Jacopo
Peracchio
I investigate hybrid threats, organized crime and illicit networks with reproducible pipelines built entirely from open sources — every claim traceable, every dataset auditable.
Nothing on this page is classified — hover the bars to verify.
Key judgements
- Proof of work: a DOI-cited electoral-data pipeline and a local-first knowledge-graph engine, both in production — not portfolio pieces.
- Field-tested method: network analysis of a live criminal-economics case, the Prato district (MA thesis, 110/110 cum laude).
- Looking for: senior analyst roles where method rigor and tradecraft are the product.
§ 01
Research Programmes
Four standing lines of work where social theory meets measurable, reproducible evidence.
-
HT-01
Hybrid Threats
Influence operations, coordinated inauthentic behaviour, astroturfing, disinformation dynamics.
-
OC-02
Organized Crime
Social network analysis of illicit ecosystems, money-laundering typologies, transnational flows.
-
TAG-03
Terrorism & Armed Groups
Mobilisation dynamics, radicalisation pathways, network structures of armed organisations.
-
SOO-04
Social Organisation Online
Astroturfing, coordinated inauthentic behaviour and synthetic influence networks: how online collectives organise, signal and mobilise.
§ 02
Method & Tradecraft
Mixed-methods tradecraft: ethnographic fieldwork and qualitative interviewing combined with computational analysis — developed on a live case study of transnational organised crime in the Prato industrial district (MA thesis, 110/110 cum laude) and applied since in OSINT collection and analysis pipelines.
Web scraping & data pipelines (Python)
Social network analysis (Gephi, ForceAtlas2, PageRank, modularity)
Machine learning & regression (R, Cook's distance, Breusch-Pagan)
OSINT: dorking, court records, Wayback Machine, digital footprint
Qualitative: ethnography, key-informant interviews
LLM multi-agent systems & sociological theory
Local-first & privacy-preserving AI orchestration
Source evaluation
Admiralty-style reliability and credibility ratings applied to every source before it enters a dataset.
Chain of custody
Provenance tracked from collection to final product — every figure traces back to its origin.
Reproducibility
Pipelines are versioned, tested and CI-checked. Results are re-runnable, not screenshots.
Privacy by design
Public sources only. Collection and storage are GDPR-aware and local-first where it matters.
Core stack
- Python
- R
- Gephi / networkx
- pandas
- SQLite
- Web scraping
- OSINT
- SNA
- Machine learning
- LLM agents
- Docker
- GitHub Actions
§ 03
Tools & Analytic Products
Built for real collection and analysis work — documented, tested, and running on hardware you control.
-
Tool · CLI
TallyHo
Italian electoral time series from 1946 to today, municipality by municipality, straight from ministerial open data. Python CLI exporting CSV/JSON/XLSX/Parquet.
-
Engine
Arachne-Scholar
Local-first knowledge-graph engine for academic literature: PDF ingestion, SVO dependency parsing, network metrics, OCR. Runs fully offline on your own hardware.
-
Research code
hybrid-kg-paper
Experiments in hybrid knowledge-graph construction: deterministic NLP backbone with LLM enrichment. Reproducibility GED 0.0.
-
Working paper
Socio-Economic Spatial Analysis of Savelli
Computational sociological study of demographic, administrative and spatial structures across the Sila municipalities.
-
MA thesis
Economic Networks and Illicit Practices in the Prato District
Sociological analysis of the Chinese presence in Prato between mafia-type infiltration and community self-regulation — fieldwork and network analysis on a live case.
§ 04
Featured Dossier — Savelli (KR), 1996–2023
Analytic product · OS
Bottom line
- Municipal-election turnout in Savelli fell from 68.5% (1996) to 49.3% (2023) — a 19-point drop that crossed the 50% line in 2018 and has stayed under it.
- The registered electorate peaked at 1,892 in 2013 and has since fallen to 1,662 — disengagement compounds depopulation.
- Every figure below was collected from the Ministry of the Interior portal by TallyHo (see § 03) — the same pipeline, re-runnable by anyone.
Fig. 05 — Turnout, municipal elections 1996–2023. Dashed line: 50%.
Method
Collection via TallyHo v1.0.0 against the ministerial historical-elections portal: 163 electoral dates scanned, 12 consultations retrieved for the municipality, zero errors.
Why it mattersSavelli is one of the small Sila municipalities under study in the working paper below (§ 03). Turnout decline and electorate shrinkage are the demographic backbone of that analysis.
Verify itThe full series — candidates, votes, lists — is one command away in the TallyHo repository.
SOURCE — elezionistorico.interno.gov.it (DAIT, Ministero dell’Interno), municipal elections, Savelli (KR) · extracted 19 Aug 2026 with TallyHo v1.0.0, cache 51% · processing: dedup by date, first-round turnout = votanti/elettori
§ 05
Education
2023–2026
Università degli Studi di Torino
MA in Sociology and Social Research
110/110 cum laude — thesis on illicit economic networks in the Prato district (full entry under Analytic products).
2025
Università degli Studi di Torino, Future Scenarios Lab
Certificate, Apr–Jun 2025
Frameworks for future-scenario analysis: which factors and events correlate and show causal links with a phenomenon in its future development. Skills: information analysis, risk management.
2025
Università degli Studi di Torino, Algorithms & Society Lab
Certificate, Feb–Apr 2025
2019–2023
Università degli Studi di Firenze
BA in Political Science
103/110 — thesis: “New Technologies, Old Inequalities: a Critical Analysis of Digital Human Resources”.
§ 06
Languages
Italian
Native
English
B2 · CEFR
Spanish
B1 · CEFR
Chinese
A2 · CEFR
Public work is written in English; fieldwork is conducted in Italian.