How a publication record accumulates
Publication counts mostly measure how long ago someone graduated, so the question is not "how many articles" but "how many by year t". The line is the median graduate; the shaded band is the 25th to 75th percentile, and the dashed line the 90th. These are quantiles of the observed distribution, not confidence intervals — the band shows how spread out people are, not how uncertain the median is. At each year only graduates who are actually that far past the PhD are counted, so the sample shrinks to the right; the count is printed under the axis.
Cumulative journal articles, by years since PhD
Rate against seniority
Each dot is one graduate. If publication rate were purely a function of seniority the cloud would slope; the spread within each column is the part that is about the person rather than the calendar.
Articles per year vs years since PhD
Where they work now
Current sector
Career shape
Where each graduate started and whether they are still there. Unlike a by-cohort share, every person here is followed individually.
Where careers start, and where they end up
Every graduate with a known first job and a known job today, drawn from their first sector on the left to their current sector on the right. No estimation, no censoring, nobody excluded — each ribbon is a count of actual people.
What they work on
Research division
One bucket per graduate, from their published journals and supervisor.
Field areas
Up to three tags per person, so these sum to more than the headcount.