10 tips for collaborative data journalism projects by Mar Cabra

Mar Cabra, award-winning data journalist, former head of the Data & Research Unit at ICIJ, and Global Editors Network Board member, has kindly shared ten top tips for collaborative data journalism projects, to introduce the Data Journalism Den. These include the importance of building a network, advice on how to be a better manager, and why collaborative projects are like dinner parties.

What is a den? A cosy place for humans and for animals? That’s exactly what we want to create with the Data Journalism Den, launching today. The Den is a hub for the data journalism community to find information, resources, jobs, and most importantly, to start working together with people they didn’t know before.

The Den has a budding voice on social media, and a strong newsletter published every Tuesday and Friday. For all data journalists, we want to offer a unique database of projects and initiatives on the hub. We will only improve data journalism if we share resources and opportunities, which is the goal of this new hub.

The main innovation at the Den is our Matchmaking section where journalists can find partners, if the former are missing a skill or need an extra pair of hands. Obviously, you will not create a Panama Papers team on the Den – such a project would need to be kept secret – but collaborative data journalism cannot be reduced to major international enquiries.Vibrant data journalism will be made possible through open sourcing and collaboration and we hope that our matchmaking section will be used for local, regional, and national initiatives.

Two sections will follow in April and May: Firstly, a job section, because we believe data journalists deserve better recognition… and better salaries. Secondly, we will follow ProPublica’s datastore initiative. They were pioneers in selling their data and we want to develop the monetisation of data journalism worldwide.

Thanks for your feedback, let us know if you consider our initiative to be relevant. Please write to us at theden@globaleditorsnetwork.org

Many thanks to Mar Cabra, GEN Board Member, for introducing this new service with her ‘10 tips for collaborative data journalism projects’. We believe in collaboration, and we also believe that scaling the best projects is the best way to build a sustainable future for journalism. BP

Mar Cabra’s 10 tips for collaborative data journalism projects:

1. Build a network even before you start a collaboration

  • Network a lot even before you have a project. This way, you can call on people within the community once you want to build a collaboration.
  • Go to conferences. You can find other data journalists at The European Investigative Journalism & Dataharvest Conference and NICAR. You can also network at other events that are not just data journalism focused, such as the GEN Summit.
  • Follow other data journalists on Twitter to see what they’re up to.
  • Connect with developers by attending meetups in your city about open data or for Python or R developers. Going to these type of meetups has been helpful in my career. You can do them in your city, so you don’t even need to travel.

2. Think of your teammates as people first

The main reason why collaborations can go wrong is that the people in the team don’t get along. And that’s regardless of how good they are as professionals.

Marina Walker Guevara, the deputy director of the ICIJ, often says that organising a collaborative team that works together is like organising a dinner party. You know that this person and this person get along well, so you sit them next to each other. You also know that these two may not get along, so you don’t want to put them too close to each other. It could be that one is an extrovert and the other an introvert, so they just don’t understand each other.

It’s about understanding how different personalities fit together and that you’re going to be dealing with people first. I think we forget that a lot.

3. Check people’s track records

In data journalism we predominantly focus on skills: ‘I need a front-end person’, or ‘I need someone who knows R’. Technical skills are important, but when building a team, don’t forget to think about people skills, too.

It’s all about understanding the person you’re going to work with both as a professional and as a person, so try to get references to see whether the person is trustworthy and whether they meet deadlines.

I would actually compare it to dating. If you’re on Tinder, before you go on a date with somebody, you want to understand whether this person has the same values and the view of the world as you do.

On the Den, collaborations are called Matchmaking, which is a great name, because it really is a bit like dating. You can ask questions and research another person, you can look at what they post on social media and their previous projects. Never start a collaboration with somebody you don’t know anything about.

How the ICIJ Used Neo4j to Unravel the Panama Papers by Mar Cabra

 

4. Look for people who are collaborative

Try to see if your contacts are going to work well in a team. There are some people who are very good professionals, but only work well when they’re doing their own projects. If it’s a collaboration, you need people who collaborate – obviously! If there’s a very stubborn person in the team who is going to fight until they get their ideas through, you know that it may be a problem further down the line or something that you’re just going to have to manage.

First and foremost you should think about people management. I’ve underestimated this in my past sometimes, and I’ve seen others fail because they underestimate the issue. I now get the newsletter of the Harvard Business Review, and they have booklets about how to manage your team, how to get your team to give you feedback, and so on. It’s therefore good to look outside of journalism too to figure out how to manage teams.

5. Be clear about responsibilities

Decide from the beginning if somebody should take the position of coordinator. I think it always helps if somebody is a project coordinator or manager. Horizontal structures sometimes work, but with people that don’t know each other, it’s always better if there is somebody who has the last say.

