Fostering a Local Culture that Values Data

In my first essay I noted that data and data-driven decision-making are transforming government and politics and that we must involve the local community to ensure that it builds us up rather than damages us. I suggested three efforts in support of this.

This essay starts to tackle the first: to foster a local culture that values data in driving policy decisions and expects decision-makers to share that data and to engage with impacted communities around it.

I want to linger awhile on the word culture. It’s easy to rush past that one without a lot of thought: of course we should build a culture that supports the premise of the essay. Moving on …

But seriously, why do we need to build a local culture? Why not simply demand that local institutions become more data-driven and community-centered? They can draw from a large and growing body of knowledge about best practices in open data, outcomes-focused performance management and community engagement.

If you have local institutions that are capable of doing that, by all means go for it. But in most places, including where I live, our institutions really don’t have that capacity. It must be created.

Easy to say. Really, really hard to do. Not because some people will resist. They will, of course, but that’s not the main problem. The biggest obstacle is that becoming inclusively and equitably data-driven turns out to require more than a willingness to do something differently; it requires the willingness to do everything differently. We have to let go what we “know” in favor of what the data tells us, listen to people and truths that challenge us and make us deeply uncomfortable, act before we feel ready because the data is clear, or hold back when we’re raring to go because we don’t actually know enough.

In other words, becoming deeply and inclusively data-driven requires a major change in the culture of our local institutions. Culture is critical because it is the only way to coordinate the actions of a large number of people who operate mostly independently. The alternative, trying to police everyone’s behavior, is extremely expensive and almost never works.

Here locally, both the County and the City will soon have new managers. Isn’t it their job to lead that change?

Yes. But.

Real, deep, long-term change in a democratic society never comes by getting them to do it. That sort of change remains shallow and vulnerable until we accept that we are responsible, that we are a critical part of this, perhaps the critical part. Institutions that don’t have to be accountable to the community eventually won’t be. If we want to see sustained data-driven policy and decision-making in our community, we must accept our own responsibility for expecting it of our institutions. That shared sense of responsibility and shared expectation is the culture I’m talking about.

The next question, of course, is how? Culture isn’t something we can wish, decree or cajole into existence, especially not in a large, diverse community. How are we supposed to go about building it?

First, it’s worth noting that it’s already started, at least here in Asheville. I suspect we’re not alone. I have been surprised and impressed over the last couple years by how often activists and local community leaders talk about data, about the need to use data in their advocacy and solution-finding. These aren’t policy wonks and data geeks, they’re activists from Black Lives Matter and the NAACP, local homeless advocates, transportation activists, advocates for the elderly. The word is out. And, like it or not, local institutions are having to change to accommodate them.

That is precisely the kind of catalyst to institutional change that I’m talking about.

But can we take this even broader? Can we make data-driven culture popular? The question seems absurd. But I don’t think it is, and, in fact, I think we have an excellent tool to help answer it.

The Results-Based Accountability (RBA) framework was developed by Mark Friedman around the turn of the century. Use of the framework has grown dramatically since the publication of his book, Trying Hard Is Not Good Enough. It has become a nationally accepted standard for effective data-driven decision-making that keeps things centered on the communities that are impacted.

The framework maintains a clear distinction between community indicators, which measure “the progress a community is making towards achieving community well-being,” and performance measures, which assess how well an organization or program performs for the specific customers who benefit from its services. Institutions are held accountable for their performance, but we always also attend to the impact of their programs on the broader community.

There is much depth to explore in RBA, but I think its power lies especially in its simplicity. To get at the most important performance measures for a program or service, RBA asks three simple questions:

  • How much did we do?
  • How well did we do it?
  • Is anyone better off?

I don’t know about you, but that looks to me almost like a chant: “How much did you do? How well did you do it? Is anyone better off? Show us the numbers!” The beauty is that it is stated so anyone can understand while simultaneously aligning with a nationally accepted performance management framework.

What if local activists, local media, individuals and community groups consistently asked these questions of government staff, elected officials, university administrators, local anchor employers, any institution who claims a role in improving outcomes for the community through its programs? Could it eventually just be a given that these questions must be answered for any program or initiative? Could it just become an integral part of our local culture?

Perhaps it’s a silly thought, but I’m having trouble letting it go. What do you think?

