Tuesday 4 February 2014, 10:52
This is one of the reasons why data journalism is still a nascent concept in the emerging hyperlocal media sector. However, examples of activity do exist - particularly in the US where a number of websites have harnessed data journalism to help tell the stories of the communities they serve.
Perhaps the easiest way to do this is to use off-the-shelf tools like SeeClickFix, a US equivalent of FixMyStreet which allows citizens to report issues to local government via third-party sites. By embedding these widgets media organisations are able to crowdsource content which can help to highlight issues, generate stories and promote discussion.
As the US journalist Amy Gahran has noted, SeeClickFix is “an example of community news that doesn't necessarily come packaged in story form”, but with a little bit of creative use it has the potential to be a tool that offers more than “cool eye candy”. For publishers it’s potentially a very effective means to promote civic engagement, a characteristic at the heart of many local publishers’ ambitions. At the same time it can be great for story gathering, too, by providing journalists with a real-time means to tap into issues that matter to their audiences.
Using maps to illustrate stories
Maps are a further source of “cool eye candy” which can also lend themselves to data journalism: taking publicly available data but presenting it in new and interesting ways, either as a standalone item or to help illustrate wider stories.
In California, The Bay Citizen created an interactive Bike Accident Tracker (above) that made use of data from 14,113 incidents reported to the police between January 2005 and December 2009. The map (which takes some time to load) enables users to filter by road conditions, lighting and other factors such as ‘who is at fault’ to determine the safest routes to use and which to avoid.
Meanwhile San Jose-based NeighborWebSJ used maps to report on the ‘Streetlight Shutoff Program’: a cost-saving initiative to permanently switch off certain streetlights. Although it saved the city $77,000 a year, many residents and businesses argued the scheme increased the risk of crime. “There are some major streets that took some major hits,” resident Davide Vieira was quoted as saying after he reported numerous lights out on Alum Rock Avenue, including at 13 bus stops.
In response to these community concerns the website included a Google Map to indicate where lights were out across the city, as well as information on how to report issues to the authorities.
This reporting helped to identify that some lights had been turned off by accident while others were turned back on as a result of public pressure, including those at the bus stops on Alum Rock Avenue. In February 2013 the city agreed to reconnect 900 of the streetlights that had been previously shut off as part of the budget cuts in 2008-09.
The team at NeighborWebSJ has also produced a Google Map (top image) which shows all of the 2012 homicides in San Jose, with data coming from police press releases, and links to other media.
This type of activity has been taken a stage further by specialist sites such as Homicide Watch DC which covers every murder in the district of Columbia. The site uses “original reporting, court documents, social media and the help of victims”, as well as suspects’ friends, family, neighbours and others in an effort to “cover every homicide from crime to conviction”. It was awarded the Knight Public Service Award by the Online News Association in 2012 in recognition of its efforts to explore a single issue in a single geographic area.
Lessons from EveryBlock: Why data alone is not enough
All of these examples involve hyperlocal publishers using data to illustrate particular issues or community considerations - often harnessing maps as a means to tell stories or provide opportunities for audiences to analyse data in a manner that’s meaningful to them. The sites are not purely data driven, and they all provide context and interpretation which audiences clearly value.
These are important considerations for publishers, and the failure of the EveryBlock - a US-based website which aggregated local information produced by government and state agencies - helps to reinforce this point.
The site, which offered a data-centric approach to hyperlocal, launched in 2008 and was bought
Steve Johnson, assistant professor of electronic journalism at Montclair State University, was not overly surprised by this when he blogged about his experience of using EveryBlock’s data for lower Manhattan: “There were reports on what graffiti the city said it had erased each month, by neighbourhoods. But what was missing was context, and photos. If I’m a reporter doing a story on graffiti, I want to show before and after photos AND, more importantly, I want to know whether the city is successfully fighting the graffiti artists, ie who is winning. The raw data didn’t provide that.”
To some extent the EveryBlock team acknowledged this when in 2011 it moved the site in a new direction, telling its audience: “As valuable as automated updates of crime, media mentions and other EveryBlock news are, contributions from your fellow neighbours are significantly more meaningful and useful. While we're not removing our existing aggregation of public records and other neighbourhood information… we’ve come to realise that human participation is essential, not only as a layer on top but as the bedrock of the site.”
Ultimately, however, this pivot was unsuccessful and, although some protested that EveryBlock was “clearly the future”, the site failed to resonate enough with audiences or advertisers to survive.
What this shows is that while data is a useful way for journalists to access - and tell - stories, raw data alone will not suffice. Both context and analysis remain vitally important. Publishers overlook these elements at their peril.
Damian Radcliffe writes here in a personal capacity. This is an adapted extract from his contribution to Data Journalism: Mapping the Future?, published by Abramis Academic Publishing.