A map can be a powerful tool when it comes to finding and reporting the news and there is currently no shortage of mapping software currently available to journalists None however are more suited to uncovering geographic patterns in data than a geographical information system (GIS).
By interviewing those who know best in this this area, this blog draws on the combined knowledge of leading computer assisted reporting (CAR) experts to provide insight into what a GIS can offer the investigative reporter and the advantages it has over the more readily available and simpler web mapping services such as Google Maps and Tableau.
When a cholera epidemic swept London in 1854, ‘bad air’ was declared to be the cause by all leading medical practitioners – bar one. Dr John Snow was not convinced by the orthodox explanation.
He noticed that up to 500 people had died in just 10 days at one particular location – the intersection of Cambridge Street and Broad Street in the Soho area of London.
Dr Snow carried out an extensive survey of his neighbourhood, recording all cholera deaths (black lines) in the area and then plotted his results on a map. He noticed that there were a large number of deaths surrounding the water pump which confirmed his belief that the outbreak was due to the contaminated water supply at the intersection.
When the pump was investigated it was found that raw sewage had been running into it. The pump was removed and cholera deaths declined immediately, finally debunking the miasma (airborne) theory of disease and paved the way for the eventual acceptance of germ theory of disease dissemination.
This use of a map to investigate the cause of this epidemic against the established medical theory of the age changed how the medical profession would research and understand disease outbreaks and ensure Dr Snow’s place in history as one of medicine’s most celebrated pioneers.
Fast forward nearly a century and a half to 1992 and journalism would get its very own Dr Snow in the form of Steve Doig, then of the Miami Herald.
On August 14, a wind of change in how journalists would investigate stories was brewing off the west coast of Africa in the form of a tropical wave.
Two days later it had become a tropical depression heading towards Florida and Miami on the east coast of America, building up speed and intensity. Finally on August 16 Hurricane Andrew made a direct hit on Florida City.
Stormfacts.net estimated that the damage at 28,066 homes destroyed with a further 107,380 damaged and 82,000 businesses either damaged or destroyed. One quarter of a million people were left homeless, 700,000 people were evacuated, and 1.4 million homes left without electricity.
In today’s money losses amounted to $43.67 billion and the death toll put at between 32 and 44. Up until Hurricane Katrina in 2005 it was the most devastating hurricane ever to hit the US.
Miami resident and reporter Steve Doig, spurred on by losing the roof of his new home, decided to look closer at the damage, asking himself if the scale was totally down to an unpredictable and unstoppable act of God or was Hurricane Andrew given help on its path of destruction by reasons that had originated much closer to home?
In his quest for answers Doig decided to utilise a new investigative method he had a couple of years earlier adopted as part of his toolkit. He had familiarised himself with a geographic information system (GIS), the use of which would prove crucial in his investigation into the damage caused by Hurricane Andrew.
To explain exactly what Doig uncovered using GIS one must understand what a GIS is. The definition used by leading GIS company, the Environmental and Social Research Institute (ESRI), states:
“A GIS integrates hardware, software, and data for capturing, managing, analysing, and displaying all forms of geographically referenced information…GIS allows us to view, understand, question, interpret, and visualize data in many ways that reveal relationships, patterns, and trends in the form of maps, globes, reports, and charts.”
Clem Henriksen, senior marketing analyst of ESRI, with more than 30 years’ experience in the industry believes that:
“A big part of journalism is taking the data out of the tabular spreadsheet form and displaying it in ways that are more accessible to your audience so there’s the visualization element that a map has traditionally provided.”
“Now in the 21st century we make our maps on computers and GIS is a great tool for that. In addition GIS has the analytical capabilities of providing insight into the data so that a story can be made clearer”
A GIS allows a journalist to place various layers of information onto a map in order to conduct spatial analysis and thereby uncover patterns that may either prove a theory or to uncover something new which warrants further investigation.
The oft spoken principle in spatial analysis is that that 80% of all data can be spatially referenced and a GIS can help make sense of this data especially in an age where journalists have access to more digital data than ever before.
A GIS uses, manages and creates geo-referenced data in vector (points, lines and polygons – also known as shapefiles) and raster format (image tiles such as mapping and imagery).
The data will have attributes assigned which the GIS can use for spatial analysis and to uncover geographical patterns.
