Shortly after Cesar Hidalgo created Data Viva, a Brazilian open data portal, he noticed something strange. Hidalgo, Director of the Collective Learning Group at the MIT Media Lab, built tools that allow users to make custom visualisations with economic data from all over the country. Immediately after launch, the hits started streaming in from social media – but users were not sticking around.
“Social media traffic was very short-lived – the bounce is really high, while with the organic search traffic, the people stayed around three times longer,” he said. “That was telling us that these were people who were actually looking for the exports of Brazil or the industries in Minas Gerais on Google – and when they found this page, they struck gold.”
Opening public data isn’t an end in itself. Portals created without a clear sense of purpose or any plan to attract visitors often lie underused: a 2016 report by Open Data Maturity in Europe determined that less than half of all national portal operators had a clear idea of who their visitors were.
To make portals work better, they need to fulfill the needs of their users, present up-to-date data attractively and, most importantly, make it easy for users to find the information they need.
We’ve picked out some of the best examples to help you get the most out of open data.
Build the portal around its users
Open data is useless if nobody uses it. Regular hackathons and events can overcome this by connecting data with the people who need it.
For the past five years, the City of Chicago has been doing just that. Chi Hacknight, a weekly event, brings together technologists and city officials to develop ideas and implement projects using the city’s data. Among their success stories is a tool which predicts where on Chicago’s beaches E. coli is at risk of breaking out. Chi Hack night has also given rise to non-profit organizations such as M-Relief, which alerts vulnerable people to food support programs based on their postcode.
Make sure data is well-maintained
For a portal to attract and retain visitors, it needs to offer accurate and up-to-date information.
New York’s DataBridge platform collects data from around 20 city agencies and other organizations through direct feeds. Geocoding is used to create a common standard for identifying features in the data, eliminating discrepancies in the way different agencies record data.
DataBridge is used by city agencies to transform the way they work. By drawing on data from the Department of Buildings, Health, Finance and Environmental Protection, among others, DataBridge helps feed the Fire Department’s risk modelling tool, Firecast, which predicts which of the city’s buildings are most at risk of fire. Agencies have a direct interest in ensuring that city data is accurate and regularly updated, rather than releasing ad-hoc datasets they don’t use themselves.
A picture is worth a thousand numbers
Raw data in spreadsheet form is difficult to navigate if you don’t know what you’re looking for. Visualisations, graphs and maps help to draw useful information to the surface.
In Cincinnati, the city’s Performance and Data Analytics team developed CincyInsights, a website containing nearly 25 different dashboards that show how public spending, business activity, ambulance call-outs and other services are affecting any part of the city. The dashboards contain interactive maps and graphs, which allow users to manipulate the data, track trends by street and see how things are changing across hours, days, weeks or months. During the winter of 2016, CincyInsights’ snowplough tracker received widespread coverage, allowing residents to see where ploughs were every five minutes.
Visualising data also made it far easier for public services to use. To combat the city’s opioid crisis, medical officials began using CincyInsights’ heroin tracker to see where overdoses were occurring and when they were most likely to take place. This information was then used to pre-emptively deploy roaming medical units. Thanks to CincyInsights, in under a year the volume of data downloaded from Cincinnati’s open data portal spiked.
Cut a clear path to the data
Not everyone knows to go to a data portal for particular information, so making datasets accessible via other routes is important to get them into the right hands.
One option for portal designers is to make their data accessible via search engines. This has worked particularly well in the case of Data USA, which has built “profile pages” for datasets around specific topics or locations. It was initially found that, while data visualisations might attract social media attention, they were not generating the type of traffic that would stick and become readers.
To overcome this, web pages were created within the platform to host datasets. These profile pages are designed to drag in traffic from search engines, with algorithms used to generate titles that best match search terms people use. As well as displaying data by cities, states and occupations, profile pages also package data as stories linked to commonly-searched topics. For instance, a page displaying the average number of hours worked by residents of different cities is headlined “The hardest working people in America”, increasing chances of it being found via search engines.
Portals should be part of a wider agenda
Bringing these elements together means conceiving portals as tools to improve government, engage citizens and spur development.
To see how this might work in practice, look no further than Kansas. The city’s KCStat program uses data from 311 calls to identify the areas of public concern, such as transport, economic development and safety, then set performance targets for city departments. Feedback is delivered through monthly public meetings and the city’s portal and data visualisations are used to show how policy is addressing citizens’ concerns.
For portals to deliver the maximum return on investment, they should be created with a plan to make content relevant, accessible and engaging to both government and citizens.
(Picture credit: Flickr/Open Data Institute Knowledge for Everyone)