Source: /home/coolhand/html/datavis/data_trove/data/quirky/aurora.json
Saturn profiled 300 rows across 8 columns. The stats below are deterministic and machine-readable; the prose is a language-model interpretation of those stats (opt-in, added after the fact, never sees raw rows).
!pip install saturn-dissect
import subprocess
subprocess.run([
"saturn", "analyze", "/home/coolhand/html/datavis/data_trove/data/quirky/aurora.json",
"--findings", "quirky-aurora.json",
"--llm", "anthropic:claude-opus-4-7",
])This dataset captures 300 minute-by-minute aurora and solar wind observations starting 2026-01-20, with 8 columns covering geomagnetic indices (kp_index, estimated_kp, intensity), solar wind conditions (speed, density), and a categorical activity label. The activity field is heavily skewed toward 'Moderate Storm' (172 of 300, ~57%), with only 8 'Quiet' readings — worth a closer look since this dominates the storyline. The kp_index and intensity columns are left-skewed and pile up at their max values (median equals max), with ~15% flagged as low-side outliers, suggesting the sample is a sustained storm period rather than a balanced range. Solar wind speed is also unusually elevated (min 881, max 1051 km/s), reinforcing that this is a storm-window snapshot rather than typical conditions.
citing: activity.top_values · activity.top_rate · kp_index.stats · estimated_kp.stats · intensity.stats · solar_wind_speed.stats · solar_wind_density.stats · row_count
saturn.schema() · 8 columns
| column | kind | n | null% | unique | alerts |
|---|---|---|---|---|---|
| time_tag | categorical | 300 | 0.0% | 300 | long_tail |
| kp_index | numeric | 300 | 0.0% | 7 | outliers |
| estimated_kp | numeric | 300 | 0.0% | 18 | outliers |
| activity | categorical | 300 | 0.0% | 5 | |
| intensity | numeric | 300 | 0.0% | 7 | outliers |
| solar_wind_time | categorical | 300 | 0.0% | 299 | long_tail |
| solar_wind_speed | numeric | 300 | 0.0% | 266 | |
| solar_wind_density | numeric | 300 | 0.0% | 215 |
| value | count | share |
|---|---|---|
| Moderate Storm | 172 | 57.3% |
| Minor Storm | 65 | 21.7% |
| Unsettled | 38 | 12.7% |
| Active | 17 | 5.7% |
| Quiet | 8 | 2.7% |
| bin | count |
|---|---|
| 0 – 0.3529 | 4 |
| 0.3529 – 0.7059 | 0 |
| 0.7059 – 1.059 | 4 |
| 1.059 – 1.412 | 0 |
| 1.412 – 1.765 | 0 |
| 1.765 – 2.118 | 16 |
| 2.118 – 2.471 | 0 |
| 2.471 – 2.824 | 0 |
| 2.824 – 3.176 | 22 |
| 3.176 – 3.529 | 0 |
| 3.529 – 3.882 | 0 |
| 3.882 – 4.235 | 17 |
| 4.235 – 4.588 | 0 |
| 4.588 – 4.941 | 0 |
| 4.941 – 5.294 | 65 |
| 5.294 – 5.647 | 0 |
| 5.647 – 6 | 172 |
| bin | count |
|---|---|
| 0 – 0.03924 | 4 |
| 0.03924 – 0.07847 | 0 |
| 0.07847 – 0.1177 | 4 |
| 0.1177 – 0.1569 | 0 |
| 0.1569 – 0.1962 | 0 |
| 0.1962 – 0.2354 | 16 |
| 0.2354 – 0.2746 | 0 |
| 0.2746 – 0.3139 | 0 |
| 0.3139 – 0.3531 | 22 |
| 0.3531 – 0.3924 | 0 |
| 0.3924 – 0.4316 | 0 |
| 0.4316 – 0.4708 | 17 |
| 0.4708 – 0.5101 | 0 |
| 0.5101 – 0.5493 | 0 |
| 0.5493 – 0.5885 | 65 |
| 0.5885 – 0.6278 | 0 |
| 0.6278 – 0.667 | 172 |
| bin | count |
|---|---|
| 881.6 – 891.6 | 6 |
| 891.6 – 901.6 | 24 |
| 901.6 – 911.5 | 68 |
| 911.5 – 921.5 | 22 |
| 921.5 – 931.5 | 17 |
