koodiklinikka/palkkakysely
1
0
Fork 0
mirror of https://github.com/koodiklinikka/palkkakysely.git synced 2026-03-14 00:03:58 +00:00
Code Issues Packages Projects Releases Wiki Activity

Merge pull request #24 from akx/claude/fix-salary-data-errors-5leXg

Browse Source 
Fix salary data entry errors in 2025 survey data
This commit is contained in:
Aarni Koskela
2026-03-13 12:21:39 +02:00
committed by GitHub
parent a6479f4af3 474a413ad9
commit e79aa5d126
1 changed files with 21 additions and 7 deletions
Download Patch File Download Diff File Expand all files Collapse all files

28
pulkka/data_ingest.py
View File

@@ -184,6 +184,27 @@ def read_data() -> pd.DataFrame:
errors="coerce", errors="coerce",
).fillna(0) ).fillna(0)
# Fix known bogus data (before salary synthesis so vuositulot is computed correctly)
df = apply_fixups(
df,
[
# Yearly salary entered in monthly field (confirmed by lomaraha being
# ~50% of base/12, which is the standard Finnish lomaraha ratio)
({ID_COL: "e901f47f4b92bc4a"}, {KKPALKKA_COL: 90000 / 12}),
({ID_COL: "8e20ca36952cc1c7"}, {KKPALKKA_COL: 95000 / 12}),
({ID_COL: "231d88e2c60ba704"}, {KKPALKKA_COL: 91000 / 12}),
({ID_COL: "610c49a8d22c01a6"}, {KKPALKKA_COL: 92881 / 12}),
({ID_COL: "e2df338adcf80f15"}, {KKPALKKA_COL: 56117 / 12}),
# Yearly salary in monthly field (no lomaraha to cross-check, but
# the monthly values are implausible)
({ID_COL: "85f388bb23703a66"}, {KKPALKKA_COL: 90000 / 12}),
({ID_COL: "bd2e597bb1b77994"}, {KKPALKKA_COL: 110000 / 12}),
({ID_COL: "b7c22a67f755f545"}, {KKPALKKA_COL: 110000 / 12}),
# Placeholder lomaraha/bonus values (1 EUR is clearly not real)
({ID_COL: "fdfb08998ac86dee"}, {LOMARAHA_COL: 0, BONUS_COL: 0}),
],
)
# Fold commission into monthly salary so KKPALKKA = base + commission # Fold commission into monthly salary so KKPALKKA = base + commission
df[KKPALKKA_COL] = ( df[KKPALKKA_COL] = (
pd.to_numeric(df[KKPALKKA_COL], errors="coerce").fillna(0) + df[COMMISSION_COL] pd.to_numeric(df[KKPALKKA_COL], errors="coerce").fillna(0) + df[COMMISSION_COL]
@@ -219,13 +240,6 @@ def read_data() -> pd.DataFrame:
# Round työvuodet # Round työvuodet
df[TYOKOKEMUS_COL] = df[TYOKOKEMUS_COL].round() df[TYOKOKEMUS_COL] = df[TYOKOKEMUS_COL].round()
# Fix known bogus data
df = apply_fixups(
df,
[
# ({ID_COL: "..."}, {VUOSITULOT_COL: 62000}),
],
)
# Fill in Vuositulot as 12.5 * Kk-tulot if empty # Fill in Vuositulot as 12.5 * Kk-tulot if empty
df[VUOSITULOT_COL] = df.apply(map_vuositulot, axis=1) df[VUOSITULOT_COL] = df.apply(map_vuositulot, axis=1)