own time sitting in a waiting room and help balance the health system by choosing the clinics closest to them and those likely to be least crowded.

However, if hospitals and clinics intentionally published information as quality open data, much more could be built. We could learn where clinics are best for different conditions and develop real-time and predictive views of when to go to particular locations. Past the technical design, people could contribute their experiences to help measure problems and successes. This would result in a low-cost harmonious feedback loop for individuals and their health system. With open data, lightweight Internet tools, and crowd-sourcing, the budget impact would be minimal and the effects profound. Because hospitals are fragmented, we may never have an official comprehensive system, though with a minimal level of open data support we can have useful, constantly developing systems that institutions could never build in the foreseeable future.

Many people like me are able to create this kind of system in an afternoon, because it's what we do during the day. We work with free world-scale systems that let us put interactive data into the best and most recognized web interfaces in the world. The proprietary and custom interfaces often used by institutions usually can't compete with this. They make the user relearn a system that's usually not nearly as good as the best in the web, and cut and paste an address to get transit directions or see what's nearby. They don't let users easily add information that can be helpful to others.

In the last few days I have read two news items about governments not taking advantage of the best the Internet has to offer. In one case, the U.K. government paid a consulting firm £200,000 to create a system that collapsed under load when put online. An individual wrote a system in eight spare hours that was more robust. In another case, the BBC announced that it had to shut down 172 content websites for budget reasons. An individual scraped and archived them using a $4-a-month hosting plan.

Using the best low-cost tools online today for free, people use digital maps to find restaurants and bus routes that suit them perfectly. Craigslist demolished the newspaper classified business with a free, easy-to-use, volunteer-based service. People count on looking up information on the collaboratively created Wikipedia. Fine-grained news travels quickly in social networks, with personalized comments. Sites like openparliament.ca publish and allow finer examination of proceedings. These are examples of the benefits of digital networks. A basis of open data enables people to effectively reuse information, to participate in democratic processes, and to enable lifelong learning.

In a generation, the Internet will be deeply embedded in everything we do. We'll continue to see problem-solving waves of innovation from the best and most motivated minds around the world. Most people may not profoundly interact, but some will, and it will affect everyone.

All of this potential is based on the existing features and design of computer data in the breakthrough web created by Tim Berners-Lee, who leads open data development in the United Kingdom government. Berners-Lee's mandate is to make data open and accessible, including individual direct involvement.

Openly learning from, using, and advancing efforts in standards around the world must be a key part of the Canadian approach. We know there are qualities of open data ranging from the opaqueness of a PDF to richly organized and connected data using open standards and licences.

“Accessible” means that data need to be consistently organized according to many perspectives in a culture that embraces the idea that this is the right thing to do. Although most people are online and computers can be equalizers for vision- and mobility-disabled people, one-third of Canadians are not online and may never be, so we look to social networks to connect people.

Many two-way knowledge translators will be required inside and outside government. This is an enormous undertaking, but it's an investment that will yield smarter, more capable people, and genuine quality-of-life improvements in a knowledge economy. There will be short-term rewards, but we need to create long-term goals, visions, and concrete milestones with the open involvement of many people.

If we think about real steps forward, we see that as more information becomes available, it needs to be carefully organized using systems like CKAN; otherwise, it will never be found, or it will be redundant and opportunities will be lost. Data directories that don't use these structured standards are a step backwards.

Licences need to be determined. For many reasons, Creative Commons by attribution could be considered the best; it's well-recognized and creates links with the origins of data.

Government needs to negotiate openly with firms like Google to make sure that data available in cloud-based services don't become dependent on any provider, and instead become standards like those developed for transit services. My experience in hospital systems informs me that there are clear aggregated sets of data that can be shared, and others that can't. Government departments need to enable their existing experts and appoint people to determine how to draw clear lines in data reuse, as well institute an open data culture.

Getting people to widely understand how data are reused is a harder problem, but government could serve many purposes by working with media, producing an awareness and participation campaign, and supporting privacy and anti-fraud interests to instill an entertaining and realistic culture of inquiry in social networks. That attitude is the best starting point to create a trustworthy, participatory culture.

Finally, if government is going to conduct an e-consultation on this topic, that sounds like a great opportunity to work openly in a first real step to organize issues and truly involve individuals in these discussions as first-class participants.

Thank you.

[The Chair, Liberal, Shawn Murphy]

Thank you very much, Mr. Mason.

We're going to start the second round with Mr. Easter. You have five minutes.

[Wayne Easter, Liberal, Malpeque, PE]

Thank you, Mr. Chair.

Thank you to both the witnesses.

Tracey, thank you for your very extensive paper and recommendations.

One of the problems with data--and I don't disagree with you that we need it in an open fashion, and that it needs to be available--is which data believe. How do you get around what's accurate and what's not?

In the most recent example, Stats Canada crime information was disputed--I think wrongly disputed--by a research think tank the other day. It seems the numbers in that crime data were cherry-picked out of another report. In any event, now you have both sets of statistics out there.

How do you see open government and information being made in a way that you can have confidence in the data itself?

[Tracey Lauriault]

First I would hope that any data our government officials produce to inform their business practices would be reliable, accurate, and authentic. I presume, from the start, that the data are of a high enough quality that they already use these data to inform their practices.

When I'm talking about administrative data, how many people receive student loans? I would presume that number is fairly well discussed.

With regard to the crime data, it was one not-quite-centre type of organization that did that analysis, and it was only one institution of the many hundreds and hundreds of institutions to whom Statistics Canada already sells such data. If the data are good enough to sell, the data must also be good enough to share.

While I don't like Statistics Canada's regressive cost-recovery policies or its restrictive licensing practices, Statistics Canada is one of the best statistical agencies in the world. I know that their data are accurate, reliable, very rigorously and methodically collected, wrapped in fantastic privacy practices to ensure that no private information is revealed, and good. I'm not asking you to share data that government officials would not already be using in their work.

I know that my government officials are professionals, so I expect their data to already have gone through quality checks.

Thank you.

[Wayne Easter, Liberal, Malpeque, PE]

I don't disagree with you on Statistics Canada, but we've already seen the decision made on the census, which will in fact jeopardize the reliability of the census information.

You're suggesting in your paper—and I think you mentioned it again now—that there are regressive cost-recovery policies at StatsCan and other organizations. Are you suggesting that in the future there would be no cost recovery applied?

[Tracey Lauriault]

At the moment, a municipality has to buy the data, and I pay municipal taxes. Provincial and territorial governments also have to buy that data, and I pay provincial or territorial taxes. The federal government and all of their myriad departments and divisions also all have to purchase the same data.

If I go to purchase it, I have to purchase it, and I've already paid for it with my taxes. As well, research organizations such as university libraries on the Data Liberation Initiative also purchase those data. I, as a data user, and with these different organizations that I work with, have purchased the same data at least 10 times, and these are resources that don't diminish with use.

What I'm suggesting is that cost recovery might in fact be more expensive than the cost of sharing the data generally. If you go back and look at Natural Resources Canada and the decisions they've made with GeoGratis, with GeoBase, with the Atlas of Canada, with the geospatial data infrastructure, with topographic maps, etc., they discovered the cost of managing and selling royalties and managing ATIP requests. I think each $5 ATIP request costs $75. Think each time you purchase data about the cost of managing all of that.

What I'm suggesting is that it's incredibly cost-prohibitive to sell us the data that we have already paid for through taxation, and I'm asking that we share back with us our own public data.