| 931.5 – 941.5 | 14 |
| 941.5 – 951.5 | 10 |
| 951.5 – 961.5 | 19 |
| 961.5 – 971.4 | 14 |
| 971.4 – 981.4 | 9 |
| 981.4 – 991.4 | 6 |
| 991.4 – 1001 | 7 |
| 1001 – 1011 | 11 |
| 1011 – 1021 | 14 |
| 1021 – 1031 | 18 |
| 1031 – 1041 | 31 |
| 1041 – 1051 | 10 |
| bin | count |
|---|---|
| 0.23 – 0.7059 | 6 |
| 0.7059 – 1.182 | 15 |
| 1.182 – 1.658 | 28 |
| 1.658 – 2.134 | 29 |
| 2.134 – 2.609 | 37 |
| 2.609 – 3.085 | 33 |
| 3.085 – 3.561 | 34 |
| 3.561 – 4.037 | 38 |
| 4.037 – 4.513 | 34 |
| 4.513 – 4.989 | 31 |
| 4.989 – 5.465 | 6 |
| 5.465 – 5.941 | 4 |
| 5.941 – 6.416 | 1 |
| 6.416 – 6.892 | 2 |
| 6.892 – 7.368 | 1 |
| 7.368 – 7.844 | 0 |
| 7.844 – 8.32 | 1 |
| column | kind | null % |
|---|---|---|
| time_tag | categorical | 0.0% |
| kp_index | numeric | 0.0% |
| estimated_kp | numeric | 0.0% |
| activity | categorical | 0.0% |
| intensity | numeric | 0.0% |
| solar_wind_time | categorical | 0.0% |
| solar_wind_speed | numeric | 0.0% |
| solar_wind_density | numeric | 0.0% |
| kp_index | estimated_kp | intensity | solar_wind_speed | solar_wind_density | |
|---|---|---|---|---|---|
| kp_index | +1.00 | +0.98 | +1.00 | +0.60 | +0.44 |
| estimated_kp | +0.98 | +1.00 | +0.98 | +0.65 | +0.50 |
| intensity | +1.00 | +0.98 | +1.00 | +0.60 | +0.44 |
| solar_wind_speed | +0.60 | +0.65 | +0.60 | +1.00 | +0.53 |
| solar_wind_density | +0.44 | +0.50 | +0.44 | +0.53 | +1.00 |
This is an ISO-8601 timestamp column at minute resolution, with all 300 values unique (entropy_ratio 1.0) and zero nulls. The visible top values form a contiguous one-minute sequence starting 2026-01-20T00:06:00, consistent with a regular time index rather than event timestamps. Cardinality equals row count, so it acts as a row key over time.
Treatment: Parse to datetime and use as the time index; do not one-hot encode.
saturn.columns["time_tag"].stats
| stat | value |
|---|---|
| n | 300 |
| nulls | 0 (0.0%) |
| unique | 300 |
| top_value | 2026-01-20T00:06:00 |
| top_rate | 0.003333 |
| cardinality | 300 |
| entropy | 8.229 |
| entropy_ratio | 1 |
| alert: long_tail | 300 singleton categories |
time_tag.
| value | count | share |
|---|---|---|
| 2026-01-20T00:06:00 | 1 | 0.3% |
| 2026-01-20T00:07:00 | 1 | 0.3% |
| 2026-01-20T00:08:00 | 1 | 0.3% |
| 2026-01-20T00:09:00 | 1 | 0.3% |
| 2026-01-20T00:10:00 | 1 | 0.3% |
| 2026-01-20T00:11:00 | 1 | 0.3% |
| 2026-01-20T00:12:00 | 1 | 0.3% |
| 2026-01-20T00:13:00 | 1 | 0.3% |
| 2026-01-20T00:14:00 | 1 | 0.3% |
| 2026-01-20T00:15:00 | 1 | 0.3% |
| 2026-01-20T00:16:00 | 1 | 0.3% |
| 2026-01-20T00:17:00 | 1 | 0.3% |
| 2026-01-20T00:18:00 | 1 | 0.3% |
| 2026-01-20T00:19:00 | 1 | 0.3% |
| 2026-01-20T00:20:00 | 1 | 0.3% |
| 2026-01-20T00:21:00 | 1 | 0.3% |
| 2026-01-20T00:22:00 | 1 | 0.3% |
| 2026-01-20T00:23:00 | 1 | 0.3% |
| 2026-01-20T00:24:00 | 1 | 0.3% |
| 2026-01-20T00:25:00 | 1 | 0.3% |
Numeric column on a small 0-6 integer scale with only 7 distinct values, consistent with a Kp geomagnetic index reading. The distribution is heavily concentrated at the high end (median and Q3 both 6.0, mean 5.09) and strongly left-skewed (skew -1.68), with 46 low-side outliers flagged (15.3%). No nulls, and zeros are rare (1.3%).
