OpenConstructionEstimate-DDC-CWICR

DDC CWICR - Construction Work Items, Components & Resources
+ Pipelines n8n for calculating estimates based on descriptions, photos, and CAD (BIM)

🇬🇧 English • 🇨🇳 中文 • 🇪🇸 Español • 🇧🇷 Português • 🇷🇺 Русский • 🇯🇵 日本語 • 🇩🇪 Deutsch • 🇫🇷 Français

⚡ n8n Workflows

Choose your input → Get cost estimate

📝 Text Quick scope-to-estimate from a short description Input: Telegram / chat message Output: Matched work items + estimate 📖 Documentation

📷 Photo / PDF Site photos, scanned BOQ, photo-PDF from the field Input: Image or PDF pages Output: Extracted scope → estimate 📖 Photo Docs · 📖 Universal Bot

🧊 CAD / BIM Revit / IFC / DWG-based quantification & estimating Input: Model export Output: 4D/5D estimate + breakdown 📖 Documentation

🏗️ OpenConstructionERP — Use This Database End-to-End

Open-source ERP that ships this database pre-loaded. Professional BoQ editor, AI-powered photo / PDF / CAD takeoff, 4D/5D planning, dashboards — all running on the 30-track CWICR cost data below.

Get up and running in 30 seconds

pip install --upgrade openconstructionerp
openestimate                      # opens the desktop app at http://localhost:8080

The app launches with the 30 CWICR country tracks pre-installed — start a new project, pick a region, and the BoQ editor instantly suggests rates from ddc_de_berlin, ddc_au_sydney, ddc_hr_zagreb, etc. with proper localised prices and translated work descriptions.

Feature previews

AI Photo → Estimate Snap a site photo, GPT-4 Vision builds a BoQ	BoQ Editor Excel-like editor backed by 55 K rates × 30 tracks
PDF Takeoff Auto-extract quantities from architectural PDFs	4D Schedule × BIM Tasks linked to IFC objects, costs flow to dashboards

DataDrivenConstruction clients and users

📑 Table of Contents

🤖 AI Integration

• Perfect Fuel for AI — Why this database is ideal for AI
• Claude Code & Google Antigravity — AI coding assistants
• n8n — Workflow automation
• Universal Use Cases — What you can build

📊 Database & Data

• About — What is DDC CWICR
• Available Formats — Excel, Parquet, CSV, Qdrant
• Data Schema — 85 fields structure
• Field Groups — Classification, Resources, Labor, Machinery
• Methodology — Resource-based costing principles
• Historical Context — 100+ years of standards

⚡ n8n Workflows

• n8n Workflows Overview — Choose your input type
• Try It Now — Live Demo Bots — Test instantly in Telegram
• Workflow 1: Text Estimator Bot — Telegram bot for text input
• Workflow 2: Photo Cost Estimator — Web form with AI Vision
• Workflow 3: Universal Bot — Text + Photo + PDF
• Workflow 4: CAD/BIM Pipeline — Revit/IFC/DWG to estimate
• Workflows Quick Start — Setup in 4 steps
• n8n 2.0+ Setup — Enable Execute Command node

🏗️ CAD/BIM Pipeline

• Prerequisites — Required components
• Pipeline Stages — 10-stage processing
• LLM Model Selection — OpenAI, Claude, Gemini, Grok
• Output Files — HTML & Excel reports
• Troubleshooting — Common issues

🔍 Vector Database

• Vector Database — Semantic search with Qdrant
• Qdrant Snapshots — Pre-built vector databases
• Collections — 11 language collections
• Docker Deployment — Self-hosted setup

🌐 API

• Pricing Search API — Free REST API for construction pricing
• API Endpoints — Search, Languages, Statistics
• Code Examples — cURL, Python, JavaScript

🚀 Getting Started

• Quick Start - Python — Tabular data & semantic search
• Developer Examples — Python, JavaScript, Rust, R, Shell, n8n guides
• Integration Use Cases — Entry to Advanced level

👥 Community

• Resources & Community — Links & channels
• Consulting & Training — Professional services
• Contributing — Submit your workflows
• License — Dual (CC BY 4.0 data / Apache-2.0 code)
• Security — Responsible disclosure
• Support the Project — Sponsor & donate
• 🤖 AI Instructions — Documentation for AI assistants

🚀 Perfect Fuel for Your AI Products

Just clone the repo and describe what you want — AI does the rest

DDC CWICR is not just a database — it's ready-to-use fuel for AI-powered applications. Whether you're building cost estimation bots, automating construction workflows, or creating intelligent assistants — this data works out of the box with modern AI tools.

Why This Database is Ideal for AI

📋 Ready-Made Work Descriptions for Any System

DDC CWICR provides complete, structured work descriptions that can be displayed in any system or format. Each work item contains all the information needed by different stakeholders:

Export to Excel, PDF, HTML, ERP systems, BIM platforms — the structured 85-field schema ensures data integrity across all outputs.

