Other pages in this series (34)
Token shares of the weather cell and agreement with two datasets
Each cell decomposed with the pre-registered token rule R1 into day-level flags (clear-only, cloudy, rain, snow, mixed), by reign and lunar month, and the same-day agreement with the Lyu et al. 2024 dummies and the Lee et al. 2024 category.
Three quarters of the days carry nothing but '晴' (76,308 with the bare character, 181 with a spelling variant such as 〈晴〉 or 朝晴). This site's rule-based decomposition agrees with the Lyu et al. 2024 dummies on 99.5–100.0% of days and with the Lee et al. 2024 category on 97.4–99.97%; the remaining 776 and 1,258 differences are published as lists. These are entries in a secretariat form, not Seoul's weather.
- Formula
- Class tokens (晴 陰 雨 雪 雹 霧 雷 風 霜) plus time and degree modifiers decompose all 896 distinct strings with zero residue; rain_any etc. are true if present anywhere in the day (precip_any = rain, snow or hail). Denominators are weather-bearing day nodes (101,793; a date with two nodes counts twice). Agreement = days with the same flag / days in both (Lyu joined on the NIKH day ID, 102,671 days; Lee on the Gregorian date, 45,674 of the 47,240 matched days that carry category 1, 3, 7 or 8 — 1,566 missing or no-document days excluded; of the four dates that appear twice in Lee's file, the row with a gauge value is kept).
- How to read it
- The shares describe entry practice. Agreement in the 99% range means the decomposition is stable, not that the weather was right. The difference lists (f_compare_*_diff) carry the original strings and take no side.
- Caveats
- Lyu et al. 2024 digitized the same cell first. What is new here is the published rule and the three-way check; the rule was not adjusted after seeing the results.
- External check
- Lyu et al. 2024: 102,689 rows in the Zenodo release (doi:10.5281/zenodo.8382243, CC BY 4.0), while the paper (§6) states 100,620 days with weather records; Lee et al. 2024 (Zenodo, CC BY 4.0), 47,480 days (Methods, quality assurance).
- Reliability card
- externalExternal match — agreement with other datasets and observations
Same NIKH day ID as Lyu et al. 2024 on 102,671 days: clear 99.53%, cloudy 99.83%, rain 99.90%, snow 99.98% agreement; identical strings on 98.3%, identical Gregorian date 99.94%.
Lee et al. 2024, joined on the Gregorian date, 47,240 days (1778–1907): clear 97.4%, rain 99.83%, snow 99.97%. Lee's single category per day disagrees on mixed days by construction.
| Reign | Days | With weather | 晴 only | Cloudy | Rain | Snow | Mixed | Blank |
|---|---|---|---|---|---|---|---|---|
| Injo | 9,060 | 8,857 | 71.6% | 15.3% | 9.7% | 1.9% | 6.9% | 203 |
| Hyojong | 3,576 | 3,523 | 74.7% | 18.5% | 6.8% | 1.1% | 3.4% | 53 |
| Hyeonjong | 5,485 | 5,412 | 73.9% | 18.3% | 8.3% | 1.5% | 6.3% | 73 |
| Sukjong | 16,104 | 15,984 | 75.7% | 18.7% | 5.6% | 0.9% | 5.1% | 120 |
| Gyeongjong | 1,527 | 1,513 | 69.9% | 17.3% | 13.4% | 3.5% | 19.7% | 14 |
| Yeongjo | 19,556 | 19,358 | 68.5% | 15.8% | 16.9% | 3.2% | 17.5% | 187 |
| Jeongjo | 8,760 | 8,718 | 76.0% | 8.6% | 16.1% | 1.3% | 5.1% | 38 |
| Sunjo | 12,518 | 12,509 | 77.2% | 8.4% | 14.3% | 0.8% | 3.2% | 5 |
| Heonjong | 5,198 | 5,193 | 73.9% | 8.5% | 15.8% | 2.0% | 2.8% | 3 |
| Cheoljong | 5,263 | 5,108 | 80.1% | 3.9% | 15.9% | 0.3% | 2.3% | 155 |
| Gojong | 15,904 | 14,510 | 82.5% | 4.3% | 13.5% | 0.4% | 2.6% | 1,390 |
| Sunjong | 1,123 | 1,108 | 82.0% | 9.1% | 8.2% | 0.7% | 0.4% | 15 |
| Lunar month | Days | 晴 only | Rain | Snow | Thunder | Fog |
|---|---|---|---|---|---|---|
| 1.0 | 8261.0 | 0.8219343905096236 | 0.0370415203970463 | 0.0401888391235927 | 0.0001210507202517 | 0.0007263043215107 |
| 2.0 | 8451.0 | 0.775529523133357 | 0.0853153472961779 | 0.0186960123062359 | 0.0 | 0.000591645959058 |
| 3.0 | 8802.0 | 0.7530106793910475 | 0.1325835037491479 | 0.0029538741195182 | 0.0002272210861167 | 0.0001136105430583 |
| 4.0 | 8895.0 | 0.7695334457560428 | 0.1323215289488476 | 0.0001124227093872 | 0.0001124227093872 | 0.0005621135469364 |
| 5.0 | 8798.0 | 0.643669015685383 | 0.2111843600818367 | 0.0 | 0.0001136621959536 | 0.0003409865878608 |
| 6.0 | 8616.0 | 0.5572191272051996 | 0.2862116991643454 | 0.0 | 0.0002321262766945 | 0.0003481894150417 |
| 7.0 | 8560.0 | 0.6508177570093457 | 0.2378504672897196 | 0.0 | 0.0003504672897196 | 0.0008177570093457 |
| 8.0 | 8334.0 | 0.7853371730261579 | 0.1241900647948164 | 0.0001199904007679 | 0.0005999520038396 | 0.0007199424046076 |
| 9.0 | 8341.0 | 0.82184390360868 | 0.0980697758062582 | 0.0017983455221196 | 0.0010790073132717 | 0.0011988970147464 |
| 10.0 | 8282.0 | 0.7971504467519923 | 0.0879014730741366 | 0.0198019801980198 | 0.0015696691620381 | 0.0007244626901714 |
| 11.0 | 8255.0 | 0.8110236220472441 | 0.0387643852210781 | 0.0488188976377952 | 0.0001211387038158 | 0.0009691096305269 |
| 12.0 | 8198.0 | 0.8465479385215906 | 0.0189070505001219 | 0.0492803122712856 | 0.0001219809709685 | 0.000243961941937 |