CiteVector forecasts how each of a researcher's papers will keep gathering citations, then recomputes the h-index, i10-index and total citations at each year-end for the next five years. It works on any public Google Scholar profile, with no sign-in.
What the forecast shows
h-index in 3 years. A headline figure next to today's h-index.
Projected h-index and i10-index. A chart from the current year to five years ahead, drawn dashed so it is not mistaken for observed data.
Cumulative citations with projection. Observed year-end totals, then the projected totals.
Citation growth of the most-cited papers. Curves for up to eight papers through today's count and over the next five years.
Per-paper forecasts. The publication table lists projected citations at +1, +3 and +5 years for every paper, and they can be downloaded as CSV.
How the prediction works
Each paper is projected with Equation 3 of Wang, Song and Barabási, "Quantifying long-term scientific impact" (Science, 2013). The equation describes a paper's cumulative citations through three quantities: fitness (how attractive the paper is), immediacy (how soon it peaks) and longevity (how slowly it fades).
A Google Scholar profile gives only each paper's current citation count and publication year, not the dates of individual citations. So immediacy and longevity are held at the typical values reported in that paper, and fitness is solved so the curve passes exactly through the paper's current count. The h-index for a future year is then the h-index of the projected per-paper counts at that year-end, anchored so the current year matches the h-index Scholar shows today. The full formula and parameter choices are on the methods page.
Limits of an h-index prediction
Only existing papers are projected. Papers not yet published are not modelled, so for an active researcher the forecast is likely to understate future growth.
Typical parameters, not fitted ones. Without per-paper citation histories, every paper gets the same immediacy and longevity. Papers that age unusually fast or slowly will be off.
Uncited papers stay at zero. A paper with no citations yet has zero fitness, so it is never projected to gain citations.
Partial publication lists. When the data comes from a fallback with a short list (a Wayback capture shows only 20 papers) and that list is no longer than the h-index or i10-index, the dashboard says so, hides the 3-year figure and projects only forward from today's real values.
Publication dates are mid-year, because Scholar exposes only the year; a paper published in the current year is assumed to be six months old.
Treat the result as an estimate, not a promise. Past values on the projection chart are model reconstructions; for the h-index Scholar actually displayed on earlier dates, use the h-index history.