Pitfalls & Potential: Building Community Capacity to Use Data

The idea of data-driven decision-making and its potential benefits is hardly new, nor is awareness of its dystopian potential. What is new today is the explosion of data being collected as a result of conversion to digital systems and the technological capacity available to process and learn from that data. New industries are being built on these capabilities and existing ones are being transformed, including politics and government.

These changes areexciting. We have the opportunity to use data to drive positive change, increase transparency, and achieve greater accountability in ways that have never been possible in the past. We should seize that opportunity.

But these changes are also dangerous. As we grapple with the ways social networking data are used to manipulate politics and civic discourse, we are also plagued by cases of data-driven decision-making gone wrong, from predictive policing software that simply automates the system’s existing biases to databases of gang members or debtors that devastate people’s lives on the basis of demonstrably inaccurate information. Too often we run well ahead of our capacity to safely manage and use the data we collect.

This is not just a national issue. It is also very much a local one. As cities pursue dreams of becoming “smart;” as police are tempted by the promise of automating hard, dangerous or costly parts of their jobs; as cities implement new practices in performance management and accountability, both the good and the bad of this new data-rich world are happening right here in our own communities and it is here that we must grapple with them.

And make no mistake: data is a critical part of the issue, but this is about more than data. In fact, I believe the challenge we face today is nothing less than how to rebuild our democracy for the 21st century. And I believe that any meaningful effort to face that challenge necessarily starts locally.

Indeed, in Democracy in America, Alexis de Tocqueville identifies “local government, that prolific seed of free institutions,” as a critical factor in the formation of the unique institutions of American democracy. It was then and continues to be today.

I should clarify just what is meant here by local government. Obviously the term includes city and county governments, but it is by no means limited to them. It includes the entire ecosystem of individual residents, community groups, governments, media, nonprofits, and businesses that collectively support and engage in local decision-making. A community where governance is limited just to the formal institutions of government is a poor community indeed.

I include the broader community for a very important reason. One of the important lessons learned over the last few decades of community organizing, civic engagement and civic tech is that achieving effective and equitable outcomes for all groups in our communities requires that we proactively and deeply involve all groups in the decision-making that yields those outcomes. That lesson is especially relevant to our ability to safely use data to achieve better and fairer outcomes for Asheville and for our neighbors throughout western North Carolina.

So what does that mean in concrete terms? Obviously that’s not a question I can answer fully in a brief essay, but I would suggest that our work centers on three primary efforts:

  1. Foster a local culture that values data in driving policy decisions and expects decision-makers to share that data and to engage with impacted communities around it.
  2. Build the capacity of community members to understand the value of data-driven decision-making and the dangers that we must guard against.
  3. Create tools and resources that empower community leaders to use data to inform policy and policy advocacy and establish baselines of agreed-upon authoritative data.

I will expand on this over the next few weeks. In the meantime, I would love to hear your thoughts. You can find me on Twitter as @ejaxon.

Treating Data Like a Strategic Asset

My recent post on shared data systems in the City of Asheville, NC introduced the idea that adding a dataset to our management and reporting repository is also a chance to be more proactive about how we manage that dataset. It’s an opportunity to decide exactly how to represent and document the data, who should have access to it, and how we keep it up-to-date. This is obviously a good idea — it only remains to figure out how.

Not surprisingly, that how turns out to be a challenge. Fortuitously, the day after I published the post, open data leader Andrew Nicklin of the Johns Hopkins Center for Government Excellence wrote that “the most important step you can take [to address data quality problems] is starting to treat your data like a strategic asset.” That idea — treating data as a strategic asset — turns out to provide a helpful foundation for figuring out how to realize our vision.

What makes data strategic?

Treating data like a strategic asset sounds great, but what does it actually mean? More fundamentally, just what makes a particular set of data “strategic”?

Internally it seems straightforward that the strategic value of a dataset should be tied to its ability to measure performance and to support decision-making in areas that the organization deems strategic. Externally, I believe the value of data is best measured by its ability to empower entrepreneurial activity for economic gain and improved civic engagement, something explicitly called out in many open data policies, including ours. Briefly, then, we may say data is strategic if:

  • It is used to improve decision-making that influences outcomes tied to strategic goals;
  • It is used by external actors to create economic or social value for the community.

