Skills and Tools
Journalism is the area in which data journalists feel they most excel, with machine learning on the opposite end of the spectrum. While three in five consider themselves advanced at journalism, only one in five does that in data analysis. Most people in data journalism rate their skills as “Novice” or “Intermediate” in data analysis (78%) and data visualisation (77%). Data wrangling and scraping are instead areas in which most consider to have none to little skill in (65% and 71% respectively).
When breaking it down by occupation, most profiles follow the same trends in terms of rating skill level. Overall, students tend to be the most inexperienced in nearly all areas. The biggest gap between students and industry is in journalism, where only 30% of students have rated their skill level in journalism as advanced.
Experience in data journalism positively correlates with higher skill level rating, and this is particularly the case in data analysis and data visualisation. The areas with least improvement over time are machine learning and statistics.
As we see in the next section on work practices, programming is not a mainstream task for data journalists (chosen by 29% of respondents, identical to 2021). The most popular programming language in data journalism is Python (62%), followed by HTML / CSS (49%), and R (42%). In terms of frequency, the vast majority uses programming daily (58%). Nearly four in five said they are self-taught using online resources. Just over one in five of respondents are solely self-taught in terms of coding.
Experience in programming varies from having more than 16 years (19%) to having between 3-5 years (28%). Similarly to last year, within data journalism some people have been programming for a long time and some are relatively new to it.
Most people in data journalism rely on mostly or exclusively external software when it comes to graphic tools (45%). However, compared to 2021 there has been a slight shift towards in-house software, as 15% use mostly or exclusively the latter (as opposed to 11% in 2021). The data shows, like last year, no correlation between company size and type of software used.
Excel and Google Sheets are the most popular tools among data journalists (used by 72% and 58% of respondents respectively). The gap between Datawrapper and Flourish grew closer in 2022 (from five to two percentage-points). Nearly one in four said they use Python for work.
Excel and Google Sheets top in usage among educators, Flourish and Datawrapper among employees, and R and Python among students. Geographically, among the top ten countries in terms of number of respondents we found the United States, Spain, and the Netherlands ranking high in terms of usage of programming languages. While Flourish is more commonly used in the United Kingdom and Brazil, Datawrapper is more popular in Germany and the United States.
Training, upskilling, and demanded training
Educators reported that the area in which most demanded training this year was data visualisation (78%). This also coincides with the area in which most desire upskilling (57%). Data analysis follows in second place in both questions. When it comes to received training, just over half have received training in journalism, while 43% have received it in data visualisation and analysis, respectively.
The area with the largest gap between receiving training and desiring upskilling are machine learning and data wrangling. While around half of respondents desire upskilling in these areas, just 13% and 14% of respondents have received training in these areas.
As in 2021, a disconnect continues to exist between the areas in which data journalists desire to be trained in, and the ones in which they actually receive training in. The areas in which educators see a demand for training happen to be in between the other two.
- Rogers, Simon, J. Schwabish, and D. Bowers. “Data journalism in 2017.” Google News Lab (2017).