🛠️ Works Perfectly With

Claude Code AI coding assistant CLI

Google Antigravity Google Antigravity

n8n Workflow automation

Dify LLM app development

Sim AI & Others AI platforms

⭐ If you want to see new updates and database versions and if you find our tools useful please give our repositories a star to see more similar applications for the construction industry. Star DDC workflow on GitHub and be instantly notified of new releases.

🎯 DDC Skills — 196 Ready-to-Use AI Automations

DDC Skills for AI Agents in Construction — 196 automation skills with direct integration into this CWICR database. Clone, open with AI coding assistant, describe what you need.

💻 Claude Code & Google Antigravity — AI Coding Assistants

The fastest way to work with DDC CWICR. Just open the repository in Claude Code or Google Antigravity and ask questions in natural language.

Getting Started:

# Clone the repository
git clone https://github.com/datadrivenconstruction/OpenConstructionEstimate-DDC-CWICR.git

# Open with Claude Code
cd OpenConstructionEstimate-DDC-CWICR
claude

Example Prompts:

Pro Tips:

• Point Claude to specific files: "Analyze the Parquet file and summarize the cost distribution"
• Ask for explanations: "Explain how the resource-based costing methodology works in this database"
• Request modifications: "Modify the n8n workflow to add email notifications"

⚡ n8n — Visual Workflow Automation

Build powerful automation pipelines without coding. Connect DDC CWICR to 400+ apps and services.

Use Cases:

Quick Start:

1. Download workflow JSON from this repo
2. Import into n8n: Workflows → Import → From File
3. Configure credentials (OpenAI, Qdrant, Telegram)
4. Activate and test

See n8n Workflows section for detailed setup.

📋 Universal Use Cases

No matter which AI tool you choose, DDC CWICR enables:

About

DDC CWICR (Construction Work Items, Components & Resources) is an open database for construction cost estimation, covering the full spectrum of construction activities - from earthworks and concrete placement to specialized installation work.

The database draws on sources describing modern construction practices across Eurasia and the Asia-Pacific region, where a unified technical standardization ecosystem serves as a common engineering language for more than ten dynamically developing economies. DDC CWICR represents an effort to harmonize open standards by establishing a single regulatory framework for capital project management in multiple languages.

The structured data can be accessed through tabular formats (XLSX, CSV, Parquet) or queried conversationally via LLM, enabling specialists to integrate construction work descriptions (QDRANT vector database) into automated pipelines and workflows using plain language or concise queries.

Available Formats

A live demo is available at openconstructionestimate.com, where you can explore the data and see the vector database in action for semantic search.

Data Schema

The database contains 85 fields organized into logical groups. Each record represents either a work item (rate) or a resource with full cost breakdown.

erDiagram
    RATE ||--o{ RESOURCE : contains
    RATE ||--o{ LABOR : requires
    RATE ||--o{ MACHINERY : uses
    RATE ||--o{ PRICE_VARIANT : has

    RATE {
        string rate_code PK "MEKA_KASA_KAKATO_KAME"
        string rate_original_name "Einbau von Trennwänden..."
        string rate_unit "100 m2"
        string category_type "BAUARBEITEN"
        string collection_name "Holzkonstruktionen"
        string department_name "TRENNWÄNDE..."
        string section_name "Einbau von Trennwänden..."
        text work_composition_text
    }

    RESOURCE {
        string resource_code PK "KAME-NE-KAME-KARI"
        string resource_name "Gipskartonplatten"
        string resource_unit "m2"
        float resource_quantity "632.0"
        float resource_price_per_unit_eur "5.02"
        float resource_cost_eur "3170.73"
        boolean is_material
        boolean is_abstract
    }

    LABOR {
        string resource_code FK
        float labor_hours_workers "172"
        float labor_hours_operators "1.67"
        int count_workers_per_unit "172"
        int count_operators_per_unit "2"
        float cost_of_working_hours "3088.11"
    }

    MACHINERY {
        string machine_class2_name "Krane"
        string machine_class3_name "Krane auf Fahrgestellen"
        float electricity_consumption_kwh "0.23"
        float price_operator_wages "13.56"
        float total_value_machinery "64.18"
    }

    PRICE_VARIANT {
        float price_est_median "5.02"
        float price_est_min "3.03"
        float price_est_max "7.99"
        int position_count "24"
        string variable_parts "glasfaserverstärkt..."
    }

Field Groups

The 85 database fields are organized into logical groups that reflect the resource-based cost estimation methodology. Each group serves a specific function in the cost breakdown structure: from hierarchical classification and work item identification to detailed resource consumption, labor requirements, machinery costs, and aggregated totals. This modular structure allows users to query only the relevant fields for their task - whether generating a bill of materials, analyzing labor productivity, or building a complete cost estimate.

