glam/docs/BELGIAN_ISIL_DATASET_FINDINGS.md

# Belgian ISIL Dataset Findings

**Date**: November 17, 2025  
**Status**: ✅ REGISTRIES FOUND - NO BULK DOWNLOAD AVAILABLE

---

## Executive Summary

Belgium has **TWO separate ISIL registries** managed by different authorities, but **NEITHER offers a public bulk CSV/Excel download**. However, both have web search interfaces that can be systematically queried.

---

## Belgian ISIL Registry Structure

### Authority Split

Belgium uses a **dual delegation model**:

| Sector | Authority | Website | ISIL Prefix |
|--------|-----------|---------|-------------|
| **Libraries** (bibliotheken/bibliothèques) | KBR - Koninklijke Bibliotheek / Bibliothèque royale de Belgique | https://isil.kbr.be/ | BE-[A-Z]{3}[0-9]{2} |
| **Archives** (archieven/archives) | Rijksarchief / Archives de l'État | http://isil.arch.be/ | BE-A[0-9]{4} |

**Note**: Museums can register with either authority depending on their primary function.

---

## Dataset 1: KBR ISIL Registry (Libraries)

### Overview

- **Managing Authority**: Koninklijke Bibliotheek van België / Bibliothèque royale de Belgique (KBR)
- **Primary Coverage**: Public libraries, research libraries, museum libraries, specialized libraries
- **Format**: Web database (no bulk export)
- **Access**: https://isil.kbr.be/search.php?lang=en

### Institution Types Covered

- Public libraries (openbare bibliotheken, bibliothèques publiques)
- University libraries (universiteitsbibliotheken, bibliothèques universitaires)
- Research center libraries
- Museum libraries and documentation centers
- Specialized libraries (law, medicine, art, etc.)

### Sample ISIL Codes (Extracted from Search)

```
BE-ANN00 - Antwerp (various institutions)
BE-ANN01 through BE-ANN24 - Antwerp sub-institutions
BE-BRL00 - Brussels institutions
BE-BRL07 - Archives & Bibliothèques de l'Université libre de Bruxelles
BE-BRL10 - Agentschap Kunsten en Erfgoed - Collectie van de Vlaamse Gemeenschap
BE-BRL13 - Archives-Musée du CPAS de Bruxelles
BE-GET08 - Arteveldehogeschool (AHS)
BE-TOI01 - Bibliothèque d'architecture (Tournai)
BE-EKN00 - Antwerp International School (AIS)
BE-TEN00 - Koninklijk Museum voor Midden-Afrika (Tervuren)
```

### Data Fields Available

Based on search results, each ISIL record includes:
- ISIL code (BE-XXXXX format)
- Institution name (official name)
- Alternative names/acronyms
- Walk-up address (street address)
- City/location
- Telephone (area code + number)
- Institution type
- Communication details

### Access Methods

**✅ Available**:
- **Web Search Interface**: https://isil.kbr.be/search.php?lang=en
  - Search by: ISIL code, name, location
  - Results displayed in HTML tables
  - Individual record detail pages

**❌ NOT Available**:
- Bulk CSV/Excel export
- API endpoint
- RSS/Atom feed
- SPARQL endpoint

**🔧 Possible Approach**:
- Systematic web scraping of search results
- Empty search query returns all records: `https://isil.kbr.be/search.php?lang=en&query=`
- Estimated records: 500-1,000 institutions

---

## Dataset 2: Rijksarchief ISIL Registry (Archives)

### Overview

- **Managing Authority**: Rijksarchief in België / Archives de l'État en Belgique
- **Primary Coverage**: Archives (government, corporate, ecclesiastical, private)
- **Format**: Web database (no bulk export)
- **Access**: http://isil.arch.be/

### Institution Types Covered

- National archives (Algemeen Rijksarchief / Archives générales du Royaume)
- Regional/provincial state archives (Rijksarchief in de Provinciën)
- Municipal archives (Stadsarchieven / Archives communales)
- Ecclesiastical archives (church, abbey, seminary archives)
- Corporate archives (bedrijfsarchieven / archives d'entreprise)
- Private archives (particuliere collecties / archives privées)

