A day in the life of a data journalist

The Hindu

22-05-2025

At 10 a.m., I begin scanning the papers for the day's news. Exactly 12 hours later, a colleague from the data team and I will need to deliver a 600-word piece with supporting graphs for the 'Data Point' section on the oped page. Like a cyclone far out at sea, the deadline feels distant and harmless at this time of the morning, but the rumblings of the storm can be heard.
Having long abandoned the practice of presenting our data stories as mere collections of facts, we now seek news pieces that allow for a narrative arc. The numbers must come together as a symphony. Every story needs a peg and a purpose. It must be built around a central argument, supported by evidence that reinforces the core takeaway.
And so we hunt for stories that can branch into compelling data narratives. We find that dog bites are on the rise in Kerala; at least three children have died in the past month. It is a strong story, no doubt, but we have covered it before. With no new angle to cover, we move on. The U.K. Prime Minister is rolling out new immigration rules; how will this affect Indians? It is a viable lead, but the data resides on U.K. websites and foreign datasets take time to decode. This cannot be done in a day. The Supreme Court has slammed the Enforcement Directorate. Promising, but the data is patchy at best. What about Virat Kohli's poor form? The Reserve Bank of India's latest report? The record in the sales of electric vehicles? The shrinking of Arctic ice? Or the highest single-day death toll in Gaza?
A dozen ideas are discussed, but there are many more reasons why we cannot implement them. There is no data. Or there is patchy data. There is good data trapped in bad PDFs. Or there is bad data on well-maintained portals. There is data but no narrative arc. Or there is a great arc but no data. And sometimes we ask ourselves: why do this story now? Post-lunch, often when we are on the verge of giving up, an idea clicks.
At around 3 p.m., the data team shuffles into office. One member is working on a prospective video, another on an interactive, and yet another on smaller graphics across the paper. Soon the major focus will be on the centre piece of the day — Data Point.
The data sources are ready for extraction. Using AI-generated Python scripts, we begin pulling in the numbers. But the data is messy — there are columns with no labels, empty rows, odd characters, misaligned tables, and missing meta data. All this needs to be cleaned and understood before any real analysis can begin. It is a task that will take several hours. There are seven hours to deadline and the winds are picking up speed.
Finally, with the data in shape, the visualisations begin. Using tools such as Flourish, Tableau, and Google Sheets, the graphs start to take form. We discuss whether a line chart may work or a scatter plot. Will they fit the space?
Then come the deliberations with the print design team. The data and design teams often agree to disagree, pushing back and forth until a consensus is reached.
Post-dinner, a clean white document awaits. It is story time. Over 600 words need to be written with the news peg, context, and data. One and a half hours later, the story is sent to the page editor.
By 9.30 p.m., we find ourselves standing in the eye of the deadline storm. The newsroom is buzzing. The page editor is firing questions at us, the design team wants annotations on the charts, and we need to sign off on news page graphics. All the i's are dotted and the t's crossed. With each passing minute, tensions rise. By 10:15 p.m., we finally send the page for printing after taking some last-minute decisions. The storm has finally passed.
At 10 a.m. the next morning, I begin looking at the papers again. Another storm is brewing. It is harmless — for now.
[e-mail]