Swedish — Personal Identifiable Information (PII)

CURRENT version:

• RegEx
• Set Lookups from JSON files

FUTURE version:

• Fine-tuned Named Entity Recognition (NER) model https://huggingface.co/FacebookAI/xlm-roberta-base

---

Dataset sources:

• Names: https://github.com/swedishdata/personal-names/tree/master SCB 2020 First names: https://www.statistikdatabasen.scb.se/pxweb/en/ssd/START__BE__BE0001__BE0001G/BE0001FNamn10/ SCB 2020 Last names: https://www.statistikdatabasen.scb.se/pxweb/en/ssd/START__BE__BE0001__BE0001G/BE0001ENamn10/

• Education Programs:

  • https://github.com/swedishdata/education-work-social/blob/master/scb-sun-2000.csv

• Professions:

  • https://github.com/swedishdata/education-work-social/blob/master/arbetsformedlingen-job-titles.csv

• Marital Status:

  • https://github.com/swedishdata/education-work-social/blob/master/scb-family-marital-status-and-consensual-union-terms.csv

• Sexual Orientation:

  • https://glaad.org/reference/terms

• Addresses:

  • Street & Postal (DOESN'T CONTAIN ALL): https://github.com/beshrkayali/sverige_postnummer
  • Municipalities & Counties: https://github.com/swedishdata/geography/blob/master/counties-municipalities.csv

• Most up-to-date addresses:

  • https://www.postnummerservice.se/
  • https://www.postnord.se/en/our-tools/search-postcode-and-address/
  • https://download.geofabrik.de/europe/sweden.html

WIP extraction from OpenStreetMap data:

osmium tags-filter sweden-latest.osm.pbf -o addresses.osm.pbf n/addr:street n/addr:postcode n/addr:city n/admin_level=7 n/admin_level=4 w/addr:street w/addr:postcode w/addr:city w/admin_level=7 w/admin_level=4

osmium export addresses.osm.pbf -f geojson -o addresses.geojson

jq -r '.features[] | select(.properties["addr:street"] != null and .properties["addr:postcode"] != null and .properties["addr:city"] != null) | "\(.properties["addr:street"]),\(.properties["addr:postcode"]),\(.properties["addr:city"])"' addresses.geojson | sort | uniq | sed -E 's/([0-9]{3})([0-9]{2})/\1 \2/' > unique_addresses.txt

jq -r '.features[] | select(.properties.admin_level == "7" and .properties.name != null) | "\(.properties.name)"' addresses.geojson | sort | uniq > municipalities.txt

jq -r '.features[] | select(.properties.admin_level == "4" and .properties.name != null) | "\(.properties.name)"' addresses.geojson | sort | uniq > counties.txt

---

Identifiers & labels

• Person First Name (PER-FIRST)
• Person Last Name (PER-LAST)
• Personnummer (ID-PNR)
• Samordningsnummer (ID-SNR)
  —
• Marital Status (MARITAL)
• Biological Sex (SEX)
• Nationality (NATION)
  —
• Education Program (EDU-PROGRAM)
• Profession (PROF)
  —
• Disabilities (DISAB)
• Ethnicity (ETHNIC)
• Sexual Orientation (SEXOR)
  —
• Political Opinions (POL)
• Religious Beliefs (REL)
  —
• Phone Number (PHONE)
• Email (EMAIL)
• Social Media Profiles (SOCM)
  —
• Street Address (ADDR-STREET)
• Postal Code (ADDR-POSTAL)
• Municipality (ADDR-MUNICIPALITY)
• City (ADDR-CITY)
• County (ADDR-COUNTY)
  —
• Bank Account Number (FIN-BANKNUM)
• IBAN (FIN-IBAN)
• BIC/SWIFT Code (FIN-BIC)
• Credit Card Number (FIN-CC)
  —
• Organization Number (ORG-NUM)
• Company Name (ORG-WORK)
• Education Institute (ORG-EDU)
  —
• IP Address (IP)
• MAC Address (MAC)
  —
• Date (DATE)
• Time (TIME)
  —
• Vehicle Registration Number (VEH)