saturn

This dataset covers 3,222 U.S. counties with population, poverty, and SNAP participation estimates alongside FIPS and state identifiers. Population and SNAP-related counts are extremely right-skewed — total_pop has a skew of 13.4 and a max of 9.78M against a median of just 25,174, with similar long tails in poverty_pop and snap_participants_est. The poverty_rate column is more behaved (median 13.55%, max 66.32%) and is probably the most useful field for cross-county comparison without log-scaling. Note that snap_eligible_est appears to be an exact duplicate of poverty_pop (identical stats), which is worth verifying before using either as an independent variable. State coverage spans 52 distinct values, so DC and territories are included.

This column holds U.S. county names with state qualifiers — 2999 of 3222 values contain the token 'county,' followed by state names like Texas (256), Virginia (189), and Georgia (159). Every one of the 3222 rows is unique with zero nulls or duplicates, and lengths cluster tightly between 16 and 31 characters (mean 24.3, 3.2 words). The near_unique alert plus the 'County, State' pattern strongly suggests this is a geographic identifier rather than free text.

This is a population count by geographic unit, with 3,222 rows, no nulls, and 3,173 unique values, suggesting one row per area (likely U.S. counties given the count). The distribution is extremely heavy-tailed: median is 25,174 but the mean is 101,340 and the max reaches 9,782,602, producing a skew of 13.36 and kurtosis of 297.59. About 13.9% of rows (449) flag as outliers, reflecting a small number of very large jurisdictions dominating the tail.

This is a count of the population in poverty per record (likely county or similar geography), ranging from 3 to 1,343,978 with a median of 3,799.5. The distribution is extremely right-skewed (skew 14.73, kurtosis 342.21) with 362 outliers (11.2%) reflecting a few very populous areas dwarfing the rest. No nulls or zeros, and 2,839 unique values across 3,222 rows suggest near-record-level granularity.

Stored as numeric but the values look like FIPS-style state codes: 52 unique integers spanning 1 to 72 across 3222 rows with no nulls and no zeros. The roughly uniform spread (mean 31.27, median 30, std 16.29, skew 0.16, kurtosis -0.63) and the count of 52 distinct codes are consistent with US states/territories rather than a continuous measurement.

Despite being labeled 'county', this column holds small integers ranging from 1 to 840 with 330 unique values across 3,222 rows, suggesting it's a numeric county code (likely a FIPS-style identifier) rather than a true measurement. The distribution is heavily right-skewed (skew 2.87, kurtosis 11.6) with 178 outliers (5.5%), which is expected for ID codes but meaningless as a numeric signal. The median (79) sits well below the mean (103), and there are no nulls or zeros.

This is the FIPS county/state code identifier — values span 1001 to 72153 and every one of the 3222 rows is unique with no nulls. The distribution is near-uniform across the code range (skew 0.16, kurtosis -0.63), consistent with a geographic key rather than a measured quantity.

This is a numeric poverty rate, almost certainly a percentage given the range from 1.6 to 66.32 and a median of 13.55, plausibly at a county or tract level given n=3222. The distribution is right-skewed (skew 2.10, kurtosis 6.89) with 137 outliers (4.25%) sitting well above the Q3 of 17.91, indicating a long tail of high-poverty areas. No nulls and no zeros, and 1719 unique values suggest some rounding to one decimal.

Numeric estimate of SNAP-eligible population per record, with all 3222 rows populated and 2839 unique values. The distribution is extremely right-skewed (skew 14.73, kurtosis 342.21): median is 3799.5 but the mean is 13001.22 and the max reaches 1,343,978, roughly 31x the standard deviation above the mean. About 11.2% of rows (362) flag as outliers, so a small number of very large geographies dominate the tail.

Estimated SNAP participant counts per row, ranging from 2 to 900,465 with a median of 2,546 and mean of 8,710.83. The distribution is severely right-skewed (skew 14.73, kurtosis 342.21) with std 28,987.13 dwarfing the IQR of 5,522, and 362 rows (11.24%) flagged as outliers — likely a few very large geographies pulling the tail. No nulls or zeros, and 2,636 of 3,222 values are unique.