By knowing where patterns occur by using a GIS and adding more layers of data, the journalist can start to answer the other journalism questions; Who? What? When? How?” and including, perhaps most importantly, Why?
Which is exactly what Doig did…
Henriksen remembers Doig’s initial success with GIS:
“Doig as a journalist dug into it [the data] and he found that there was negligence on the part of the government to allow developers to create shoddy construction that was very vulnerable to the conditions of the hurricane that was not obvious until it was mapped and it was not obvious until the two data layers of wind speed and damage assessment were overlaid on top of one another,” he said.
Steve Doig says: “With Hurricane Andrew we could look at the map, we could see the patterns … to show that basically it was the changes in the building codes that magnified the extent of the disaster of Hurricane Andrew.”
Much like Dr Snow before him, Doig was not prepared to accept what he had been told, deciding instead to look deeper using mapping techniques to spot trends that warranted further reporting.
The outcome of this reporting led to several high-profile political resignations and Doig being awarded The Pulitzer Prize for Public Service (1993) for: ‘What Went Wrong’, his reporting on the disaster.
More significantly for journalism Doig inspired a generation of reporters to seriously consider adding GIS to their investigative toolkit in an age that was just starting to embrace Computer Assisted Reporting (CAR) and that which predated the arrival of web mapping services such as Google Maps, Tableau and Bing Maps.
While GIS had been used in other industries such as policing, marketing, transportation and epidemiology for a lot longer, Doig, now a professor and the Knights Chair in Journalism specialising in CAR methods at the University of Arizona, recounts how he went about exploiting GIS.
He said: “I got interested in finding a way to do data mapping when the US 1990 census was about to come out… It posed an interesting problem: how do you tell a story of demographic change?”
GIS offered Doig the solution he was after.
Doig says that he was not the only pioneer who was using GIS to uncover trends at the time.
“Back in the ‘90s …. there was me and a handful of other people who were starting to do this kind of thing beyond a level of simply colouring maps using graphic artists.”
One of these people was Jennifer LaFleur, currently director of CAR at ProPublica, ‘an independent, non-profit newsroom that produces investigative journalism in the public interest, and previously the Investigative Reporters and Editor’s (IRE) first training director.
She says: “I started using GIS when in the 1990s the census came out and at that time I used a programme called Atlas GIS which doesn’t exist anymore and started using it doing census stories and then became more interested in GIS.”
LaFleur was then able to use this new skill in 1998 while working for the San Jose Mercury News when she uncovered homes that were built on landslide areas.
“With the big problems with landslides in the area it was really important to look at it before these storms started coming…I think the real power of GIS to investigative reporting is being able to place different layers of information on top of each other and do stories that way.” she said .
Her investigative work with GIS on this story won her national acclaim.
David Donald is data editor at the Centre of Public Integrity and world’s foremost authority in data journalism methods. He believes:
“The investigate journalist to do a real thorough investigation that will hold up to potential outside attacks that may come need to look for evidence, evidence of whatever it is that you are investigating and GIS can provide evidence and statistical mapping that will help bullet proof your you investigation.”
David Herzog is the foremost authority on the use of GIS in reporting the news, a professor at the Missouri School of Journalism and an academic advisor to the National Institute for Computer-Assisted Reporting (NICAR), a joint programme of the Missouri School of Journalism and Investigative Reporters and Editors. In 2003 he wrote a pioneering book ‘Mapping the News – Case Studies in GIS and Journalism’.
Then he firmly believed: “Over the last decade some enterprising journalists in newsrooms across the US have been using a new reporting tool – GIS – to dig deeper as they try to uncover exactly where the news is happening.”
Herzog also believed thie increased of GIS by journalists would happen because of, firstly the continuing evolution of CAR [facilitating] the analysis of databases compiled by government agencies, secondly the evolution of GIS itself from mainframes and powerful network servers to desktop computers, thirdly the deployment of GIS in governmental agencies, from City Hall to the state house to Federal government and fourthly to the wide availability of GIS data on the internet. (Mapping the News 2003)
Asked whether she considered GIS to be more of a investigational tool or visualisation too, Megan Luther, current training director at IRE said:
“The GIS software advanced options are used more I believe by journalists as a reporting tool more than a visualisation tool for the audience aimed towards the audience. Journalists use the advanced GIS to see what the story is for reporting news rather than visualisation.”
