How to: do data journalism on a budget

Far from always being a resource-heavy endeavour, data journalism can also be the driving force behind frequent, if not daily news stories and topical features, required to be turned around at some speed.

This serves to challenge the idea, therefore, that data journalism always has to place a significant burden on the time, staff and financial resources of a newsroom.

But people often "associate data journalism with the most high-profile and famous examples", journalism academic Paul Bradshaw explained. "So they think of huge investigations like WikiLeaks and the MPs expenses, and they associate data journalism with investigative journalism, or with huge scoops."

"But that's not necessarily the case," he added. "Data journalism can take as many shapes as journalism itself. It can be very quick and cheap and simple, it can be just as flawed and ill-informed as normal journalism. So I don't think data journalism has to be expensive in any way".

David Ottewell, who is head of data journalism at Trinity Mirror, and leads the team of three who make up the publisher's data journalism unit, added that there is also the misconception that data journalism is just "a broadsheet thing". But, as their projects have shown them, such as their Real Schools Guide and "football ratings systems", it is also possible to deliver data journalism in an engaging way to "a tabloid audience".

In this feature, which follows a Journalism.co.uk podcast on the subject last month, we hear practical tips on how to most efficiently carry out data projects, and avoid wasting time unnecessarily, with the support of free tools.

Prepare, prepare, prepare

Preparation is largely about being organised. And that also saves you from making mistakes as well, or at least reduces the chance of thatPaul Bradshaw, journalism academic

Of course much of the work to ensure data-projects are run as efficiently as possible can be done before the data arrives on the scene.

Bradshaw, who has blogged about greater efficiency in data journalism, suggests using bookmarks and spreadsheets to keep track of things when coming across useful data online.

"Spreadsheets are very easy to re-order, to filter, so if you're trying to time things quickly, having it in a structured format allows you to drill down much more quickly.

"Preparation is largely about being organised. And that also saves you from making mistakes as well, or at least reduces the chance of that."

And one way to ensure you are prepared, is to gain a thorough understanding of your patch and what data may be available, he said, and when new information may be due for release.

"A general exploration of your area and an understanding of what's recorded in that area, who records it and just talking to people who deal with that data can make a big difference in thinking about possibilities and not overlooking things," Bradshaw added.

Reuse where possible, and appropriate

In a presentation at the Digital Journalism World conference last month, Nicolas Kayser-Bril, co-founder and chief executive of Journalism++, advised delegates to look at opportunities to reuse material, such as "the code base and all the technology", instead of repeating processes each time.

Speaking to Journalism.co.uk, he gave the example of Helsingin Sanomat, and the way they have built reusable templates to help them turn around data visualisations at speed, as well as the way the Guardian is "using Google docs to aggregate all the data they came across in the past three years, so they have this huge resource of data which they know inside out".

This means "anybody can go in this archive and get information really fast," he added, "and I think they are processes we need to look at when we want to do data journalism without the millions that the New York Times is pouring into that".

Keep data structured and organised

When it comes to building newsroom databases for everyone to use, it is important to ensure the contents are accessible, Kayser-Bril added.

"When you work with a database, investigate it, you have to know who was behind it, for which reason this data was aggregated this way and not in another way, what are the biases in the data, then you have to clean it because maybe the formats that are used in this original database are not the right ones.

"You have this huge amount of work that goes in before you can even use the data and if you have to do this work every time you want to use the data set you're actually losing money."

Instead, journalists should "structure all these data sets in a way [so] anyone in the news organisation can reuse them". He pointed to the Panda Project and Local Focus (built by the Amsterdam-based arm of Journalism++) as "solutions that have been made for the newsroom" to help to organise data in one place.

Maintain clear editorial boundaries

If you have a small team, then it is worth thinking carefully about what your editorial focus is going to be, as well as what is going to constitute an effective end product for you, and your news outlet.