Treatment: Treat as ordinal categorical (7 levels) rather than continuous; outliers are real low-Kp readings, do not clip.
saturn.columns["kp_index"].stats
| stat | value |
|---|---|
| n | 300 |
| nulls | 0 (0.0%) |
| unique | 7 |
| min | 0 |
| max | 6 |
| mean | 5.09 |
| median | 6 |
| std | 1.396 |
| q1 | 5 |
| q3 | 6 |
| iqr | 1 |
| skew | -1.683 |
| kurtosis | 2.149 |
| n_outliers | 46 |
| outlier_rate | 0.1533 |
| zero_rate | 0.01333 |
| alert: outliers | 15.3% rows beyond 1.5 IQR |
kp_index. Vertical dash marks the median.
| bin | count |
|---|---|
| 0 – 0.3529 | 4 |
| 0.3529 – 0.7059 | 0 |
| 0.7059 – 1.059 | 4 |
| 1.059 – 1.412 | 0 |
| 1.412 – 1.765 | 0 |
| 1.765 – 2.118 | 16 |
| 2.118 – 2.471 | 0 |
| 2.471 – 2.824 | 0 |
| 2.824 – 3.176 | 22 |
| 3.176 – 3.529 | 0 |
| 3.529 – 3.882 | 0 |
| 3.882 – 4.235 | 17 |
| 4.235 – 4.588 | 0 |
| 4.588 – 4.941 | 0 |
| 4.941 – 5.294 | 65 |
| 5.294 – 5.647 | 0 |
| 5.647 – 6 | 172 |
This looks like an estimated Kp geomagnetic index value, bounded between 0.0 and 6.33 across 300 rows with only 18 unique values, consistent with the discrete third-step Kp scale. The distribution is heavily left-skewed (skew -1.30) with the median pinned at the maximum 6.33 and Q3 also at 6.33, meaning at least half the rows sit at the ceiling. 18 low-side outliers (6% rate) and a near-zero zero_rate (0.67%) confirm a long thin tail toward quiet conditions.
Treatment: Treat as ordinal/discrete and consider binning or a rank transform before modelling given the ceiling-heavy left skew.
saturn.columns["estimated_kp"].stats
| stat | value |
|---|---|
| n | 300 |
| nulls | 0 (0.0%) |
| unique | 18 |
| min | 0 |
| max | 6.33 |
| mean | 5.183 |
| median | 6.33 |
| std | 1.538 |
| q1 | 4.67 |
| q3 | 6.33 |
| iqr | 1.66 |
| skew | -1.295 |
| kurtosis | 0.9485 |
| n_outliers | 18 |
| outlier_rate | 0.06 |
| zero_rate | 0.006667 |
| alert: outliers | 6.0% rows beyond 1.5 IQR |
estimated_kp. Vertical dash marks the median.
| bin | count |
|---|---|
| 0 – 0.3724 | 4 |
| 0.3724 – 0.7447 | 2 |
| 0.7447 – 1.117 | 2 |
| 1.117 – 1.489 | 0 |
| 1.489 – 1.862 | 6 |
| 1.862 – 2.234 | 4 |
| 2.234 – 2.606 | 6 |
| 2.606 – 2.979 | 6 |
| 2.979 – 3.351 | 16 |
| 3.351 – 3.724 | 10 |
| 3.724 – 4.096 | 4 |
| 4.096 – 4.468 | 3 |
| 4.468 – 4.841 | 63 |
| 4.841 – 5.213 | 2 |
| 5.213 – 5.585 | 0 |
| 5.585 – 5.958 | 4 |
| 5.958 – 6.33 | 168 |
Categorical descriptor of geomagnetic or space-weather activity level, with 5 ordered classes from 'Quiet' to 'Moderate Storm'. The distribution is heavily skewed toward storm conditions: 'Moderate Storm' alone accounts for 57.3% (172/300) of rows, while 'Quiet' appears just 8 times, suggesting the dataset is filtered to disturbed periods rather than representing typical activity. Entropy ratio of 0.73 confirms moderate concentration on the top class.