### Sample ISIL Codes (Extracted from Search)

```
BE-A0510 - Algemeen Rijksarchief (ARA) - National Archives
BE-A2000 through BE-A2003 - Regional state archives
BE-A3000 through BE-A3006 - Provincial state archives
BE-A4000 - Archief en Documentatiecentrum voor het Vlaams-nationalisme (ADVN)
BE-A4001 - Archive (unspecified)
BE-A4002 - Archive (unspecified)
BE-A4003 - Archief voor Hedendaagse Kunst in België, KMSKB (Contemporary Art Archive)
BE-A5000 - Archives de l'Evêché de Liège (Liège Diocesan Archives)
BE-A5001 - Archief Grootseminarie Brugge (AGSB) - Bruges Seminary Archives
BE-A5002 - Aartsbisschoppelijk Archief Mechelen (AAM) - Archdiocese of Mechelen Archives
```

### Data Fields Available

Based on web interface analysis:
- ISIL code (BE-AXXXX format)
- Institution name (Naam/Nom)
- Location (Plaats/Localité)
- Institution type (archive designation)
- Contact information (login required)

### Access Methods

**✅ Available**:
- **Web Search Interface**: http://isil.arch.be/?view=searchisil
  - Public search by ISIL code or name
  - Login system for institutions to manage their records
  - "Already have an ISIL code?" page: http://isil.arch.be/?view=bestaande

**❌ NOT Available**:
- Bulk CSV/Excel export
- API endpoint
- RSS/Atom feed
- Full directory listing

**🔧 Possible Approach**:
- Web scraping of public search results
- Institution login required for full details
- Estimated records: 100-300 archives

---

## ISIL Code Format Patterns

### KBR Library Codes

**Format**: `BE-[CITY_CODE][INST_NUM]`

Examples:
- `BE-ANN00` - Antwerp (ANN = Antwerpen) base institution
- `BE-BRL07` - Brussels (BRL = Bruxelles) institution #7
- `BE-GET08` - Ghent (GET = Gent) institution #8
- `BE-TEN00` - Tervuren (TEN = Tervuren) base institution

**Pattern**:
- 3-letter city code (geographic identifier)
- 2-digit institution number (00 = primary institution, 01+ = additional)

### Rijksarchief Archive Codes

**Format**: `BE-A[TYPE][NUM]`

Examples:
- `BE-A0510` - National Archives (05 = national level, 10 = sequence)
- `BE-A2000` - Regional state archives (20 = regional level)
- `BE-A3000` - Provincial state archives (30 = provincial level)
- `BE-A4000` - Specialized/private archives (40 = non-governmental)
- `BE-A5000` - Ecclesiastical archives (50 = church institutions)

**Pattern**:
- Letter 'A' = Archive
- First 1-2 digits = Archive type/level
- Remaining digits = Sequential number

---

## Comparison with Dutch ISIL System

| Feature | Netherlands | Belgium |
|---------|-------------|---------|
| **Managing Authority** | Nationaal Archief (archives), KB (libraries) | Rijksarchief (archives), KBR (libraries) |
| **ISIL Format** | NL-[CityCode][InstType] (alphanumeric) | BE-[Code][Num] (mixed alphanumeric/numeric) |
| **Bulk Export** | ✅ CSV available (archives) | ❌ NOT AVAILABLE |
| **Public Search** | ✅ Available | ✅ Available |
| **Documentation** | ✅ Good | ⚠️ Limited |

---

## Data Acquisition Strategy

### Option 1: Contact Authorities (RECOMMENDED)

**KBR Contact**:
- Email: isil@kbr.be
- Contact person: Imke Hansen
- Address: Keizerslaan 4, 1000 Brussel
- Phone: +32 (0)2 519 57 41

**Request**:
> Subject: Request for Belgian ISIL Registry Dataset
> 
> Dear KBR ISIL Office,
> 
> I am working on a global heritage custodian data project (https://github.com/kempersc/glam) 
> that aims to create a comprehensive, open dataset of GLAM institutions worldwide using 
> LinkML and Linked Data standards.
> 
> We have successfully integrated ISIL datasets from the Netherlands (Nationaal Archief) 
> and are now seeking to include Belgian heritage institutions.
> 
> Could you provide a bulk export (CSV, Excel, or JSON) of the Belgian ISIL registry 
> maintained by KBR? We would also appreciate any similar export from the Rijksarchief 
> archive registry.
> 
> If bulk export is not available, would you permit systematic web scraping of the 
> public search interface at https://isil.kbr.be/search.php for research purposes?
> 
> The resulting dataset will be published under an open license with full attribution 
> to KBR and Rijksarchief as authoritative sources.
> 
> Thank you for your consideration.