At Trinity Mirror Regionals, for example, head of data journalism David Ottewell said he is "quite brutal about what we can and can't do".

Our job is to really add value, and if we're not doing that there's no point to us being hereDavid Ottewell, Trinity Mirror Regionals

As of earlier this year, the news outlet has its own data journalism unit, but as this serves a network of numerous local newspapers and websites, the team still has to be strict about what they commit themselves to.

"I've always said I'd rather we do fewer things brilliantly, than a huge number of things reasonably well," he added. "Our job is to really add value, and if we're not doing that there's no point to us being here."

The unit has two main functions, the first is to share daily news leads with the company's regional outlets, and the second is to build larger-scale data "resources", such as its recent Real Schools Guide.

In terms of the news tips, in order to ensure the process is effective – and manageable – "I have to be very careful about where our role starts and finishes" he added. This also means a local newspaper's readers are still being served by their newspaper's own journalists, instead of just a centralised unit, he said.

"For example, in the generation of news, we pick out the lines, we provide visualisations, we analyse the data, we make life very easy for the individual centres, but we don't actually write the copy. That has to be done, as far as I can see certainly at the moment, locally. It shouldn't be done by us."

Issue thorough FOI requests

One place where time can be wasted is in the process of issuing a Freedom of Information request, which perhaps neglects to ask an important question and so has to be re-issued. In this case, really doing thorough work in advance can help reduce delays.

"One of the problems with making an FOI request is that they only give you what you ask for," Bradshaw observed, "and so it's very important to have as detailed an understanding of what's possible before you make that request, and having those discussions about 'what do you record, how do you record it'."

At The Detail, journalist Kathryn Torney also added that thinking ahead is key.

"I certainly find that it's worth putting the time into researching what data there is, how it's held, what the column headings are, problems you may anticipate with confidential information, for example, and to tackle that as part of your FOI request.

"Because the worst thing is putting a request in, that getting refused after 20 working days because of some quirk to do with the data, then you have to lodge it again and before you know it, it can be quite a few months before you get your information."

Harness the expert network

The Detail regularly looks out for opportunities to collaborate with other media outlets on data projects, Torney explained.

"If I have a data project that I am working on that I know I have regional breakdown [for], there are certain newspapers that I have maybe contacted in advance and said 'are you interested in this', and to talk about the timing of when we launch that story and what I can do for them. Or we simply find that local papers go to our website, pick up the information that's of interest to them and then they run a story further down the line."

And the value of collaboration was another of the points stressed by Kayser-Bril last month. He said there is ample opportunity for journalists to engage with others who have expertise in relevant areas, such as statisticians. And it does not have to have a cost implication either.

"Journalists actually have something to offer to these people who have these specific skills, it's not about stealing their work for free, it's really about providing them with a meaningful mission that they'll be happy to do beside their day job."

If you come to that project with an egotistical attitude that essentially you represent a particular brand and therefore people are going to fall before your feet, it isn't going to happenPaul Bradshaw, journalism academic

Bradshaw said one challenge to this can be those newsrooms which want "to own everything about a story". But he added, "to get things done well, particularly with data journalism, you're likely to have to reach out to other people with very different skill sets".

And when you do so, he added, be mindful to consider your approach.

"There will have to be something in it for them as well," he explained. "They might believe in the story, but there needs to be a dialogue, so I think if there isn't that dialogue, if you come to that project with an egotistical attitude that essentially you represent a particular brand and therefore people are going to fall before your feet, it isn't going to happen."

Make the most of free tools

There are numerous free tools out there in the growing world of digital data journalism, from scraping and cleaning to visualising.

Ottewell stressed that Google's collection of data-related tools are a good place to start. If you are not sure of what is on offer Google recently launched a Google Media Tools site which highlights all its tools which can be of use to journalists, with a section dedicated to visualisations. Here is more on using Google Fusion tables. Others he suggested to bear in mind include Tableau, Infogram and GetBulb, the latter of which came out of an Irish Times digital challenge.

