Nine columns are numeric in practice, but seven of them are stored as text. I parsed those to numbers, and the unparseable values were excluded. Figures come from the clean profile in step 4, which matched my earlier profiles. "n" is the count of values that parsed. **Spread and shape by column** | Column | n | Min | Max | Mean | Median | Std dev | |---|---|---|---|---|---|---| | Age | 263 | 16 | 97 | 30.18 | 27 | 12.47 | | Monthly_Income | 290 | 0 | 250,000 | 2,304.38 | 521.5 | 14,743.44 | | AI_Usage_Hours_Per_Day | 271 | 0.5 | 23 | 3.01 | 2.2 | 2.45 | | Monthly_AI_Cost | 288 | 0 | 100 | 9.84 | 0 | 13.75 | | Productivity_Score | 284 | 1.6 | 10 | 5.23 | 5 | 1.82 | | Accuracy_Rating | 285 | 2 | 5 | 3.43 | 4 | 0.86 | | Satisfaction_Score | 293 | 3 | 10 | 6.15 | 6 | 1.52 | | Time_Saved_Hours_Per_Week | 272 | 0 | 19.4 | 5.81 | 5.05 | 4.11 | | Work_or_Study_Hours_Per_Day | 279 | 2.1 | 20 | 7.29 | 7.2 | 2.14 | **What the distributions show** - **Heavily right-skewed:** Monthly_Income, Monthly_AI_Cost, Age and AI_Usage_Hours_Per_Day all have a mean above the median. - **Monthly_Income:** the mean is about 2,304 against a median of 521.5, with a standard deviation of about 14,743. A few extreme earners (maximum 250,000) dominate the mean, so the median is the better summary. Income is person-level data, so I only report aggregates. - **Monthly_AI_Cost:** the median is 0, so most users pay nothing. The mean of 9.84 comes from a minority of paid subscribers, up to 100. - **Age:** the median is 27 and the mean is 30.2. The maximum of 97 stretches the right tail. - **AI_Usage_Hours_Per_Day:** the median is 2.2 and the mean is 3.0. The maximum of 23 hours is implausible and probably a data-entry issue. - **Roughly symmetric:** Productivity_Score (mean 5.23, median 5), Satisfaction_Score (6.15 vs 6), Time_Saved_Hours_Per_Week (5.81 vs 5.05, mildly right-skewed) and Work_or_Study_Hours_Per_Day (7.29 vs 7.2) are fairly well-behaved. - **Narrow rating scales:** Accuracy_Rating only spans 2–5, with a median of 4 above its mean of 3.43, so it leans toward higher ratings. Satisfaction_Score starts at 3 and never goes below it. **Data quality caveats** - **Unparseable text:** the numeric-looking text columns have non-numeric entries: Age 37 of 300, AI_Usage_Hours_Per_Day 29, Time_Saved_Hours_Per_Week 28, Work_or_Study_Hours_Per_Day 21, Monthly_AI_Cost 12 and Monthly_Income 10. - **Nulls:** Productivity_Score is null for 16 rows, Accuracy_Rating for 15 and Satisfaction_Score for 7. - **Outliers:** the extremes in income, age, daily AI hours (23) and daily work/study hours (20) deserve a validity check before they feed averages or correlations.