Research, patent, capital and search-interest signals for engineering, computing and business innovation in Bangladesh, 2017–2025 (Fliar BI Prism)

Aggregate tables behind the article "Wired, Built and Sustained: Research, Patent, Capital and Search Signals in Bangladeshi Engineering, Computing and Business Innovation, 2017–2025" (Kumar, 2026). All counts were produced from Fliar BI Prism, an index maintained by Fliar BI that links scholarly works, patent publications, reported investment rounds and search-interest series, and were taken on 2026-10-23 with the full index available. The deposit gives the yearly and city-level research counts, the world baseline, every patent family with its CPC codes, the search-interest series, the theme summary, the exact theme and affiliation rules, and the figures printed in the article, so that every number in the article can be checked. research_by_year.csv: National papers per year, theme paper counts and theme share of national output (%) research_world_by_year.csv: World baseline: papers per year and theme paper counts, same rules research_by_city.csv: Papers and theme papers by city or division of the affiliated institution, 2017–2025 patent_families.csv: One row per patent family with a national assignee: family id, publication number, year, title, CPC codes investment_rounds.csv: Investment rounds reported in business news: company, round type, amount, currency, sector label, city, year search_interest_monthly.csv: Monthly relative search interest (0–100) within the country for each theme term search_interest_by_region.csv: Relative search interest by division for each theme term theme_summary.csv: Direction of each theme in each stream, as reported in the article's summary table theme_rules.json: Affiliation, city, exclusion and theme rules (title patterns, CPC prefixes, search terms) and the direction rule figures/: Figures in the article Method: Cross-Domain Signal Triangulation (CDST). Each theme is defined once and measured separately in each stream; directions come from a log-linear fit over 2017–2025. See README.md.