How themes are defined and measured
AlgovestIQ tracks twelve curated investment themes — spanning AI software and infrastructure, semiconductor design, enterprise cloud, fintech, pharma, defense, utilities, and consumer — each with a maintained constituent list drawn from the ~1,400-symbol tracked universe. Classification is by material business exposure, not keyword matching, and constituent lists change through review rather than drifting with headlines. For every theme, every market day, the pipeline computes basket returns, signal breadth (the share of constituents with strengthening AIQ signals), strong-conviction counts, and a momentum trend for the theme as a whole.
What "activity" means on this page
The activity ordering combines the size of the basket move (in either direction), signal breadth, and the magnitude of momentum-score change. A theme can be highly active while falling — activity measures how much is happening, and the state label (Accelerating, Narrowing, Weakening, Reversing) says which direction it is happening in. The two together replace the "top themes" framing with something closer to a market map.
FAQ
Why twelve themes rather than fifty?
Because theme-level statistics are only meaningful over defined, adequately sized constituent lists. Each tracked theme carries enough members for breadth and conviction counts to be statistically interpretable — a standard most long-tail "theme" lists cannot meet. Additional themes are added when they can meet the same bar.
How fresh is this page?
The thematic index recomputes after every market close. The snapshot date is shown in the Pulse block, and weekly comparisons use the same pipeline's week-lookback mode rather than an approximation.