**Rijksarchief Contact**:
- Website: http://arch.arch.be
- General inquiry: http://arch.arch.be/index.php?l=en&m=contact
- Same request as above, adapted for archives

### Option 2: Web Scraping (Fallback)

**Technical Approach**:

1. **KBR Library Registry**:
   ```python
   # Scrape all ISIL records from KBR
   base_url = "https://isil.kbr.be/search.php?lang=en&query="
   
   # Empty query returns all records
   response = requests.get(base_url)
   
   # Parse HTML tables
   soup = BeautifulSoup(response.text, 'html.parser')
   
   # Extract ISIL codes, names, locations
   for row in soup.find_all('a', href=re.compile('data.php')):
       isil_code = extract_code(row)
       fetch_detail_page(isil_code)
   ```

2. **Rijksarchief Archive Registry**:
   ```python
   # Scrape archive ISIL records
   base_url = "http://isil.arch.be/?view=searchisil"
   
   # Search form submissions
   # Extract BE-A#### codes from results
   ```

3. **Rate Limiting**:
   - Respectful delay: 2-3 seconds between requests
   - User-Agent header identifying project
   - Log all requests for reproducibility

4. **Data Validation**:
   - Cross-check with Wikidata (P791 ISIL code property)
   - Geocode addresses using Nominatim
   - Verify institution types

### Option 3: Wikidata Enrichment

**Query Wikidata for Belgian institutions with ISIL codes**:

```sparql
SELECT ?item ?itemLabel ?isil ?viaf ?location WHERE {
  ?item wdt:P791 ?isil .            # Has ISIL code
  FILTER(STRSTARTS(?isil, "BE-"))   # Belgian institutions
  OPTIONAL { ?item wdt:P214 ?viaf } # VIAF ID
  OPTIONAL { ?item wdt:P131 ?location } # Located in
  SERVICE wikibase:label { bd:serviceParam wikibase:language "en,nl,fr" }
}
```

**Current Wikidata Coverage** (as of Nov 2025):
- Estimated: 50-100 Belgian heritage institutions with ISIL codes
- Quality: Variable (some records incomplete)
- Usefulness: Good starting point, but not comprehensive

---

## Estimated Dataset Size

| Registry | Estimated Institutions | Completeness |
|----------|----------------------|--------------|
| KBR Libraries | 500-1,000 | High (required for public libraries) |
| Rijksarchief Archives | 100-300 | Medium (many archives unregistered) |
| **Total** | **600-1,300** | **Medium-High** |

**Comparison**:
- Netherlands ISIL registry: ~364 institutions (Nationaal Archief archives only)
- Belgium would add significant European coverage

---

## Integration with GLAM Project

### Schema Mapping

Belgian ISIL records map to the GLAM project's LinkML schema as follows:

**Core HeritageCustodian Fields**:
```yaml
- id: https://w3id.org/heritage/custodian/be/ann00
  name: "Institution Name" (from ISIL registry)
  institution_type: LIBRARY | ARCHIVE (from registry sector)
  alternative_names: ["Acronym", "Alternative Name"]
  
  locations:
    - city: "Brussels" (from ISIL Plaats/Localité field)
      street_address: "Walk-up address" (from ISIL registry)
      country: "BE"
      geonames_id: null  # To be geocoded
  
  identifiers:
    - identifier_scheme: "ISIL"
      identifier_value: "BE-BRL07"
      identifier_url: "https://isil.kbr.be/data.php?lang=en&id=BRL07"
  
  provenance:
    data_source: CSV_REGISTRY
    data_tier: TIER_1_AUTHORITATIVE
    extraction_date: "2025-11-17T..."
    source_url: "https://isil.kbr.be/"
    verified_by: "KBR ISIL Office"
```

### GHCID Generation

Belgian GHCID format: `BE-[REGION]-[CITY]-[TYPE]-[ABBREV]`

Examples:
- `BE-BRU-BRX-L-KBR` - KBR, Brussels (Library)
- `BE-VLG-ANT-M-MAF` - Museum aan de Stroom, Antwerp (Museum)
- `BE-WAL-LIE-A-AEL` - Archives de l'Evêché de Liège (Archive)