Treatment: Treat as ordinal target; consider class-weighting or stratified sampling given the imbalance toward 'Moderate Storm'.
saturn.columns["activity"].stats
| stat | value |
|---|---|
| n | 300 |
| nulls | 0 (0.0%) |
| unique | 5 |
| top_value | Moderate Storm |
| top_rate | 0.5733 |
| cardinality | 5 |
| entropy | 1.69 |
| entropy_ratio | 0.7278 |
activity.
| value | count | share |
|---|---|---|
| Moderate Storm | 172 | 57.3% |
| Minor Storm | 65 | 21.7% |
| Unsettled | 38 | 12.7% |
| Active | 17 | 5.7% |
| Quiet | 8 | 2.7% |
Numeric 'intensity' column bounded between 0.0 and 0.667 with only 7 unique values across 300 rows, suggesting a discretised or quantised measurement rather than a continuous reading. Distribution is heavily left-skewed (skew -1.68) with median equal to the max (0.667), and 46 rows (15.3%) flagged as outliers on the low end. The handful of zeros (1.3%) and the ceiling at 0.667 hint at a capped or normalised score.
Treatment: Treat as ordinal/categorical given only 7 unique values, or bin explicitly before modelling.
saturn.columns["intensity"].stats
| stat | value |
|---|---|
| n | 300 |
| nulls | 0 (0.0%) |
| unique | 7 |
| min | 0 |
| max | 0.667 |
| mean | 0.5658 |
| median | 0.667 |
| std | 0.1553 |
| q1 | 0.556 |
| q3 | 0.667 |
| iqr | 0.111 |
| skew | -1.682 |
| kurtosis | 2.139 |
| n_outliers | 46 |
| outlier_rate | 0.1533 |
| zero_rate | 0.01333 |
| alert: outliers | 15.3% rows beyond 1.5 IQR |
intensity. Vertical dash marks the median.
| bin | count |
|---|---|
| 0 – 0.03924 | 4 |
| 0.03924 – 0.07847 | 0 |
| 0.07847 – 0.1177 | 4 |
| 0.1177 – 0.1569 | 0 |
| 0.1569 – 0.1962 | 0 |
| 0.1962 – 0.2354 | 16 |
| 0.2354 – 0.2746 | 0 |
| 0.2746 – 0.3139 | 0 |
| 0.3139 – 0.3531 | 22 |
| 0.3531 – 0.3924 | 0 |
| 0.3924 – 0.4316 | 0 |
| 0.4316 – 0.4708 | 17 |
| 0.4708 – 0.5101 | 0 |
| 0.5101 – 0.5493 | 0 |
| 0.5493 – 0.5885 | 65 |
| 0.5885 – 0.6278 | 0 |
| 0.6278 – 0.667 | 172 |
This column is a minute-resolution timestamp of solar wind observations, all falling on 2026-01-20 with values like '2026-01-20 03:58:00.000'. With 299 unique values across 300 rows and entropy ratio 0.9997, it is effectively a per-row time index; only '2026-01-20 03:58:00.000' repeats (twice). No nulls, but the long_tail alert reflects this near-unique structure rather than meaningful categories.
Treatment: Parse to datetime and use as the time axis; do not treat as a categorical feature.
saturn.columns["solar_wind_time"].stats
| stat | value |
|---|---|
| n | 300 |
| nulls | 0 (0.0%) |
| unique | 299 |
| top_value | 2026-01-20 03:58:00.000 |
| top_rate | 0.006667 |
| cardinality | 299 |
| entropy | 8.222 |
| entropy_ratio | 0.9998 |
| alert: long_tail | 298 singleton categories |
solar_wind_time.
| value | count | share |
|---|---|---|
| 2026-01-20 03:58:00.000 | 2 | 0.7% |
| 2026-01-20 00:06:00.000 | 1 | 0.3% |
| 2026-01-20 00:07:00.000 | 1 | 0.3% |
| 2026-01-20 00:08:00.000 | 1 | 0.3% |
| 2026-01-20 00:09:00.000 | 1 | 0.3% |
| 2026-01-20 00:10:00.000 | 1 | 0.3% |
| 2026-01-20 00:11:00.000 | 1 | 0.3% |
| 2026-01-20 00:12:00.000 | 1 | 0.3% |
| 2026-01-20 00:13:00.000 | 1 | 0.3% |
| 2026-01-20 00:14:00.000 | 1 | 0.3% |
| 2026-01-20 00:15:00.000 | 1 | 0.3% |
| 2026-01-20 00:16:00.000 | 1 | 0.3% |
| 2026-01-20 00:17:00.000 | 1 | 0.3% |
| 2026-01-20 00:18:00.000 | 1 | 0.3% |
| 2026-01-20 00:19:00.000 | 1 | 0.3% |
| 2026-01-20 00:20:00.000 | 1 | 0.3% |
| 2026-01-20 00:21:00.000 | 1 | 0.3% |
| 2026-01-20 00:22:00.000 | 1 | 0.3% |
| 2026-01-20 00:23:00.000 | 1 | 0.3% |
| 2026-01-20 00:24:00.000 | 1 | 0.3% |
Numeric measurements of solar wind speed across 300 records, all populated and tightly clustered between 881.6 and 1051.3 with a mean of 955.79 and median of 940.15. The distribution is mildly right-skewed (0.45) and platykurtic (kurtosis -1.32), suggesting a flatter-than-normal spread with no outliers flagged. With 266 unique values out of 300, the column behaves as a continuous physical feature rather than a categorical one.
