Image Acquisition Analysis Report Executive Summary for 168-Record Dataset

This report analyses image acquisition results across the 168 news article drafts compiled for hokkaidopress.com through Trial and Batches 1 through 11. The dataset spans publication dates from December 2024 to August 2026.

Headline figures

• Total records analysed: 168.
• Image acquisition success: 104 records, 61.9 percent.
• Records without image: 64 records, 38.1 percent.

Image status distribution

• sourced: 75 records, 44.6 percent. Government or national-institution imagery cleared for publication reuse.
• needs_review: 29 records, 17.3 percent. Commercial or press-release imagery requiring licence confirmation before publication.
• missing: 64 records, 38.1 percent. OGP tag not defined or source is a direct PDF link.

Key findings

Batch-level success rates ranged from 0 percent to 100 percent, indicating that source selection was the single largest determinant of image acquisition outcomes. Batches 1 through 3 and Batch 8, which drew primarily from visit-hokkaido.jp and ainu-upopoy.go.jp, achieved 100 percent acquisition. Batches 4 through 7, which drew heavily from Hokkaido subprefectural bureau PDFs, dropped to between 0 and 15 percent. Batches 10 and 11 recovered to 75 to 85 percent by deliberately prioritising OGP-compliant sources.

Practical recommendation

Continuing the Batch 10 and 11 source-selection strategy is expected to raise the overall acquisition rate above 80 percent in future batches. Supplementary detail is available in the companion records covering domain-level, category-level and batch-level breakdowns.

Companion records:

• hbtp-image-analysis-breakdown-2026-08 (detailed breakdowns).
• hbtp-image-analysis-recommendations-2026-08 (improvement plan).