These are three questions the dataset can answer, based on its structure and the profile query I ran. I haven't run the analyses themselves, so none of them comes with results yet. **What the data covers.** The table has 422,729 daily price rows for 152 companies and tickers, from 2012-01-02 to 2025-08-04. Each row has open, high, low, close and volume. About 99,984 rows fall in the period since ChatGPT's launch, which gives a large "after" sample to set against the long "before" history. Median close is about $35.57 and average daily volume is about 8.5M shares. ### 1. Did the AI surge change how semiconductor stocks behave, and for whom? - **Question:** Did returns, price levels and volatility shift after late 2022, and was the shift concentrated in a few AI-exposed names or spread across the sector? - **Method:** Compare each company's cumulative return and its average (high − low) / close range before and after the ChatGPT launch. Then rank companies by the change. - **Why it matters:** It separates real AI winners from names that simply rode a sector-wide move. That bears on overweight or underweight decisions and on which companies belong in a sector basket. ### 2. Has risk changed, and does it differ by company? - **Question:** How does volatility compare across companies and across time, and has the surge made the sector riskier or more correlated? - **Method:** Compute rolling daily-range volatility and return correlations by company and year. Use the year and month gold tables for the time trend. - **Why it matters:** It shows whether a semiconductor basket still diversifies, or whether the names now move together. That matters for sizing risk. ### 3. When does trading volume signal unusual interest? - **Question:** Do volume spikes, relative to a company's own baseline, precede or coincide with large price moves, and has the AI era raised the baseline? - **Method:** Compute each company's volume z-score against its trailing average. Test whether spikes line up with next-day or next-week returns, and compare baselines before and after the surge. - **Data caveat:** 4,888 rows have zero volume. These are likely halted days, missing data or illiquid names, and they should be excluded or flagged first so they don't distort baselines. If you want, I can start with #1. The per-company, per-year gold tables can answer it quickly.