Treatment: Use as-is for modelling; optional standardization given the narrow range and mild skew.
saturn.columns["solar_wind_speed"].stats
| stat | value |
|---|---|
| n | 300 |
| nulls | 0 (0.0%) |
| unique | 266 |
| min | 881.6 |
| max | 1051 |
| mean | 955.8 |
| median | 940.1 |
| std | 51.98 |
| q1 | 908.5 |
| q3 | 1008 |
| iqr | 99.8 |
| skew | 0.4506 |
| kurtosis | -1.323 |
| n_outliers | 0 |
| outlier_rate | 0 |
| zero_rate | 0 |
solar_wind_speed. Vertical dash marks the median.
| bin | count |
|---|---|
| 881.6 – 891.6 | 6 |
| 891.6 – 901.6 | 24 |
| 901.6 – 911.5 | 68 |
| 911.5 – 921.5 | 22 |
| 921.5 – 931.5 | 17 |
| 931.5 – 941.5 | 14 |
| 941.5 – 951.5 | 10 |
| 951.5 – 961.5 | 19 |
| 961.5 – 971.4 | 14 |
| 971.4 – 981.4 | 9 |
| 981.4 – 991.4 | 6 |
| 991.4 – 1001 | 7 |
| 1001 – 1011 | 11 |
| 1011 – 1021 | 14 |
| 1021 – 1031 | 18 |
| 1031 – 1041 | 31 |
| 1041 – 1051 | 10 |
This is a numeric feature capturing solar wind density, fully populated across 300 rows with 215 distinct values. The distribution is fairly symmetric (skew 0.26, kurtosis 0.12) with mean 3.10 and median 3.14, ranging from 0.23 to 8.32 and an IQR of 2.02. Only 2 outliers (0.67%) appear, so the column looks clean and well-behaved.
Treatment: Use as-is in modelling; standard scaling is sufficient given the near-symmetric distribution.
saturn.columns["solar_wind_density"].stats
| stat | value |
|---|---|
| n | 300 |
| nulls | 0 (0.0%) |
| unique | 215 |
| min | 0.23 |
| max | 8.32 |
| mean | 3.103 |
| median | 3.14 |
| std | 1.344 |
| q1 | 2.09 |
| q3 | 4.11 |
| iqr | 2.02 |
| skew | 0.2555 |
| kurtosis | 0.1168 |
| n_outliers | 2 |
| outlier_rate | 0.006667 |
| zero_rate | 0 |
solar_wind_density. Vertical dash marks the median.
| bin | count |
|---|---|
| 0.23 – 0.7059 | 6 |
| 0.7059 – 1.182 | 15 |
| 1.182 – 1.658 | 28 |
| 1.658 – 2.134 | 29 |
| 2.134 – 2.609 | 37 |
| 2.609 – 3.085 | 33 |
| 3.085 – 3.561 | 34 |
| 3.561 – 4.037 | 38 |
| 4.037 – 4.513 | 34 |
| 4.513 – 4.989 | 31 |
| 4.989 – 5.465 | 6 |
| 5.465 – 5.941 | 4 |
| 5.941 – 6.416 | 1 |
| 6.416 – 6.892 | 2 |
| 6.892 – 7.368 | 1 |
| 7.368 – 7.844 | 0 |
| 7.844 – 8.32 | 1 |
click to copy
@misc{saturn-quirky-aurora-2026,
author = {Steuber, Luke},
title = {Saturn reading: quirky aurora},
year ={2026},
howpublished = {\url{https://dr.eamer.dev/saturn/view/quirky-aurora}},
note = {Profiled with saturn-dissect v0.2.0, prompt saturn-insight-v2, model anthropic:claude-opus-4-7},
}
Steuber, L. (2026). Saturn reading: quirky aurora. Source: /home/coolhand/html/datavis/data_trove/data/quirky/aurora.json. Profiled with saturn-dissect v0.2.0 (saturn-insight-v2, anthropic:claude-opus-4-7). Retrieved from https://dr.eamer.dev/saturn/view/quirky-aurora