What’s striking about those definitions is what they share: use. It is the use of the data that makes it strategic (or not). Thus, our approach centers on use and users.

Cultivating and empowering data users

If the strategic value of data lies in its use, then data with no users obviously can’t be strategic. Perhaps then we should avoid adding a dataset until a compelling use and set of users are identified?

That’s certainly an option, but we believe treating data as a strategic asset aligns better with the proactive approach proposed in the recent GovLab/Omidyar Network report on open data impact, which recommends that governments “take steps to increase the capacity of public and private actors to make meaningful use of open data”. Their recommendation pertains to external users of open data, but it applies equally well to internal users. The key is to create a relationship with data users and to actively support and expand their ability to make effective use of data.

Our work in this area is just beginning, but we are experimenting with several ideas:

  • Talk about it. Every chance we get we talk up our efforts to make data easier to access and use. What’s gratifying is that word has begun to spread: people are starting to come to us to talk about opportunities they see to use data to manage performance and and to communicate and collaborate with citizens.
  • Give users ways to tell us what they’re interested in. The new version of SimpliCity will let people subscribe to specific topics and datasets so that we can reach out to them when changes are in the works or to get user feedback on what we’re providing. We are also launching a new public records request portal that not only lets citizens access data from prior requests, but gives us a better window into the kinds of data that people are interested in.
  • Identify and connect with key data users in the community. We plan to hold outreach events for frequent open data users in the community, such as the local Code for America brigade, news organizations, advocacy groups and professional groups.
  • Provide ways for users to hold us accountable. There is no better quality-control mechanism than to have active users who depend on the quality of the data, but it is important to communicate clearly what they can expect and to give them an easy way to communicate issues to us. In addition to our efforts above, the infrastructure discussed in the next section will play a vital supporting role in accomplishing this.

Let metadata drive the data infrastructure

Creating relationships and maintaining conversation with users of the data is important, but what about the actual mechanics of maintaining high-quality metadata for our data? No matter how noble our plans and intentions, the minute we have to do something special to keep the metadata up to date is the minute it will begin falling behind.

Our big idea here is to turn the process around. Rather than try to keep metadata in sync with the data, why not let the metadata itself drive the entire data infrastructure? Maintaining metadata and maintaining data then become the same activity.

Behind-the-scenes look at the data infrastructure of a city.

That’s the idea that powers our new data management system, ComplexCity.* ComplexCity consists of a hierarchy of metadata directories together with a few scripts that use that metadata to create and maintain the data infrastructure. At the highest level, the system has three key design goals:

  1. Provide high-quality metadata to reporting and management data users,
  2. Maintain the integrity of the relationship between reporting and source data,
  3. Maintain the integrity of the relationship between the data and applications that use it.

The system is a work in progress, but currently there are scripts to:

  • Validate data set definitions against the associated tables in the target and (soon) source databases;
  • Create ETL jobs that move data from enterprise systems into the reporting warehouse;
  • Run the ETL jobs, accounting for dependencies between datasets; and
  • Generate API code and configuration for use in SimpliCity’s GraphQL server.

ComplexCity will also help us better hold ourselves accountable to users of the data through generated dataset dashboards. With the launch of the new version of SimpliCity, each dataset in the system will automatically get a dashboard that includes summary information about the data, quick links to APIs and downloads and, most importantly, all the metadata, including links to contact the data owners about any issues with the dataset. By exposing this metadata to our users, we hope to empower them both to make more effective use of the data and to help us ensure that it is high-quality and serves the needs of the community.

The road ahead

These plans are simple enough to state, but will entail an enormous amount of work in the months (and years) ahead. In carrying out that work, we will undoubtedly discover major gaps in our thinking as well as exciting opportunities to leverage what we’re building for even greater value. We’d love to hear your own ideas and critiques and would love even more to find ways to bring this approach to other local governments.


*The name was initially triggered by a joke — SimpliCitypowered by Complexity — but the more we thought about it, the more it grew on us. There is no getting away from the fact that the data infrastructure of a city is complex. ComplexCity is our approach to managing it.

Photo credit: The image above is Complexity by Mark Skipper.

Note: This post has been cross-published on DigitalSimplicity.io, the City of Asheville IT Services blog.