Update dataset (9 interviews)

README.md

@@ -1,70 +1,59 @@
 ---
 license: cc-by-4.0
 language:
-  - en
   - sv
 tags:
   - oral-history
-  - swedish-american
-  - emigration
   - biography
   - interview
+  - life-narrative
 size_categories:
   - n<1K
 ---
 
-# Setterdahl Swedish Emigrant Interview Corpus
+# Sommar & Vinter i P1 Corpus
 
-Processed interviews from the Lennart Setterdahl collection at Kulturparken Småland (Växjö, Sweden). Setterdahl interviewed ~3,000 first-generation Swedish immigrants across the United States between 1960–1990.
+Personal narrative monologues from Sveriges Radio's iconic Sommar i P1 program. Public figures tell their life story in ~55-minute episodes. Running since 1959.
 
 ## Dataset Summary
 
-- **Interviews processed:** 9
-- **Source:** [DigitalMuseum](https://digitaltmuseum.se) (Kulturparken Småland, S-KPS)
-- **License:** CC-BY (original audio), generated content by Wundrblock
-- **Languages:** English (primary), Swedish (code-switching)
-- **Time period:** Interviews recorded 1960s–1990s, subjects born N/A
+- **Records:** 9
+- **Languages:** sv
 
 ## Key Statistics
 
 | Metric | Value |
 |--------|-------|
-| K-factor (mean unique people/interview) | N/A |
+| Records | 9 |
+| With transcript | 9 |
+| With entities | 0 |
+| K-factor (mean) | N/A |
 | K-factor (median) | N/A |
-| K-factor (range) | N/A |
 | Total processing cost | $0.00 |
-| Avg cost/interview | $0.00 |
 
 ## Fields
 
 | Field | Type | Description |
 |-------|------|-------------|
-| interview_id | string | LS identifier (e.g., "LS 0238") |
-| subject_name | string | Interviewee name |
-| birth_year | int | Subject's birth year |
-| birth_place | string | Subject's birthplace (Sweden) |
+| interview_id | string | Unique identifier |
+| source_id | string | Source-specific ID |
+| subject_name | string | Subject/speaker name |
+| birth_year | int | Birth year (if known) |
+| birth_place | string | Birthplace (if known) |
 | interview_year | int | Year of recording |
 | duration_minutes | float | Audio duration |
 | transcript_words | int | Word count |
-| transcript_corrected | string | Full corrected transcript with speaker labels |
-| entities | list | Extracted people and places with mention counts |
+| transcript_corrected | string | Full transcript |
+| entities | list | Extracted people and places |
 | k_factor | int | Unique people mentioned |
-| biography_markdown | string | Generated biography in Pandoc markdown (if available) |
-| biography_words | int | Biography word count |
-| processing_costs | list | Per-step cost breakdown |
-| total_cost_usd | float | Total processing cost |
-| audio_url | string | Direct MP3 download URL |
-| dimu_url | string | DigitalMuseum page URL |
+| biography_markdown | string | Generated biography (if available) |
+| audio_url | string | Audio download URL (if available) |
+| source_url | string | Source page URL |
 
-## Pipeline
+## Citation
 
-1. Audio download from DigitalMuseum (CC-BY)
-2. Transcription via OpenAI Whisper
-3. Speaker labeling via Claude Haiku
-4. Transcript correction via Claude Sonnet (Swedish names/places)
-5. Entity extraction via Claude Sonnet
-6. Biography generation via Claude Sonnet (optional)
+Source audio from Sveriges Radio (sverigesradio.se/sommar-i-p1). Transcription and processing by Wundrblock.
 
-## Citation
+## Processing
 
-This dataset is derived from the Lennart Setterdahl interview collection held by Kulturparken Småland / Svenska Emigrantinstitutet, Växjö, Sweden. Original recordings are CC-BY licensed.
+Processed by [Wundrblock](https://wundrblock.com) using Claude (Anthropic) and Whisper (OpenAI).