Archaios Data Intelligence (A.D.I.) is a data consultancy platform that converts declassified U.S. Department of Defense aerial bombing records into actionable geospatial intelligence for archaeological field teams. The platform was built to serve a specific and practical need: before any excavation campaign in Normandy can begin, directors must know where to dig, what hazards to expect, and how to allocate limited prospection budgets across a landscape that was profoundly reshaped by industrial-scale aerial warfare between 1943 and 1944. A.D.I. answers those questions by ingesting the THOR (Theater History of Operations Reports) database, cleaning and enriching it through a reproducible Python EDA pipeline, and surfacing the results as an interactive Power BI dashboard. The intended audience spans archaeological directors planning field campaigns, data analysts who want to extend or replicate the methodology, and heritage consultancies that need a defensible, evidence-based prioritisation tool for regulatory or funding submissions.Documentation Index
Fetch the complete documentation index at: https://mintlify.com/HelenDiMo/archaios-data-Intelligence/llms.txt
Use this file to discover all available pages before exploring further.
The Mission
Operación Normandía is the core analytical initiative within A.D.I. Its goal is to transform the historical record of Allied aerial operations — covering the period from 1943 through June 6, 1944 (D-Day) — into a model of geospatial intelligence for the Normandy region of France. The strategic logic is straightforward: the locations of the most intense Allied bombing activity are, by definition, the locations most likely to contain buried material remains of military significance. Destroyed bunkers, aircraft fuselages, impact craters, and the debris of shattered logistics infrastructure do not disappear — they subside into the soil. Eighty years on, those concentrations of disturbed ground and ferrous metal represent the highest-yield targets for systematic archaeological survey. By mapping bombing density, ordnance tonnage, mission frequency, and target-type distribution, the A.D.I. dashboard allows excavation directors to:- Prioritise survey zones based on objective density metrics rather than folklore or anecdotal field reports.
- Minimise prospection costs by eliminating low-probability areas from the survey plan before any ground work begins.
- Maximise the scientific discovery rate by focusing limited field days on statistically high-yield zones.
- Flag unexploded ordnance (UXO) risk at the planning stage, so field teams can apply appropriate clearance protocols before breaking ground.
What is THOR?
The Theater History of Operations Reports (THOR) database is a U.S. Department of Defense project that digitised tens of thousands of handwritten and typed paper mission reports from the Second World War. The result is a structured, machine-readable record of Allied air power from 1939 to 1945. THOR aggregates data from multiple Allied air forces, principally:- U.S. Army Air Forces (USAAF) — the dominant contributor to the Northwest European and Mediterranean theaters.
- Royal Air Force (RAF) — covering Bomber Command, Coastal Command, and tactical operations.
- Commonwealth air forces — including missions from Australian, New Zealand, and South African units operating under RAF and USAAF command structures.
How A.D.I. Uses THOR
The A.D.I. pipeline transforms five raw THOR source files into a finished Power BI dashboard through four reproducible stages.Raw THOR Data Ingestion
Five source CSV files stored in the
Each record in
/data/ directory form the foundation of the pipeline. These files are ingested without modification to preserve the integrity of the primary source record:| File | Contents |
|---|---|
1_cleanbombww2.csv | Bombing sorties — a pre-cleaned THOR extract (1,109 records, 21 variables) |
2_operations.csv | Allied operations register |
3_thor_wwii_aircraft_gloss.csv | Aircraft type glossary |
4_thor_wwii_data_clean.csv | Full THOR data clean derived from 1_cleanbombww2.csv |
5_thor_wwii_weapon_gloss.csv | Bomb and weapon type glossary |
1_cleanbombww2.csv carries fields including Mission ID, Mission Date, Aircraft Series, Takeoff Base, Takeoff Latitude/Longitude, Target City, Target Latitude/Longitude, Airborne Aircraft, Bombing Aircraft, High Explosives Type, High Explosives Weight (Tons), and Total Weight (Tons).Data Cleaning and Enrichment
The
desembarco_normandia.ipynb Jupyter notebook performs the full EDA (Exploratory Data Analysis) and enrichment pass across all five source datasets. The notebook uses pandas and numpy to handle type casting, null imputation, deduplication, and temporal filtering. It also applies geographic filters to isolate the Normandy theater and joins the glossary tables to resolve aircraft and ordnance codes into human-readable labels.The output of this stage is three ADI master datasets stored at the repository root — one per chronological phase of the campaign — which serve as the direct data sources for the Power BI report.Semantic Modelling
The three ADI master CSV files feed a Power BI Fabric semantic model composed of three tables, one per campaign phase. The model defines typed columns, calculated measures (total tonnage, mission count, bombing aircraft density), and the relationships needed to support cross-filter interaction across all dashboard visuals, including the geospatial heat map layer.The semantic model is embedded inside
wwii_analisys_dashboard.pbix at the repository root and requires no external database or cloud connection — it reads directly from local CSV files.Interactive Dashboard Delivery
The finished report exposes three dashboard views, each corresponding to a distinct phase of the air campaign:
- 1943–1944 — a strategic panorama of the full European Theater of Operations, identifying macro-level target concentrations and high-impact zones across the entire campaign window.