Other visualisation tools include mapping platform BatchGeo and chart-builder Datawrapper – both of which were recommended by Bradshaw and Torney. Bradshaw also added OpenHeatMap to the mix, as well as pointing to Google Spreadsheets or Open Office for collecting data, and tools such as Outwit Hub or Import.io to scrape from the web.

Finally he also suggested taking a look at FreeDive, "which will turn a Google spreadsheet into a searchable database, so you can publish that database and allow people to search the data to find out how it affects them".

Ottewell added that when working with external tools, consider "repeatability".

"We try to make sure everything is either cost-free, or very very cheap, and we try and make sure everything is repeatable and everything is compatible," he said. "So using something like Google and the Google suite of software, it's really useful to us".

And it is worth mentioning here that the individuals using the tools are just as vital, if not more so. For small teams in particular, having sharp data journalists is key, Ottewell said. "If I had the wrong staff we would not have been able to do anything like what we have been able to do".

"So being ruthless about the resources that we do have, having the right staff and picking the right ideas are the three most important things in what we do."

Make use of templates where effective

As mentioned earlier, the data journalism team at Helsingin Sanomat, have created a series of templates to power their visualisation creation.

"The idea from the beginning was that we make news apps very fast," Esa Makinen, news editor for data journalism, told us. "If there's some news element going on, we have to do something. We try to, whenever we do something, make it from scratch in two or three hours."

And Makinen said they will always look for opportunities to "modify" previous creations, to save time, instead of always starting anew.

Ottewell added that while "templating is really important", and is put into use in some cases by the team at Trinity Mirror, he advised journalists to be careful they are not overused to a degree that it loses the potential "impact". In some cases similarity is not a problem, such as with a "heavily templated" football tool produced by the data journalism unit, which readers can use to select their own team. But it is worth considering where repetition could be less effective, he said.

"In that instance, it doesn't matter that it's essentially the same graphic every time because it's the information which is of interest to people," he added.

And when it comes to home-made data apps, Ottewell advises newsrooms ensure they are made in a way which will require limited onward work.

"We can't be spending all our time maintaining things we've already built," Ottewell explained. "Once it's built, I don't want us spending lots of time maintaining it. I need to make sure that as far as possible things are automated."

"If they're not automated," he added, then they will at least be "very low maintenance."

Remain focused on the story at all times

When it comes to visualiations, think KISS (Keep It Simple, Stupid). While there are fantastic examples of in-depth visualisations and interactives being produced by newsrooms, there is also much to be said for those graphics which are more to the point.

Sometimes I personally can get very bogged down in how I'm going to visualise my stories, or how I can do them differently, and sometimes you have to realise that simple can be goodKathryn Torney, The Detail

And when resources – particularly time – are limited, focusing on the main story to be told with fairly straight-forward visualisations can still deliver a powerful result.

"Sometimes I personally can get very bogged down in how I'm going to visualise my stories, or how I can do them differently," Torney admitted, "and sometimes you have to realise that simple can be good".

The key is providing "layers", she added, by complementing data visualisations with other content, such as multimedia. But importantly, "keep your eye on what your key findings are and what your news line is", she said.

Ottewell also stressed that "the data has to lead the visualisation". Remember your role as a journalist, he said, and ensure that the end product is serving a purpose.

"You can tell a fantastic story sometimes just with a simple bar chart. If the story that's being told is good enough, the quality of the visualisation, or the intricacy of the visualisation, is not the most important thing. Indeed, sometimes, it can be a drawback because it takes away from the point.

"To me, yes, there's loads of great things you can do with visualisation. Yes there's some fantastic software to help us do that, but you shouldn't be bewitched by it and the most important thing, always, is the idea that drives the visualisation."

Free daily newsletter

If you like our news and feature articles, you can sign up to receive our free daily (Mon-Fri) email newsletter (mobile friendly).