Classification - category_type, collection_code, collection_name, department_code, department_name, department_type, section_name, section_type, subsection_code, subsection_name

Work Item (Rate) - rate_code, rate_original_name, rate_final_name, rate_unit, row_type, is_scope, is_abstract, is_machine, is_labor, is_material, work_composition_text

Resources - resource_code, resource_name, resource_unit, resource_quantity, parameter_resource_quantity, resource_price_per_unit_eur_current, resource_cost_eur

Labor - count_workers_per_unit, count_engineers_per_unit, count_operators_per_unit, count_total_people_per_unit, labor_hours_construction_workers, labor_hours_operators, labor_hours_engineers, total_labor_hours_workers_operators, total_labor_hours_all_personnel, cost_of_working_hours, count_people_per_day

Machinery - machine_class2_name, machine_class3_name, personnel_operator_code, personnel_operator_grade, price_operator_wages, price_relocation_included, price_cost_without_wages, electricity_consumption_kwh_per_machine_hour, electricity_cost_per_unit, electricity_cost_total_sum, cost_operator_sum, total_value_machinery_equipment

Price Variants - price_code_prefix, price_abstract_resource_common_start, price_abstract_resource_variable_parts, price_abstract_resource_position_count, price_abstract_resource_est_price_min, price_abstract_resource_est_price_max, price_abstract_resource_est_price_mean, price_abstract_resource_est_price_median, price_abstract_resource_unit, abstract_resource_tech_group

Aggregates - total_cost_per_position, total_material_cost_per_position, total_resource_cost_per_position, total_value_abstract_resources, materials_resource_cost_eur

Mass & Services - mass_name, mass_value, mass_unit, service_category, service_type, parameter_service_code, parameter_service_unit, parameter_service_name, parameter_service_quantity, service_cost_sum

Cost Calculation Formula

Methodology

The key value of Resource-Based Costing is the separation of unchanging production technology from the volatile financial component. It is based on the physical "first principles" of construction:

• Labor hours required for specific work
• Material quantities per unit of work
• Equipment time needed

Why it matters:

• Transparency - Pricing without hidden markups, full resource breakdown
• Auditability - Deep-dive capability for investment analysis and verification
• Portability - Region-independent norms applicable across markets
• Proven - Industry standard methodology established over 100+ years

flowchart TB
    subgraph Source["📦 Data Source"]
        CWICR[(DDC CWICR<br/>────────────<br/>55,719 Work Items<br/>27,672 Resources<br/>85 Fields per Record)]
    end

    subgraph Processing["⚙️ Processing Pipeline"]
        direction LR
        ETL[["🔄 ETL<br/>Extraction &<br/>Transformation"]]
        TRANS[["🌐 Translation<br/>11 Languages"]]
        EMBED[["🧠 Vectorization<br/>OpenAI 3072d"]]
        ETL --> TRANS --> EMBED
    end

    subgraph Outputs["📤 Output Formats"]
        XLSX[("📊 Excel<br/>.xlsx")]
        PARQUET[("⚡ Parquet<br/>.parquet")]
        CSV[("📄 CSV<br/>.csv")]
        QDRANT[("🔍 Qdrant<br/>.snapshot")]
    end

    subgraph Apps["🎯 Applications"]
        SEARCH["🔎 Semantic<br/>Search"]
        BIM["🏗️ BIM 5D<br/>Integration"]
        RAG["🤖 RAG<br/>Systems"]
        BI["📈 BI<br/>Analytics"]
    end

    Source --> Processing
    Processing --> XLSX & PARQUET & CSV & QDRANT
    XLSX & PARQUET & CSV --> BI & BIM
    QDRANT --> SEARCH & RAG & BIM

    style Source fill:#dbeafe,stroke:#2563eb,stroke-width:2px
    style Processing fill:#fef3c7,stroke:#d97706,stroke-width:2px
    style Outputs fill:#d1fae5,stroke:#059669,stroke-width:2px
    style Apps fill:#fce7f3,stroke:#db2777,stroke-width:2px

Historical Context

The construction work descriptions in this database are grounded in a resource-based standardization methodology with roots stretching from early 20th-century production norms to today's digital reference systems. Developed and refined continuously since the 1920s, this approach has seen especially robust evolution across the Eurasian region.

Throughout a hundred years of development, the system has transitioned from manual computations to machine-readable formats - yet its foundational principle remains intact: the precise measurement of physical resources required per unit of construction output. Modern implementations bridge historical normative data with real-time market pricing.

Regional adaptations of this methodology operate under various national designations: ENIR, GESN, FER, NRR, ESN, AzDTN, ShNQK, MKS ChT, SNT, BNbD, Dinh Muc, Ding'e.

Integration

Use Cases

• Entry Level - Cost Benchmarking, Price Indexation, Tender Estimation

• Intermediate - Localization, ETL/BI Pipelines, CO₂ Calculation

• Advanced - AI/ML Training, CAD (BIM) 5D, Deep-Dive Investment Audit

n8n Workflows — Detailed Description

Four production-ready workflows for automated construction cost estimation. Each workflow connects to the DDC CWICR vector database via Qdrant and uses AI models for intelligent parsing and matching.