Although journalists do use GIS more as a investigational tool Herzog stresses the importance and capability of using GIS to create good visualisations besides just being able to spot trends:
“I think that data visualisations are very important for data journalists who are using GIS, to be able to create maps that are visually appealing to tell stories that are engaging that people want to spend some time with.”
Donald believes there; “is a demand for data visualizations, the visualization of data is really up and coming and maps are seen as a crucial part of data visualization, not the only part, but crucial part.”
GIS is the tool to use if wanting to create an impressive static visualisation as it is the professional cartographer’s tool of choice.
Doig, looking back at the period since his Hurricane Andrew story, has seen a huge improvement in the use of GIS, particular when it came to reporting Hurricane Katrina in 2005 (Both in terms of investigation and visualisations).
“In that time span the power of GIS got so much better and the skills of journalists working for papers like the New York Times had got so much better that I think the coverage of the patterns (of the storms imapact) was extraordinarily enhanced by using GIS properly,” he said.
Other notable enthusiasts include John Bones with Verdas Gang; Gregor Aish, at Open Knowledge Foundation; Bryan Boyer on the Chicago Tribune and Nick Evershed of the Guardian.
Although Herzog was aware back in 2003 that GIS was expensive and complicated he fully believed that the benefits of GIS for investigative reporting would outweigh these disadvantages, rightfully anticipating that advances in GIS technology and its availabilty (Open source GIS) , and journalists increasing their computer skills would make GIS much more accessible to the everyday journalist.
Ten years on from writing his book Herzog believes that although GIS has come a long way in the newsroom he remains disappointed with its overall uptake.
He said: “I think the vision I had in 2003 hasn’t really come to pass for a number of reasons. Firstly the economics of using GIS … the complexity of learning the software… and lastly the emergence of web mapping and web publishing in that anybody can take data and put this up on the web.”
He could not have foreseen the rapid rise and capabilities in web mapping services (WMS) such as Google Maps and Tableau.
Now journalists can do a lot more with these free WMS than they could back in 2003, they can import their own data, embed WMS on news sites and present this data in a much more interactive, simpler and quicker way than a GIS can currently achieve.
More directly Donald states: “I really think the use of Google Mapping and Fusion Tables has been a big as impact as anything in journalism.”
This is a huge statement bearing in mind the person making it and the many changes we have seen in journalism in recent years.
WMS software such as Google Maps, Open Street Maps, Google Fusion Tables and Tableau has become the mapping tool of many UK leading data journalists such as Simon Rodgers of the Guardian, Claire Miller at Media Wales and Kathryn Tourney, The Detail in their investigative reporting.
Tourney uses Batch Geo and Google Maps extensively in her reporting and provides evidence that you don’t need complicated GIS to successfully spot trends.
Tourney said: “The mapping of this data unveiled a big gaping hole in the middle of Northern Ireland where no tickets had been issued at all. I wouldn’t have been aware of this without the map,” she added.
Donald says of WMS: “I think that it allows the reporter to move quickly and without the technical skills to try and see some things that may be useful in their reporting, whatever it is that they are doing so I think it lowers that cost and also lowers the amount of time invested and in improving skills, you can teach someone a fusion table in less than an hour which makes them functional.”
While Doig believes that WMS allows reporters to create a data visualisation that shows a pattern and provides a basis analysis tool to the reporters, there are concerns that reporters may be limiting themselves by relying solely on this software.
He said: “Ninety per cent of the kind of data mapping that journalists do is pretty straight forward. Producing dot maps or chloropleth [thematic] maps or those kinds are certainly within the reach of open source tools like Google Fusion Tables and if the free tool does what you need and you are unlikely to run into really high end problems during the course of your work then I don’t see and argument against using the free tool other than possibility of cutting off the concept of doing something more sophisticated if that opportunity comes along.”
It is this lack of sophistication that Herzog believes makes WMS a red herring, masquerading as a genuine and more convenient alternative to GIS:
“GIS programs are information systems for managing, storing, processing, analysing and visualising data. Google Maps and Tableau only allow you to visualize.”
He said: “I think things like Google Fusion Tables are great at visualisation but they are not data analysis tools. You can take your data and display it but you cannot take the data to answer underlying questions like ‘why is this happening?”
“I think they are for a different audience in terms of people and journalists who want to be able to visualise something pretty easily with pretty basic data handling skills,” he says.