In a simple project, like a collaboration between two people, you don’t need a project manager if the responsibilities are well-defined. But even then, think about who is responsible for making the final decisions. The bigger and the more complex the project gets, the more you need someone to make the final call. It really depends on the case, but I would say that the moment you have more than three people working together, it’s good to identify who is looking at the overall picture and who is responsible for making sure that deadlines are being met.

You also need to define the responsibilities in the beginning. It’s very important, because that way each person knows what they have to do, and you know who’s to blame when something doesn’t get done. Otherwise people might not realise they’re responsible for something, or they are uncertain and afraid of taking initiative. Or they take initiative too easily and its badly seen by the others. So be clear about who is responsible for what.

Never underestimate how much work and time has to go into project management. One of the first data journalism projects that I managed turned out to be a nightmare, because I didn’t realise I had two jobs: working on the reporting and managing other people.

6. Spend time planning the project

I think we’re not used to planning in journalism, but you need to identify what the needs of the project are and how long the different steps are going to take. I always have problems with this and I always underestimate, much like most other data journalists. Once you have a finished estimate of how long the different tasks are going to take, double the time!

Before carrying out a task, remember to sit down and think about the most efficient method. Say you want to get some data from a website. Instead of just diving into it, think first: Should I write code? Or should I just download it manually?

Coding is not always the right solution. It could be that coding a script would take a week, but if you just did it manually it could be done in a day. Yes, it’s painful, but it’s also faster. And sometimes it’s the opposite: you think you’ll just start doing it manually and then further down the line you realise that you should have written a script to save time. Or you may have the impulse to use R, but it would be simpler to do it in Excel. So strategise: stop, think, and then do, not the opposite way around.

From the beginning, think about how the story would be told best. Are you going to have an article or several articles? An interactive? Maybe a graphic or a searchable database? It’s important to consider this because it’s going to affect the work. The earlier you think about it, the earlier you’ll know how many people you’ll need to involve and the better the outcome will be.

 

7. Don’t underestimate the writing and editing process

It’s crucial not to overlook this. I think sometimes in data journalism we easily think something along the lines of ‘I have these numbers and this data, I’m just going to crunch them and that’s it’. Then after the data analysis, some people just sort of spit out the numbers. This does not work in article form.

8. Fact-check, fact-check, fact-check

It’s so easy to make a mistake with data. I’ve made some in the past, and whenever that’s happened, it’s because I didn’t leave enough time for fact-checking. Ideally, fact-checking should be done by somebody else in the team; somebody who didn’t work on the original analysis.

But it could be that you are the only data journalist in a collaboration and only you know the tool you used for the analysis. Then you just have to redo your work. Even the best data journalists make mistakes.

9. Before you decide on tools, do a threat modelling assessment

You need to consider the security needs of your project, so it’s important to do a threat modelling assessment. Do you need to protect your data? Do you need to protect the conversations within the collaboration? Different projects have different security needs: if you work on a Panama Papers dataset, it’s not the same as working on a fishing subsidies project.

The threat modelling analysis will define the types of tools that you’re going to use. You’ll need to ask yourself:

  • Is using email ok?
  • What about calls, especially group calls?
  • Can you use Slack for chatting or do you need to use Semaphor (an end-to-end encrypted messaging service)?
  • Where will you store the documents and datasets that you’re gathering? Can you use Dropbox, Google drive, or a more secure solution?

It’s also better to define the tools at the beginning, because it gets too chaotic afterwards. People will communicate through different means, which means that information can be lost easily.

Semaphor provides end-to-end encrypted group messaging for teams.

10. Identify how frequently you need to meet

I know people hate meetings, but I think it’s important to have them, even if you’re in the same newsroom. Identify how often you need people to reconnect. This depends on how long the project is going to be: If it’s going to take a week, maybe you’re ok without meetings, but if the project’s going to take a month, meeting once a week might be a good idea.

Often the big question is whether you need face-to-face meetings or not. I think they always help, so if it’s a long or complex project, I suggest you have face-to-face meetings if you can.

You should also have a place where people can communicate regularly, such as Slack or Semaphor. At ICIJ we had our own tool where we could update each other. These messaging platforms help to keep you up to date with others and they contribute to team-building and bonding.

The kinds of meetings you hold really depend on the project. For ProPublica’s Documenting Hate project, they have regular, short phone call meetings, where all the partners give an update. In some cases, such as software development, you might want to do stand-up meetings. But generally, treat a meeting as something that is going to be useful for the project. Set up an agenda for the meeting, put a time limit, and try to make sure they will be effective.

Harvard Business Review and other sources outside journalism have plenty of resources about team management.

 

Many thanks to Teemu Henriksson, and Freia Nahser  for their help editing the interview.