1️⃣ Text Estimator Bot

File: n8n_1_Telegram_Bot_Cost_Estimates_and_Rate_Finder_TEXT_DDC_CWICR.json

Telegram bot for text-based cost estimation. Describe construction works in natural language — the bot parses input, searches the vector database, and returns detailed cost breakdowns.

🤖 Try It Now — Live Demo Bots

Test the estimation workflows instantly in Telegram

@TextOpenConstructionEstimate_bot

Create complete cost estimates
from text descriptions

flowchart LR
    subgraph Input["💬 INPUT"]
        A[Telegram Message]
    end

    subgraph AI["🤖 AI PROCESSING"]
        B[Parse Text]
        C[Extract Work Items]
    end

    subgraph Search["🔍 VECTOR SEARCH"]
        D[Generate Embeddings]
        E[Qdrant Search]
        F[AI Rerank]
    end

    subgraph Output["📊 OUTPUT"]
        G[Calculate Costs]
        H[HTML Report]
        I[Excel Export]
    end

    A --> B --> C --> D --> E --> F --> G --> H
    G --> I

    style Input fill:#e0f2fe,stroke:#0284c7
    style AI fill:#fef3c7,stroke:#d97706
    style Search fill:#dcfce7,stroke:#16a34a
    style Output fill:#f3e8ff,stroke:#9333ea

How it works:

Features:

Required credentials:

• Telegram Bot Token (from @BotFather)
• OpenAI API Key (for embeddings + optional LLM)
• Qdrant URL + API Key

2️⃣ Photo Cost Estimator

File: n8n_2_Photo_Cost_Estimate_DDC_CWICR.json

Web form interface for photo-based estimation. Upload a construction photo — AI Vision identifies elements, estimates dimensions, and calculates costs automatically.

flowchart TB
    subgraph Upload["📷 PHOTO UPLOAD"]
        A[Web Form]
        B[Select Region]
        C[Choose Work Type]
    end

    subgraph Vision["👁️ AI VISION"]
        D[GPT-4 Vision Analysis]
        E[Identify Elements]
        F[Estimate Dimensions]
        G[Detect Room Type]
    end

    subgraph Decompose["🔧 DECOMPOSITION"]
        H[Elements → Work Items]
        I[Calculate Quantities]
    end

    subgraph Price["💰 PRICING"]
        J[Vector Search]
        K[Match DDC Rates]
        L[Apply Regional Prices]
    end

    subgraph Report["📄 REPORT"]
        M[Generate HTML]
        N[Cost Breakdown]
    end

    A --> B --> C --> D
    D --> E --> F --> G
    G --> H --> I
    I --> J --> K --> L
    L --> M --> N

    style Upload fill:#dbeafe,stroke:#2563eb
    style Vision fill:#fef3c7,stroke:#d97706
    style Decompose fill:#dcfce7,stroke:#16a34a
    style Price fill:#fee2e2,stroke:#dc2626
    style Report fill:#f3e8ff,stroke:#9333ea

How it works:

Features:

Required credentials:

• OpenAI API Key (GPT-4 Vision + embeddings)
• Qdrant URL + API Key

3️⃣ Universal Estimator Bot (Text + Photo + PDF)

File: n8n_3_Telegram_Bot_Cost_Estimates_and_Rate_Finder_TEXT_PHOTO_PDF_DDC_CWICR.json

Full-featured Telegram bot supporting all input types: text descriptions, construction photos, and PDF floor plans. The most comprehensive workflow for production use.

flowchart TB
    subgraph Input["📥 MULTI-INPUT"]
        A[💬 Text Message]
        B[📷 Photo]
        C[📄 PDF Document]
    end

    subgraph Router["🔀 SMART ROUTER"]
        D{Detect Type}
    end

    subgraph TextPath["💬 TEXT PATH"]
        E[AI Parse Text]
        F[Extract Works]
    end

    subgraph PhotoPath["📷 PHOTO PATH"]
        G[Vision AI]
        H[Identify Elements]
        I[Decompose]
    end

    subgraph PDFPath["📄 PDF PATH"]
        J[Extract Pages]
        K[Vision Analysis]
        L[Parse Content]
    end

    subgraph Common["🔍 COMMON PIPELINE"]
        M[Generate Embeddings]
        N[Qdrant Search]
        O[AI Rerank]
        P[Calculate Costs]
    end

    subgraph Export["📤 EXPORT"]
        Q[HTML Report]
        R[Excel CSV]
        S[PDF Document]
    end