- 1 A 5 JUNIO — the five-day pre-D-Day logistical softening phase (June 1–5, 1944), showing the intense targeting of inland transport infrastructure — bridges, rail yards, road junctions — designed to isolate the Normandy beaches.
- DIA-D — June 6, 1944, the D-Day assault itself: a visual record of one of the most operationally significant and historically consequential days of aerial warfare.
Project Deliverables
Power BI Dashboard
The primary deliverable:
wwii_analisys_dashboard.pbix at the repository root, containing the embedded semantic model and all three campaign-phase report views. Also published as a Microsoft Fabric report for stakeholder sharing.EDA Notebook
desembarco_normandia.ipynb — the full exploratory data analysis notebook. Documents every cleaning decision, enrichment step, and analytical derivation applied to the five THOR source files. Fully reproducible with standard Python data science libraries.ADI Master Datasets
Three enriched CSV files at the repository root:
adi_dataset_master.csv, adi_master_1943_1944_CLEAN.csv, and adi_master_1_5_junio.csv. These are the cleaned, phase-filtered, and enriched tables that Power BI reads directly.THOR Source Data
Five raw and pre-cleaned CSV files in
/data/: the unmodified THOR extracts that form the authoritative primary source layer for the entire pipeline. Preserved without alteration to allow full traceability from raw record to dashboard visual.Archaeological Value
The bombing data surfaces two findings with direct, concrete value for any field team planning operations in Normandy. Cost reduction through target-zone precision. A common assumption is that D-Day archaeology concentrates on the coastline — the beaches, the cliffs, the beach obstacles. The A.D.I. analysis of the June 1–5 logistical softening phase challenges that assumption directly. The data shows that Allied air power in the five days before D-Day was directed overwhelmingly at inland transport nodes: bridges over the Seine and Loire rivers, railway marshalling yards, and road junctions throughout the Norman interior. The bomb tonnage density in these inland zones was massive — in many cases exceeding that of the coastal assault areas. For archaeological directors, this means that a coastline-first survey strategy systematically underweights the interior, where the density of buried structural remains (destroyed bridge abutments, cratered rail infrastructure, aircraft wreckage from flak-downed bombers) is likely to be very high. Field safety through UXO risk mapping. Bomb tonnage density is a direct proxy for the probability of unexploded ordnance in the soil. Not every bomb dropped detonated: dud rates for WWII-era ordnance varied by fuse type, soil conditions, and delivery angle, and estimates of residual UXO in Northern France remain significant. By mapping the spatial distribution of total bomb weight per square kilometre, the A.D.I. dashboard gives field teams a pre-survey UXO risk layer that can inform decisions about ground-penetrating radar deployment, ordnance clearance contracting, and safe excavation sequencing before a single shovel enters the ground.This project is a consultancy deliverable, not a live API or installable library. The documentation guides you through understanding, opening, and extending the data assets.