(Point to note WordPress.com is not compatible with Fusion Tables and many other mapping software so can’t be embedded – only Google maps at the moment)
Donald also believes: “We have done some interesting investigation work using Google and Bing maps so I don’t want to downplay them and I don’t want to take sides there… I do think that they have a place … [but] … there are times when Google Fusion Tables are not sufficient so any journalist who thinks ‘well its good enough for me’ and it may be, then they are limiting themselves to the limitations of the online services and there will be limitations whether it’s the amount of data, the complexity of the analysis, how easy it is to geocode, whatever, there will be limitations . ..”
Donald continues: “Often in the more in-depth and deeper problems I run into some problems in that they don’t handle a lot of data, so that’s one the plusses in running something like Arc [GIS software] where I don’t have the same data limitations.”
these mapping services backed by internet technological giants such as Google and Yahoo, are advancing their analysis capabilities at a tremendous rate and in reply GIS companies like ESRI are increasing their presence online with many of the features of WMS but with reduced analysis capabilities of a typical GIS.
Slowly the distinction between the two is reducing as each seeks to adopt the advantages of the other. The issue is summed up by Shane Gorman on the Esri DC Development Centre blog titled: ‘Is Google Maps GIS?’ where he states:
“The truth is likely somewhere between the two, but I think the more important point is that the gap between the two is shrinking at an incredibly rapid pace. I’d predict in the next few years there will be little that differentiates the GeoWeb from GIS other than the names identified with their past like ESRI, Google, Microsoft, MapInfo etc.”
Until this happens GIS and WMS are currently two separate forms of mapping software with GIS giving the reporter a much more powerful tool for conducting spatial analysis for uncovering trends in data than currently provided by a WMS.
So should the investigative reporter ditch WMS in favour of GIS?
None of the these experts are suggesting that. While they believe GIS offers the investigative reporter a greater degree of analysis in order to uncover trends they are keen to stress that it is beneficial for journalists to adopt as many investigative methods as they can.
Donald considers: “GIS a part of my investigative toolkit, but not a primary one, not that it can’t be, but it’s not always the first tool I turn to.”
Investigative journalists need to arm themselves with as many data analysis skills as possible and be aware of which tool is most suited to a particular purpose whether it be GIS, WMS or knowledge of database and spreadsheet technologies or the more traditional forms of reporting.
Herzog, the biggest advocate of the use of GIS in news reporting, also believes GIS is just one option investigative journalists should utilise in uncovering trends and the reasons that lie behind them.
He said: “I think that GIS is just one of the tools we have to answer the ‘why’ question.”
Sometimes a GIS alone can answer the questions but not always.
When it doesnt it is still a very powerful tool in giving clues, insights and connections between different bits of data through performing spatial analysis on a multitude of different variables and comparing outputs.
Doig is less convinced on the ability of GIS to answer the why question:
“One of the points I always make when [his students] are making any kind of data story whether they are using GIS or using Excel etc… is that it will answer any number of the classic W’s and H but it won’t answer the sixth one which is why?
“By understanding where events take place you can start to answer the ‘why’ question.”
Why are there those differences in these patterns they see? What caused them is the real story is and that’s the thing I try and get my students to think about.”
Claire Miller, a leading UK data journalist states in her book: Getting Started With Data Journalism, 2013, that the role of the data journalist is;“not about carrying out research into the data that will lead to definitive answers about cause and effect.”
But rather to; “look for patterns, the outliers and the connections between data in order to ask questions about what a dataset might be telling.”
How you get answers to these questions will most probably come down using several forms of investigative methods such as GIS and spreedsheets hand in hand with traditional forms of reporting such as knocking on doors and making telephone calls.
Whether investigative journalists adopt GIS or WMS or any another investigative method is a choice they have to make dependent on the task in hand. There is no need to conduct complicated GIS analysis when a simple WMS product can achieve the same outcome in 20 minutes.
Doig warns: “You are limited in the tool you use if you are necessarily not aware of the other possibilities that may be out there. It’s like that old saying, ‘if all you have is a hammer then everything else looks like nail’.
Doig would still have no roof after Hurricane Andrew if all he had was a hammer to put it back on; much like other investigative journalists will be limited in their reporting if they only keep one type of mapping software in their journalistic toolbox.
By Richard Bedford.
I thank the following for their kind contributions:
Professor Steve Doig
Professor David Herzog
And last but not least, Kathryn Tourney