    A --> D
    B --> D
    C --> D
    D -->|Text| E --> F --> M
    D -->|Photo| G --> H --> I --> M
    D -->|PDF| J --> K --> L --> M
    M --> N --> O --> P
    P --> Q
    P --> R
    P --> S

    style Input fill:#e0f2fe,stroke:#0284c7
    style Router fill:#fef3c7,stroke:#d97706
    style TextPath fill:#dcfce7,stroke:#16a34a
    style PhotoPath fill:#fce7f3,stroke:#db2777
    style PDFPath fill:#f3e8ff,stroke:#9333ea
    style Common fill:#fee2e2,stroke:#dc2626
    style Export fill:#d1fae5,stroke:#059669

How it works:

17 Bot Actions:

Features:

Required credentials:

• Telegram Bot Token
• OpenAI API Key (embeddings)
• Gemini API Key (Vision) or OpenAI GPT-4 Vision
• Qdrant URL + API Key

4️⃣ CAD (BIM) Cost Estimation Pipeline

File: n8n_4_CAD_(BIM)_Cost_Estimation_Pipeline_4D_5D_with_DDC_CWICR.json

Automated cost estimation from Revit/IFC/DWG models. Extracts BIM data, classifies elements, decomposes into work items, and generates 4D/5D estimates with full resource breakdown.

flowchart TB
    subgraph INPUT["📁 INPUT<br/><i>CAD • photos • text description</i>"]
        CAD["📐 Project Input<br/>(text • photos • RVT / IFC / DWG)"]
    end

    subgraph EXTRACT["⚙️ EXTRACTION"]
        CONV["RvtExporter.exe / CAD Export  / ETL"]
        XLSX["📊 .XLSX<br/>(Raw Elements)"]
    end

    subgraph PREP["🔧 DATA PREPARATION"]
        PREP_AI["🤖 AI: Clean & Classify<br/><i>headers • types • categories</i>"]
    end

    subgraph STAGE_PLAN["📋 STAGES 1–3: Planning"]
        PLAN["🤖 Detect Project & Phases<br/><i>new / renovation / demolition</i><br/><i>small / medium / large</i><br/><i>elements → construction phases</i>"]
    end

    subgraph STAGE4["🔨 STAGE 4: Decomposition"]
        S4["🤖 Decompose Types to Works<br/><i>'Brick Wall 240mm' → masonry, mortar, plaster</i>"]
    end

    subgraph STAGE5["💰 STAGE 5: Pricing"]
        S5["🤖 Price via Vector DB<br/><i>OpenAI embeddings + Qdrant</i><br/><i>rate_code, unit_cost, resources</i>"]
    end

    subgraph STAGE75["✅ STAGE 7.5: Validation"]
        S75["🤖 CTO Review<br/><i>completeness • duplicates • missing works</i>"]
    end

    subgraph OUTPUT["📤 OUTPUT"]
        HTML["📄 HTML Report"]
        XLS["📊 XLS Report"]
    end

    CAD --> CONV --> XLSX
    XLSX --> PREP_AI --> PLAN --> S4 --> S5 --> S75
    S75 --> HTML & XLS

    style INPUT fill:#f4f4f5,stroke:#d4d4d8,color:#18181b
    style EXTRACT fill:#e0f2fe,stroke:#bae6fd,color:#0f172a
    style PREP fill:#ede9fe,stroke:#ddd6fe,color:#1e1b4b
    style STAGE_PLAN fill:#ecfdf5,stroke:#bbf7d0,color:#064e3b
    style STAGE4 fill:#fef9c3,stroke:#fef3c7,color:#78350f
    style STAGE5 fill:#fee2e2,stroke:#fecaca,color:#7f1d1d
    style STAGE75 fill:#e0f2f1,stroke:#bae5e1,color:#134e4a
    style OUTPUT fill:#eef2ff,stroke:#e0e7ff,color:#111827

n8n provides 400+ native integrations with platforms like Google Sheets, Notion, Slack, Airtable, databases (PostgreSQL, MongoDB), cloud storage, and more. Every node in this workflow is modular — you can:

• 🔄 Swap LLM providers (OpenAI ↔ Claude ↔ Gemini ↔ Grok)
• 📊 Connect to your ERP or project management system
• 📁 Export results to any destination (cloud storage, email, dashboards)
• 🔧 Modify any stage to match your estimation methodology

The workflow is yours to adapt. No restrictions. No licensing fees. Full control.

📋 Prerequisites

Workflows Quick Start

Step 1: Import Workflow

n8n → New workflow → Import from File → Select JSON

Step 2: Configure Credentials

In the 🔑 TOKEN node, set your API keys:

{
  "bot_token": "YOUR_TELEGRAM_BOT_TOKEN",
  "OPENAI_API_KEY": "YOUR_OPENAI_KEY",
  "GEMINI_API_KEY": "YOUR_GEMINI_KEY",
  "QDRANT_URL": "http://localhost:6333",
  "QDRANT_API_KEY": ""
}

Step 3: Load DDC CWICR to Qdrant

Download the snapshot from the corresponding language folder in this repository and import:

curl -X POST "http://localhost:6333/collections/ddc_en_toronto/snapshots/upload" \
  -H "Content-Type: multipart/form-data" \
  -F "snapshot=@EN___DDC_CWICR/EN_TORONTO_workitems_costs_resources_EMBEDDINGS_3072_DDC_CWICR.snapshot"

Step 4: Activate & Test

• Enable the workflow in n8n
• For Telegram bots: send /start to your bot
• For web forms: open the form URL provided by n8n

⚠️ n8n 2.0+ Setup Required

Starting from n8n version 2.0, the Execute Command node is disabled by default for security reasons.

