Want Value from Open Data? Think Platform, Not Portal
Socrata's Safouen Rabah, VP of product, offers four pieces of advice for governments launching open data 2.0 initiatives.
Safouen Rabah is Socrata’s vice president of product and specializes in bringing the open data company’s new technologies to market. Since 2010 Rabah has worked with governments at all levels to spearhead open data projects and initiatives. Some of these include the company’s recent Open Data Network, signature data portals and a suite of apps that includes the recently launched financial transparency suite. Asked how governments can write the next chapter to open data, Rabah has this to say.
1. Apart from selecting the right “high-value” datasets, how can jurisdictions get the most return from their open data portals?
The first step is to go beyond the “portal.” In essence, a portal that aggregates datasets is valuable, but not transformative. The “return on data” comes from connecting the data to the citizen interactions, internal decision-making, and physical infrastructure of a city or a county. We have identified two key success factors here:
- Data on a portal is much less valuable than a data-as-a-utility platform. Data literally needs to become like water: available when and where people need it, at scale.
- Context is key. A dataset – and data in general – is something few people can relate to. They want information, not data, and they want it the same way they consume information on the Web.
Here is an example that illustrates both.
During the snow season, your car will be towed if it’s parked on the snow plow’s path. Where your car was towed to is data that the city tracks and publishes as open data. This way people can consult the data portal so they know where to find their car. A more powerful and useful application would be a simple SMS service where the resident texts their license plate number and gets an address [for the towing location]. This latter example uses the APIs that the data platform exposes to deliver a consumer-friendly service in the context that’s most useful to that user.
2. What are some best practices for keeping datasets current?
In an organization where IT is centralized, get the CIO’s or the chief data officer’s team to own automation of the data publishing — in real time if necessary — right from the source. This is becoming more commonplace as the applications of open data shift from a portal or catalog paradigm to a mission-critical platform for information and service delivery.
Where IT is decentralized [governed by different sets of leadership], we tend to engage the different departments to educate them on the value of open data to their specific mission. In the past, departmental participation was based on compliance with an open data policy. That’s great, but if that data is not helping them do their jobs better or serve their constituents directly, it’s just overhead from their perspective.
Another best practice that’s becoming the norm is to put pressure on the enterprise vendors to open up APIs to the systems they sell to government. No government organization should be locked out of its own data or require an expensive or restrictive license to getting its own data out of that system.
Finally, the tools that are now available for organizations to make publishing easier have evolved quite a bit. Many are developing open connectors to legacy enterprise systems and contributing them to the community.
3. How would you define a second-generation open data initiative in government?
Our definition of a second-generation initiative starts with the end goal: What outcomes are we driving? If data transparency is the outcome, then a portal will suffice. However, a more purposeful approach would deal with the following questions:
- How can data help us improve quality of life in our city or county or state? How do we make our [city] one of the best places to live?
- How can data help us improve the business environment in our jurisdiction? How can we make our [city] a great place to do business?
- How can data help us — the government as an organization — improve performance, improve the quality of our decision-making and ultimately improve our ability to deliver better outcomes for less?
Once you take that strategic outlook on data, the answers to what a next-gen initiative ought to look like become rather obvious. It’s really about the digital platform for a 21st-century government. The most important mind shift that government leaders can make is to stop thinking about open data as a project they do on the side and start thinking about how they can use data strategically to solve the problems they — and their constituents — care about.
4. What other tips can you add?
Here are some of the key tenets of that mind shift:
- A data portal is nice, but what we really need, if we’re going to put data to work at scale, is a platform that turns data into a true utility. If you’re a CIO, you’re thinking every data source should have a uniform API that [you] can manage and use to deliver not only inexpensive access to data, but true economies of reuse and a data foundation for service-oriented architecture. The cost differential between data as a platform and the prevailing model of bespoke custom development is staggering. Apps that cost millions can be built.
- Opening data to the public is necessary, but so is opening up data internally. As an organization, data is still locked up in department and agency silos. Collaboration is still cumbersome. We can’t even find our own data. Chances are it’s in spreadsheets and reports, scattered all over the organization in multiple duplicates. So a next-gen open data initiative also deals with internal sharing and discovery.
- Data is great, but what citizens, businesses, analysts and government decision-makers want is information they can use. The difference between the two is context. So how do we provide the context that empowers all the different stakeholders to answer questions, discover insights, visualize problems, see correlations, tell a story with data, etc.? What are the intelligent interfaces that will allow data and people’s lives to intersect in a meaningful and productive way?
- How do we make business processes smarter with data? How can we apply data to solving traffic congestion problems? Or parking? Or school performance? Those are the questions that ultimately matter.
- Last but not least, how can a government, any government at any level, benefit from the data economy and ecosystem of innovation happening around them? How can we connect our data to the consumer Web to increase our reach? How do we share predictive models to solve urban problems? How do we use comparative analytics to assess where we are versus our peers? Those are questions that a data-driven leader must ask.