Skip to main content

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 A.D.I. project draws on eight CSV dataset files organized into two tiers. The first tier consists of five THOR source files in the /data/ directory — declassified records from the U.S. Department of Defense’s Theater History of Operations Reports database, covering Allied aerial bombing operations throughout World War II. The second tier consists of three ADI master files at the repository root — cleaned and enriched subsets produced by A.D.I.’s Jupyter notebook EDA pipeline, scoped to the chronological phases of the Normandy campaign (August 1943 through June 6, 1944). Only the three ADI master files are ingested by the Power BI semantic model.

​
ADI Master Datasets

These three files at the repository root are the direct inputs to the desembarco_normandia_dashboard.SemanticModel. Each corresponds to a specific dashboard page and a distinct phase of the campaign timeline.
Purpose: The primary analytical dataset for the June 6, 1944 D-Day operations. This is the most fully enriched of the three master files — it contains all standard THOR mission fields plus three columns derived by A.D.I.’s pipeline: he_tons, frag_tons, and weapon_type. The weapon_type classification categorizes each mission’s ordnance mix into human-readable Spanish-language labels such as "Ataque Combinado (HE + Fragmentación)", "Solo Explosivos", or "Solo Fragmentación".Location: Repository root — adi_dataset_master.csvPower BI table: adi_dataset_masterDashboard page: “Día D” (D-Day analysis view)Source encoding: UTF-8 (65001), 49 columnsColumn list (from CSV header):
Mission ID, Mission Date, Theater of Operations, Country, Air Force, Unit ID,
Aircraft Series, Callsign, Mission Type, Takeoff Base, Takeoff Location,
Takeoff Latitude, Takeoff Longitude, Target ID, Target Country, Target City,
Target Type, Target Industry, Target Priority, Target Latitude, Target Longitude,
Altitude (Hundreds of Feet), Airborne Aircraft, Attacking Aircraft,
Bombing Aircraft, Aircraft Returned, Aircraft Failed, Aircraft Damaged,
Aircraft Lost, High Explosives, High Explosives Type,
High Explosives Weight (Pounds), High Explosives Weight (Tons),
Incendiary Devices, Incendiary Devices Type,
Incendiary Devices Weight (Pounds), Incendiary Devices Weight (Tons),
Fragmentation Devices, Fragmentation Devices Type,
Fragmentation Devices Weight (Pounds), Fragmentation Devices Weight (Tons),
Total Weight (Pounds), Total Weight (Tons), Time Over Target,
Bomb Damage Assessment, Source ID, he_tons, frag_tons, weapon_type
A.D.I. enrichment columns:
ColumnDescription
he_tonsDerived HE tonnage (int64) — extracted and normalized from raw weight fields
frag_tonsDerived fragmentation tonnage (int64) — extracted and normalized
weapon_typeSpanish-language classification label for the ordnance mix
Sample record: Mission 125300, June 6 1944, ETO, USA, 15 AF, B-24, Target: FRANCE/UNKNOWN/TACTICAL TARGET, 60 total tons (weapon_type: “Ataque Combinado (HE + Fragmentación)”)
Purpose: A cleaned subset of THOR operations covering the Allied strategic bombing campaign in the European and Mediterranean theaters from August 1943 through June 5, 1944 — the entire pre-invasion build-up period. This dataset captures USAAF and RAF missions across France, Belgium, and the Netherlands, tracking the progressive isolation of Normandy’s transport and logistics network.Location: Repository root — adi_master_1943_1944_CLEAN.csvPower BI table: adi_master_1943_1944Dashboard page: “Panorama 1943–1944” (global ETO/MTO overview)Source encoding: Windows-1252 (1252), 46 columnsDate range: 1943-08-15 to 1944-06-05Column list (from CSV header):
Mission ID, Mission Date, Theater of Operations, Country, Air Force, Unit ID,