Without the configuration below, workflows using Execute Command (especially CAD/BIM Pipeline) will not work — nodes will show with a question mark or won't be recognized.

Quick Fix

Windows (CMD) — run each time:

set NODES_EXCLUDE=[] && npx n8n

Permanent solution — create once:

Create file C:\Users\YOUR_USER\.n8n\.env with:

NODES_EXCLUDE=[]

Then just run npx n8n as usual.

Docker:

environment:
  - NODES_EXCLUDE=[]

Verify Setup

1. Start n8n
2. Click + → search for "Execute Command"
3. If the node appears → ✅ you're all set!

📚 More details: n8n 2.0 Breaking Changes

🌍 Supported Languages & Price Levels

📊 Pipeline Stages

The CAD/BIM workflow processes data through 10 stages:

⚙️ LLM Model Selection

The workflow supports multiple AI providers. Enable your preferred model in the LLM Models section:

To switch models: Enable the desired model node and Disable others.

📁 Output Files

Reports are saved to the project folder:

project_YYYY-MM-DD.html   ← Interactive report (opens in browser)
project_YYYY-MM-DD.xls    ← Excel-compatible spreadsheet

🔗 Qdrant Collections

The workflow automatically selects the correct collection based on language_code:

{LANG}_{CITY}_workitems_costs_resources_EMBEDDINGS_3072_DDC_CWICR

Example: DE_BERLIN_workitems_costs_resources_EMBEDDINGS_3072_DDC_CWICR

⚠️ Troubleshooting

Vector Database

Ready-to-use Qdrant collections with OpenAI text-embedding-3-large embeddings for semantic search across construction work items.

Vector databases allow you to "talk" to your data in natural language – using simple sentences or short phrases instead of code or complex filters. This dramatically speeds up finding the right work item or cost line, even in very large datasets.

These Qdrant collections can be connected to application via modern automation and integration workflows (for example, low-code/no-code Workflow and Pipeline tools). You can build assistants that search, filter and explain construction work items, or integrate semantic search directly into your existing estimation and project-control tools.

Qdrant Vector Database Snapshots

All Qdrant snapshots are included directly in the corresponding language folders of this repository (stored via Git LFS).

Derived tracks (19) — built by 0_Workflow and Pipelines CWICR/python/11-country-track-builder/

Norms (labour hours, machine hours, resource quantities) are bytewise-identical to the source track; only prices and translatable text differ. rate_code and resource_code are stable across all tracks.

All collections use 3072-dimensional OpenAI embeddings with full payload metadata.

Collections

Shipped (11): 🇸🇦 ddc_ar_dubai · 🇩🇪 ddc_de_berlin · 🇨🇦 ddc_en_toronto · 🇪🇸 ddc_sp_barcelona · 🇫🇷 ddc_fr_paris · 🇮🇳 ddc_hi_mumbai · 🇧🇷 ddc_pt_saopaulo · 🇷🇺 ddc_ru_stpetersburg · 🇬🇧 ddc_uk_gbp · 🇺🇸 ddc_usa_usd · 🇨🇳 ddc_zh_shanghai

Derived (19): 🇦🇺 ddc_au_sydney · 🇧🇬 ddc_bg_sofia · 🇨🇿 ddc_cs_prague · 🇭🇷 ddc_hr_zagreb · 🇮🇩 ddc_id_jakarta · 🇮🇹 ddc_it_rome · 🇯🇵 ddc_ja_tokyo · 🇰🇷 ddc_ko_seoul · 🇲🇽 ddc_mx_mexicocity · 🇳🇬 ddc_ng_lagos · 🇳🇱 ddc_nl_amsterdam · 🇳🇿 ddc_nz_auckland · 🇵🇱 ddc_pl_warsaw · 🇷🇴 ddc_ro_bucharest · 🇸🇪 ddc_sv_stockholm · 🇹🇭 ddc_th_bangkok · 🇹🇷 ddc_tr_istanbul · 🇻🇳 ddc_vi_hanoi · 🇿🇦 ddc_za_johannesburg

Docker Deployment

# docker-compose.yml
services:
  qdrant:
    image: qdrant/qdrant:latest
    container_name: ddc-cwicr-qdrant
    ports:
      - "6333:6333"
      - "6334:6334"
    volumes:
      - qdrant_storage:/qdrant/storage
      - ./snapshots:/qdrant/snapshots
    environment:
      - QDRANT__LOG_LEVEL=INFO
    restart: unless-stopped

volumes:
  qdrant_storage:

