Looking at the dataset, I need to clarify what constitutes "numeric" data here, as the structure is unusual. ## Distribution Summary The dataset has **one true numeric column** (`question_number`) and **43 VARCHAR columns containing percentage values**. Here are the distributions: ### True Numeric Column **question_number** (survey question identifier): - **Min:** 1 - **Max:** 15 - **Mean:** 7.64 - **Median:** 7.0 - **Spread (std dev):** 4.71 This represents 15-16 distinct survey questions distributed across 105 response rows. ### Percentage-Valued Columns (Sample) The remaining columns store polling percentages as text (e.g., "14.21%"). After converting them to numeric values, here's a representative sample: **Overall** (aggregate response rates): - **Min:** 1.83% - **Max:** 78.70% - **Mean:** 18.35% - **Median:** 14.21% - **Spread:** 15.71% **Gender: Male**: - **Min:** 1.09% - **Max:** 78.19% - **Mean:** 18.53% - **Median:** 14.00% - **Spread:** 15.93% **Age: 18–29**: - **Min:** 0.61% - **Max:** 71.01% - **Mean:** 18.05% - **Median:** 13.10% - **Spread:** 14.93% **Income: Under $25K**: - **Min:** 0.00% - **Max:** 68.83% - **Mean:** 17.71% - **Median:** 12.34% - **Spread:** 15.55% **Education: Bachelor+**: - **Min:** 1.02% - **Max:** 79.68% - **Mean:** 19.50% - **Median:** 14.56% - **Spread:** 16.75% **Region: South**: - **Min:** 1.95% - **Max:** 76.35% - **Mean:** 18.41% - **Median:** 12.94% - **Spread:** 15.57% ### Key Observations All demographic breakdown columns show similar distributional patterns: - Medians cluster around 12-14%, indicating most response options receive modest support - Means are slightly higher (17-19%), pulled up by a few high-percentage responses - Wide spreads (15-17%) reflect the diversity of responses across different survey questions - Some segments show 0% minimums, indicating certain demographic groups had no respondents for specific answer choices