Aircraft Series, Callsign, Mission Type, Takeoff Base, Takeoff Location,
Takeoff Latitude, Takeoff Longitude, Target ID, Target Country, Target City,
Target Type, Target Industry, Target Priority, Target Latitude, Target Longitude,
Altitude (Hundreds of Feet), Airborne Aircraft, Attacking Aircraft,
Bombing Aircraft, Aircraft Returned, Aircraft Failed, Aircraft Damaged,
Aircraft Lost, High Explosives, High Explosives Type,
High Explosives Weight (Pounds), High Explosives Weight (Tons),
Incendiary Devices, Incendiary Devices Type,
Incendiary Devices Weight (Pounds), Incendiary Devices Weight (Tons),
Fragmentation Devices, Fragmentation Devices Type,
Fragmentation Devices Weight (Pounds), Fragmentation Devices Weight (Tons),
Total Weight (Pounds), Total Weight (Tons), Time Over Target,
Bomb Damage Assessment, Source ID
Notable data characteristics:
  • Theater focus: ETO (European Theater of Operations) and MTO (Mediterranean Theater of Operations)
  • Primary target countries: France, Germany, Belgium, Netherlands
  • Aircraft series: predominantly B-17 and B-24 heavy bombers
  • Sample record: Mission 17, 1943-08-15, ETO, B-17, AMIENS/FRANCE/AIRDROME, 5 tons HE
Archaeological relevance: This dataset is the macro-level view showing the macro-objectives and zones of high bomb impact density across the European theater. It helps archaeological survey teams identify regions warranting deep prospection beyond the Normandy coastal strip.
Purpose: A tightly scoped dataset covering only the five days immediately before D-Day (June 1–5, 1944) — the critical phase A.D.I. designates as the “logistical softening” of the Normandy beachhead. During this window, Allied air forces concentrated their bombing on transport infrastructure — bridges over the Seine and Loire rivers, rail marshalling yards, and fuel depots — to sever German supply lines to the landing zones.Location: Repository root — adi_master_1_5_junio.csvPower BI table: adi_master_1_5_junioDashboard page: “1 A 5 JUNIO” (June 1–5 detailed analysis)Source encoding: Windows-1252 (1252), 46 columnsDate range: 1944-06-01 to 1944-06-05Column list (from CSV header):
Mission ID, Mission Date, Theater of Operations, Country, Air Force, Unit ID,
Aircraft Series, Callsign, Mission Type, Takeoff Base, Takeoff Location,
Takeoff Latitude, Takeoff Longitude, Target ID, Target Country, Target City,
Target Type, Target Industry, Target Priority, Target Latitude, Target Longitude,
Altitude (Hundreds of Feet), Airborne Aircraft, Attacking Aircraft,
Bombing Aircraft, Aircraft Returned, Aircraft Failed, Aircraft Damaged,
Aircraft Lost, High Explosives, High Explosives Type,
High Explosives Weight (Pounds), High Explosives Weight (Tons),
Incendiary Devices, Incendiary Devices Type,
Incendiary Devices Weight (Pounds), Incendiary Devices Weight (Tons),
Fragmentation Devices, Fragmentation Devices Type,
Fragmentation Devices Weight (Pounds), Fragmentation Devices Weight (Tons),
Total Weight (Pounds), Total Weight (Tons), Time Over Target,
Bomb Damage Assessment, Source ID
Notable data characteristics:
  • Theater: ETO exclusively
  • Air forces: primarily 9 AF (Ninth Air Force) and 8 AF tactical bombers
  • Primary target types: GUN EMPLACEMENTS, BRIDGES, RAIL YARDS, MARSHALLING YARDS
  • Sample record: Mission 119255, 1944-06-01, ETO, USA, 9 AF, B-26, LA HAYE/FRANCE/GUN EMPLACEMENTS, 2 tons
Archaeological significance: The bombing density in this window is the primary driver for identifying transport infrastructure remnants and UXO (unexploded ordnance) risk zones in the Norman interior. According to A.D.I.’s business conclusions, excavation efforts should prioritize transport nodes — not just the coastal strip — where bridge and rail destruction was concentrated.

