Scholar Citation Forecaster 作成者: jonaso

Scholar Citation Forecaster adds a single small enhancement to your Google Scholar profile page: a year-end projection on top of the citation bar chart. Nothing else changes on the page.

What you'll see

When you visit a Scholar profile (e.g. scholar.google.com/citations?user=…), the citation histogram at the bottom of the page gets one extra element: a light-shaded extension on top of the current-year bar, showing the predicted total citations by year-end, with a horizontal tick at the point estimate. Hover the projection bar (on desktop) and a small tooltip shows the predicted total, the 95% prediction interval, the model used, and how many historical years it was based on.

If the projection extends past the chart's existing y-axis maximum, the whole chart rescales proportionally so nothing gets clipped.

How the projection is computed

The extension reads the existing per-year citation counts straight from the page DOM — it doesn't query any API. It then fits four candidate growth curves (constant, linear, log-linear, quadratic) to your past full years and selects the best by AICc (a model-selection criterion that penalises overfitting on small samples). The selected model's prediction for the current year is then blended with year-to-date pace using a time-weighted average — early in the year the trend dominates; late in the year the actual count dominates. A 95% prediction interval is computed alongside.

When it runs

Only on https://scholar.google.com/citations URLs that have a ?user= query parameter (i.e. profile pages). It does nothing on Scholar search pages, the welcome page, or any other site.

Privacy

No data leaves your computer. The extension reads chart values from the page you're already viewing, runs all the math locally in JavaScript, and renders an overlay. There is no network call, no analytics, no telemetry, no storage.

Limitations

• Designed for desktop layout; may render imperfectly on the mobile Scholar layout.
• Projections are statistical estimates, not predictions of fact. The prediction interval is calibrated against synthetic linear-growth processes; real-world citation patterns can deviate.
• Requires at least one full historical year on the chart; profiles with only the current year showing will see no projection.

Source & attribution

Released under the MIT License. Source available on request.

This extension is not affiliated with or endorsed by Google. "Google Scholar" is a trademark of Google.