**Region Codes**:
- BRU = Brussels-Capital Region (Brussels Hoofdstedelijk Gewest)
- VLG = Flemish Region (Vlaams Gewest)
- WAL = Walloon Region (Région wallonne)

---

## Timeline and Milestones

### Phase 1: Data Request (Week 1)
- ✅ Contact KBR ISIL Office (isil@kbr.be)
- ✅ Contact Rijksarchief
- ⏳ Await response (1-2 weeks)

### Phase 2: Acquisition (Weeks 2-3)
- If bulk export provided: Process CSV/Excel
- If scraping required: Implement scraper scripts
- Fallback: Enhance Wikidata SPARQL query

### Phase 3: Processing (Week 4)
- Parse Belgian ISIL records
- Geocode addresses
- Generate GHCIDs
- Cross-link with Wikidata

### Phase 4: Validation (Week 5)
- Validate against schema
- Check for duplicates
- Compare with OpenStreetMap data
- Manual review of uncertain records

### Phase 5: Integration (Week 6)
- Merge with existing GLAM dataset
- Export to RDF/JSON-LD
- Update documentation
- Create Belgium-specific report

---

## Related Resources

### Belgian Heritage Ecosystem

**National Aggregators**:
- **MetaBelgica**: Linked Data platform for Federal Scientific Institutes (https://www.kbr.be/en/projects/metabelgica/)
- **Erfgoedkaart**: Heritage map of Belgium (archives, museums, libraries)
- **Archives Portal Europe**: Includes Belgian archives

**Museum Networks**:
- FARO - Vlaams steunpunt voor cultureel erfgoed (Flemish cultural heritage)
- Federation Wallonie-Bruxelles heritage services

**University Libraries**:
- LIBIS (Flemish university libraries consortium)
- CIUF (Walloon-Brussels university libraries)

### Identifier Systems

| System | Coverage | Belgian Usage |
|--------|----------|---------------|
| ISIL | Libraries, archives, museums | ✅ Widely adopted |
| VIAF | Libraries, archives | ✅ Used by KBR, university libraries |
| Wikidata | All heritage institutions | ⚠️ Partial coverage |
| GRID | Research institutions | ✅ Universities covered |
| ROR | Research organizations | ✅ Growing adoption |

---

## Key Findings Summary

### ✅ What We Found

1. **Two Official ISIL Registries**:
   - KBR (libraries): https://isil.kbr.be/
   - Rijksarchief (archives): http://isil.arch.be/

2. **Web Search Interfaces Available**:
   - Both registries have public search
   - Individual record detail pages
   - Estimated 600-1,300 total institutions

3. **Clear Authority Split**:
   - Libraries → KBR
   - Archives → Rijksarchief
   - Museums → Either registry (depends on function)

4. **Multilingual Data**:
   - Dutch (Flemish), French, German, English
   - Institution names in multiple languages

### ❌ What's Missing

1. **No Bulk Export**:
   - No CSV/Excel download available
   - No public API endpoint
   - No SPARQL/Linked Data access

2. **Limited Public Access**:
   - Full details may require institution login
   - Some contact information restricted
   - No data license specified

3. **Incomplete Coverage**:
   - Not all Belgian archives registered
   - Small museums may lack ISIL codes
   - Private collections underrepresented

### 🔧 Recommended Next Steps

1. **Immediate**: Send data request emails to KBR and Rijksarchief
2. **Parallel**: Query Wikidata for existing Belgian ISIL codes
3. **Fallback**: Prepare web scraping scripts if bulk export denied
4. **Long-term**: Contribute enhanced data back to Wikidata

---

## Questions for User

1. **Should I send the data request emails to KBR and Rijksarchief now?**
   - Draft emails are ready above
   - Can be sent via automated email or you can send manually

2. **Should I implement the web scraping fallback?**
   - Respectful scraping with rate limiting
   - Only if bulk export request is denied
   - Requires ~2-3 hours development time

3. **Should I query Wikidata for Belgian institutions now?**
   - Quick SPARQL query to get existing data
   - Can enrich with ISIL codes later
   - Good baseline for comparison

4. **Which Belgian regions should we prioritize?**
   - Brussels-Capital (central, multilingual)
   - Flanders (largest population, most institutions)
   - Wallonia (French-speaking region)
   - All regions (comprehensive coverage)

---

**Next Steps**: Awaiting your decision on data acquisition strategy.