​
THOR Source Files (/data/)

These five files are the unmodified historical source data. They are never loaded directly by Power BI but are used as input by the A.D.I. Jupyter notebook EDA pipeline to generate the three master datasets above.
Description: A subset of THOR aerial incursion records. This file has a strong geographic bias toward the Mediterranean Theater of Operations (MTO), covering the Italy/Sicily campaigns, which made it unsuitable as the primary source for a Normandy-focused analysis. It was reviewed during EDA and excluded from the main Normandy analytical pipeline.Key columns (from CSV structure):
Mission ID, Mission Date, Theater of Operations, Country, Air Force, Unit ID,
Aircraft Series, Takeoff Base, Takeoff Location, Takeoff Latitude,
Takeoff Longitude, Target Country, Target City, Target Latitude,
Target Longitude, Airborne Aircraft, Bombing Aircraft,
High Explosives Type, High Explosives Weight (Tons), Total Weight (Tons)
Data quality note: Theater bias is heavily MTO (Italy, Sicily, North Africa). Zones without data do not imply absence of activity — only absence of registration. Use with caution for any ETO inference.
Description: The broadest source file in the project — the full THOR operations dataset with 46 fields covering the complete WWII mission record format across all theaters. This is the richest raw source and the primary input for generating adi_master_1943_1944_CLEAN.csv and adi_master_1_5_junio.csv.Column list (from CSV header):
Mission ID, Mission Date, Theater of Operations, Country, Air Force, Unit ID,
Aircraft Series, Callsign, Mission Type, Takeoff Base, Takeoff Location,
Takeoff Latitude, Takeoff Longitude, Target ID, Target Country, Target City,
Target Type, Target Industry, Target Priority, Target Latitude, Target Longitude,
Altitude (Hundreds of Feet), Airborne Aircraft, Attacking Aircraft,
Bombing Aircraft, Aircraft Returned, Aircraft Failed, Aircraft Damaged,
Aircraft Lost, High Explosives, High Explosives Type,
High Explosives Weight (Pounds), High Explosives Weight (Tons),
Incendiary Devices, Incendiary Devices Type,
Incendiary Devices Weight (Pounds), Incendiary Devices Weight (Tons),
Fragmentation Devices, Fragmentation Devices Type,
Fragmentation Devices Weight (Pounds), Fragmentation Devices Weight (Tons),
Total Weight (Pounds), Total Weight (Tons), Time Over Target,
Bomb Damage Assessment, Source ID
Key statistics for France-targeted missions:
  • Total missions in France: 15,270
  • Top target cities: Brest (459 missions), St. Lô (404 missions)
  • Maximum recorded tonnage: 999.0 tons — this is likely a placeholder or data-entry error code; treat as unreliable
  • Mean tonnage: ~46.7 tons | Median: ~33 tons
Data quality anomaly: The 999.0 tons maximum value appears to be a sentinel/error code in the THOR source data rather than a real bombing weight. The EDA pipeline filters or flags these values before they propagate into the ADI master files.
Description: A lookup/glossary table mapping THOR Aircraft Series code strings (e.g., B17, B24, P51, A36) to full aircraft model names and specifications. Used during EDA to enrich and validate the Aircraft Series field across all mission tables. Not loaded into the Power BI semantic model directly — the code values from the THOR source appear as-is in the ADI master files.Use case: Cross-reference when interpreting Aircraft Series values in the Power BI visuals — e.g., B26 = Martin B-26 Marauder, B17 = Boeing B-17 Flying Fortress.
Description: A cleaned version of the full THOR WWII dataset covering all theaters of operations globally. Used as a complementary reference dataset during EDA to understand the broader target-type distribution and validate Normandy-specific findings in context.Key statistics (full dataset):
Target TypeMission Count
AIRDROME2,157
UNIDENTIFIED TARGET1,181 + 574 (two variants)
CITY AREA539
Data quality note: Some targets appear as "Unknown" or "Unidentified", which reduces classification precision for target type analysis. This is an inherent limitation of the THOR source records — not all missions were fully documented at time of filing.
Description: A lookup table mapping THOR ordnance code strings (e.g., "500 LB GP (GP-M43/M64)") to descriptive names for bomb and weapon types. Used during EDA to decode the High Explosives Type, Incendiary Devices Type, and Fragmentation Devices Type fields. Cross-reference this file when the weapon_type classification in adi_dataset_master.csv needs to be traced back to its underlying THOR ordnance codes.Use case: Supports UXO (unexploded ordnance) risk assessments — identifying which specific bomb models were used in a given target area allows archaeological safety teams to reference known fuze types and failure rates for those ordnance specifications.

Build docs developers (and LLMs) love

Get started for freeTalk to us