## Distribution of the numeric columns Nine columns are numeric, but six of them are stored as text, so some values failed to parse. The parsed counts and summary statistics are in the result blocks above. Counts are out of 300 rows. **Scores and ratings (clean numeric types, 7–16 nulls each)** - **Productivity_Score** (284 values): ranges 1.6–10, mean 5.23, median 5.0, SD 1.82. It is roughly symmetric, with the mean and median close together. - **Satisfaction_Score** (293 values): ranges 3–10, mean 6.15, median 6.0, SD 1.52. It is also roughly symmetric, and nobody scored below 3. - **Accuracy_Rating** (285 values): ranges 2–5, mean 3.43, median 4.0, SD 0.86. The median is above the mean, so the ratings are left-skewed. Most people rate accuracy at 4 or so, and a minority give low ratings. **Usage and time (text columns, partly unparseable)** - **AI_Usage_Hours_Per_Day** (271 parsed, 29 failed): ranges 0.5–23, mean 3.01, median 2.2, SD 2.45. The right skew comes from a few extreme values, and 23 hours a day is implausible. - **Time_Saved_Hours_Per_Week** (272 parsed, 28 failed): ranges 0–19.4, mean 5.81, median 5.05, SD 4.11. It is mildly right-skewed and widely spread, since the SD is about 70% of the mean. - **Work_or_Study_Hours_Per_Day** (279 parsed, 21 failed): ranges 2.1–20, mean 7.29, median 7.2, SD 2.14. It is centred tightly, with a few high outliers. **Demographics and money (text columns, partly unparseable)** - **Age** (263 parsed, 37 failed): ranges 16–97, mean 30.18, median 27, SD 12.47. It is right-skewed, and the maximum of 97 looks like an outlier or data error. - **Monthly_AI_Cost** (288 parsed, 12 failed): ranges 0–100, mean 9.84, median 0.0, SD 13.75. The median of zero shows that most respondents pay nothing, while a minority pay up to 100. - **Monthly_Income** (290 parsed, 10 failed): ranges 0–250,000, mean 2,304, median 521.5, SD 14,743. This is the most skewed column. The mean is more than four times the median, and the SD is about six times the mean, so a few very large values drive the average. The median is the better summary here. ## Caveats - **Parsing method changes two columns.** A stricter parse (step 3) gave fewer usable values for Monthly_Income (284) and Monthly_AI_Cost (273). That shifted their means slightly, to 2,329 for income and 9.76 for AI cost, with medians of 520 and 0. The other seven columns were identical under both methods. The figures above use the more permissive parse (steps 0–1). - **Treat the maxima with care.** The extremes in income, daily AI hours, age and daily work hours are probably outliers or entry errors. For those columns, the median and the spread are more reliable than the mean.