My advice to new data journalists
Read widely, write a lot, learn to research, and start a blog
I receive a lot of emails from people (especially students) asking how to enter the world of data journalism. While I don’t consider myself particularly qualified to give this advice (I just started out in this world myself) people keep asking the question, and at some point, it makes sense for me to post the ideas online—both as a time-saving technique and so that people who are afraid to email/don’t have my address can receive the same information.
By now, I have responded to so many of these inquiries that I have a sort of automated mental template for my response when people ask for my advice. In keeping with a popular axiom about software development—if you’ve done something twice, it’s time to put it in a package—I think the time is right to expand on that template and put it online.
Before saying anything, I should mention that it helps to get lucky. And I have gotten very, very lucky.
My “career” (to the extent that working professionally for just four years on something can be counted as a career) started in 2016. While I was still in college, I wrote a forecasting model for the presidential election and put the findings up on my blog. I originally only set out to do it to learn to code, but it gained a lot of traction and a few thousand followers from journalists on Twitter. I also got an academic publication out of it.
I think the story would have stopped there, frankly—had The Economist not been looking to hire a young journalist at the same time. My now-boss found my work by way of Twitter, and only because his colleague happened to be researching French election polling around the time I had been forecasting the 2017 presidential election.
It is hard to know how much of what I’ve achieved is directly attributed to what I have really earned. Maybe luck only multiplied my merit somewhat, but perhaps it exponentiated it. Who can know?
I should also acknowledge my privilege; white men aren't exactly discriminated against in the media today. I have also been quite fortunate in the mentors I have had, both academic and professional.
Okay, here is my advice:
Write a lot
Learn how to research (this may involve learning to code)
Start a blog
Why read widely? Primarily, reading is key to exposing yourself to ideas that thinkers before you have developed. Reading makes us smarter, but it also helps us create new knowledge. Perhaps you disagree with something someone else has said or want to build off their research. You can’t do either if you haven’t already read their work.
Why write a lot? Just as you go to the gym to get stronger and develop your muscles, frequently writing help to develop your writing muscles. You can learn style and substance, grammar and structure, dos and don’ts in a classroom or textbook, but writing helps embed them in your head like habits. Perhaps it’s better to say the best practices get embedded in your fingertips. Writing also allows us to share our work with others and received feedback.
It's important to note that your writing may not start great, but that's no reason to give up. Writing does not involve just putting words to paper, but also revising and editing them. This is why we draft. Write something and then rewrite it. Then rewrite it again. Some authors even argue that good writing is not what makes you a good writer; good rewriting is. (Still, the better your original copy, the fewer iterations will be needed).
Why learn how to research? Because once you have read widely and learned to write, your skills are limited to conveying only other peoples’ ideas. You need to know how to research to come up with your own.
It is in this step that I also often advise young analysts to learn a statistical programming language, such as R or Stata or the like. But I also typically parenthesize this advice as it is only generally useful for people who want to extract knowledge from data, and I know that doesn't include everyone. I think you can get by as an empirical journalist without knowing how to program for statistics, but it also worth saying frankly that you will probably be better off knowing how to code than relying on Excel or Google Sheets to run models.
Finally, I tell people to start a blog. Why? Because in the master scheme of this advice — read, write, and research: AKA absorb ideas, convey ideas, generate ideas — the next step is to have somewhere to publish your ideas. They will need a place to live so you can show them off to others; so others can give you feedback; so you can improve them. If you have no job, the blog is also your most accessible public portfolio. You get to show off all that you know about a topic and highlight your new ideas.
Blogging also helps to create habits out of this schema of advice. In The Power of Habit, former New York Times investigative reporter Charles Duhigg tells readers about how cues and rewards can reinforce cycles of action. Publishing your thoughts online allows others to reward you for your hard work. That reward can take many forms: a pat on the back, a "like" on social media, a comment that sparks further discussion of a subject. Do not underestimate the power of receiving validation for your work.
I want to end this post roughly where I began. I am young; I do not consider this advice to be the be-all-end-all guide to a profitable and professional career in data journalism. With the extremely important caveat that I represent a sample size of one, and that this analysis clearly and forthrightly selected on the dependent variable, perhaps this can be useful for you.
I still invite email solicitations from people considering this field, especially students. But I have written more here than I will in an email, and perhaps this webpage can serve as a meta-function in response to the advice question. `Rscript -e “print(get(‘advice’))”`
I am at the other end of life. Retired for 15 years. I think you state it well about writing. I have always found the writing disciplines my mind so the ideas get refined as I edit and re-write.