AI-Enhanced Web Crawler for Real-Time Discovery of Emerging Trends in Sustainable Finance
Bibliographic record
Abstract
The evolving field of sustainable finance, which encompassing ESG (Environmental, Social, Governance) investing, green bonds, and climate-related financial disclosures—demands timely, data-driven insights from vast web-based sources such as policy updates, corporate filings, news outlets, and research publications. Manual tracking is no longer feasible due to the volume and velocity of data. Traditional web crawlers are rule-based and lack contextual understanding, leading to poor relevance detection, high false positives, and missed emerging topics. This paper shows how to create an AI-powered web crawler using techniques including Natural Language Processing (NLP), subject modeling, and spotting trends. We use a mix of keyword-driven crawling and BERT/FinBERT transformer-based language models to figure out what is semantically relevant. Emerging keyword recognition takes advantage of the fact that term frequency-inverse document frequency (TF-IDF) changes over time. Temporal trend analysis over time-sliced document sets, on the other hand, uses Latent Dirichlet Allocation (LDA). We put the suggested crawler through its paces on 50,000 URLs. These included blogs established by regulatory entities, sites that simply report on financial news, and blogs run by schools and other organizations. When compared to baseline crawling, it was 87.5% accurate and 82.3% recall. This means it could find new problems with acquiring money in a way that lasts.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.005 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.005 | 0.003 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.002 | 0.002 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".