# Start
docker-compose up -d

# Import snapshot
curl -X POST "http://localhost:6333/collections/ddc_en_toronto/snapshots/upload" \
  -H "Content-Type: multipart/form-data" \
  -F "snapshot=@EN___DDC_CWICR/EN_TORONTO_workitems_costs_resources_EMBEDDINGS_3072_DDC_CWICR.snapshot"

# Dashboard: http://localhost:6333/dashboard

Linux APT Packages

Install Qdrant + construction cost data with a single command on any Debian/Ubuntu system. No Docker, no manual setup — just apt install and search.

sudo apt install ddc-cwicr-en
  │
  ├── ddc-qdrant              Qdrant v1.16.3 binary + systemd service
  │     └── localhost:6333    vector database ready on port 6333
  │
  └── postinst                downloads ~1 GB snapshot from GitHub Releases
        └── PUT /snapshots/recover → 55,719 vectors loaded

Setup APT Repository

# Add the DDC package repository
echo "deb [trusted=yes] https://pkg.datadrivenconstruction.io stable main" \
  | sudo tee /etc/apt/sources.list.d/ddc.list

sudo apt update

Install a Language Collection

# Install English construction cost database (downloads ~1 GB of vector data)
sudo apt install ddc-cwicr-en

# Verify Qdrant is running
systemctl status qdrant
curl http://localhost:6333/collections

The .deb package is only ~5 KB — the heavy vector data is downloaded directly from GitHub Releases during installation and restored into Qdrant automatically.

Available Packages

Install multiple languages side by side:

sudo apt install ddc-cwicr-en ddc-cwicr-de ddc-cwicr-fr

CLI Search Tool

The optional ddc-cwicr-cli package provides ddc-search — a command-line interface for querying the database directly from the terminal.

sudo apt install ddc-cwicr-cli

Semantic search (requires OpenAI API key):

export OPENAI_API_KEY=sk-...

ddc-search "reinforced concrete foundation 300mm"

╔══════════════════════════════════════════════════════════════════╗
║  DDC CWICR Search Results — ddc_en_toronto (55,719 items)      ║
╚══════════════════════════════════════════════════════════════════╝

  #1  [0.847]  03.01.004
      Reinforced concrete strip foundations, width 300-600mm
      Unit: m³    Labor: $45.20    Material: $189.50    Total: $287.30

  #2  [0.831]  03.01.007
      Reinforced concrete pad foundations up to 500mm depth
      Unit: m³    Labor: $52.10    Material: $195.80    Total: $312.60

  #3  [0.814]  03.02.001
      Concrete foundation walls, reinforced, 200-400mm thick
      Unit: m³    Labor: $48.90    Material: $178.40    Total: $279.50

Keyword search (no API key needed):

ddc-search --keyword "concrete"

Other options:

# Search a specific language collection
ddc-search --collection ddc_de_berlin "Stahlbetonfundament"

# List all installed collections
ddc-search --list

# JSON output for scripting and automation
ddc-search --json "floor tiles installation"

# Limit number of results
ddc-search --limit 10 "steel beam HEB 300"

Package Architecture

• ddc-qdrant is installed automatically as a dependency of any ddc-cwicr-{lang} package
• Removing a language package (apt remove ddc-cwicr-en) deletes the collection from Qdrant
• Purging ddc-qdrant (apt purge ddc-qdrant) removes all data and the system user
• Available for amd64 and arm64 architectures

🌐 Pricing Search API — BuildCalculator.io

Free REST API for searching construction work items with full cost breakdown, labor, materials, and equipment data. 55,719 items across 11 languages with 84 fields per item.

Base URL: https://buildcalculator.io/api/v1

API Endpoints

GET/POST /api/v1/search — Search Construction Items

GET /api/v1/languages — List Supported Languages

Returns all available languages with item counts.

GET /api/v1/stats — Database Statistics

Returns item counts, categories, languages, and metadata.

API Code Examples

cURL:

curl "https://buildcalculator.io/api/v1/search?q=concrete+foundation&lang=en&top=5"

Python:

import requests

response = requests.get("https://buildcalculator.io/api/v1/search",
    params={"q": "brick masonry walls", "lang": "en", "top": 5})
data = response.json()

for item in data["results"]:
    print(f"{item['name']} — {item['pricing']['total_per_unit']} EUR/{item['unit']}")

JavaScript:

const res = await fetch(
  "https://buildcalculator.io/api/v1/search?q=HVAC+ducting&lang=en&top=3"
);
const data = await res.json();

Response example:

{
  "query": "concrete foundation",
  "language": "en",
  "results_count": 5,
  "results": [
    {
      "rate_code": "KANE_KAME_KAKAME_KAMECON",
      "name": "Concrete preparation device",
      "unit": "m3",
      "currency": "EUR",
      "pricing": {
        "total_per_unit": 167.51,
        "labor_per_unit": 18.80,
        "material_per_unit": 142.92,
        "equipment_per_unit": 4.80
      },
      "cost_breakdown": {
        "labor_pct": 11.3,
        "material_pct": 85.8,
        "equipment_pct": 2.9
      }
    }
  ]
}

Error codes:

📖 Full documentation: buildcalculator.io/api-docs

Quick Start

Python - Tabular Data

import pandas as pd

# Parquet (recommended)
df = pd.read_parquet("DDC_CWICR_EN.parquet")

# Excel
df = pd.read_excel("DDC_CWICR_EN.xlsx")

print(f"Records: {len(df):,} | Fields: {len(df.columns)}")
print(df[['rate_code', 'rate_original_name', 'rate_unit', 'total_cost_per_position']].head())

Python - Semantic Search

from qdrant_client import QdrantClient
from openai import OpenAI

client = QdrantClient("localhost", port=6333)
openai = OpenAI()

# Search by natural language
query = "reinforced concrete foundation pouring"
embedding = openai.embeddings.create(
    input=query, 
    model="text-embedding-3-large"
).data[0].embedding

results = client.search(
    collection_name="ddc_en_toronto",
    query_vector=embedding, 
    limit=5
)

for r in results:
    print(f"[{r.score:.3f}] {r.payload['rate_code']}: {r.payload['rate_original_name']}")

Filtered Search

from qdrant_client.models import Filter, FieldCondition, MatchValue, Range

# By department
results = client.search(
    collection_name="ddc_en_toronto",
    query_vector=embedding,
    query_filter=Filter(must=[
        FieldCondition(key="department_name", match=MatchValue(value="Concrete and Reinforced Concrete"))
    ]),
    limit=10
)

# By price range
results = client.search(
    collection_name="ddc_en_toronto",
    query_vector=embedding,
    query_filter=Filter(must=[
        FieldCondition(key="price_est_median", range=Range(gte=1000, lte=50000))
    ]),
    limit=10
)

💻 Developer Examples

The examples/ directory contains ready-to-run code in multiple languages — from basic data loading to advanced RAG pipelines and cost estimation.

Three Ways to Start

Available Examples

Includes sample data (100-row Parquet extract), Docker Compose for Qdrant, and .env template.

→ Browse all examples

Resources & Community

Consulting & Training

We work with leading construction, engineering, consulting agencies, and technology firms around the world to help them implement open data principles, automate CAD/BIM processing, and build robust ETL pipelines. We actively support organizations seeking practical solutions for digital transformation and interoperability, focusing on data quality and classification challenges while driving the adoption of open and automated workflows.

If you would like to test this solution with your own data or are interested in adapting the workflow to real project tasks, feel free to contact us. Our team delivers hands-on workshops, provides strategic consulting, and develops prototypes tailored to real project processes.

Contributing

DDC CWICR is a free and open project dedicated to making the construction industry more efficient, transparent, and technologically advanced. We are actively looking for like-minded enthusiasts who share this mission. If you create useful solutions and are ready to share them with the community, we are here to help you be heard.

We invite you to submit your open source workflows, pipelines, and integrations based on DDC CWICR-tools that anyone can freely use in their work. The top solutions will be published with full author attribution on GitHub and announced through our newsletter and social media channels, reaching tens of thousands of professional subscribers. This places your name directly in front of an international community of estimators, BIM specialists, and project managers.

Together we are changing the industry. You can send your solution to [email protected] with the subject "DDC Open Workflow" or submit a Pull Request directly to our GitHub repositories.

Automate construction data processing with ready-made CAD-BIM n8n workflows:

🤖 AI Instructions

The AI_INSTRUCTIONS/ folder contains comprehensive documentation for AI coding assistants to work effectively with this construction cost database.

What is DDC CWICR?

DDC CWICR (Construction Work Items, Components & Resources) is an open-source construction cost database containing:

• 55,719 work items — detailed construction operations with full cost breakdowns
• 27,672 resources — materials, labor, and equipment with regional pricing
• 85 data fields — structured schema for accurate cost calculations
• 11 languages — with region-specific pricing (EUR, USD, CAD, RUB, CNY, etc.)
• Pre-computed embeddings — 3072-dimensional OpenAI vectors for semantic search

Resource-Based Methodology

The database uses a resource-based costing approach that separates:

• Technology norms (unchanging) — labor hours, material quantities, equipment time
• Regional prices (volatile) — hourly rates, material costs, fuel prices

Actual Cost = Technology Norm × Regional Price

This allows accurate estimation across different regions and time periods.

AI Instructions Files

n8n Workflows — Examples & Templates

The included n8n workflows are examples and templates demonstrating cost estimation logic. They can be:

• ✅ Used as-is for quick deployment
• ✅ Partially adapted for specific business requirements
• ✅ Studied to understand the cost calculation methodology
• ✅ Referenced when building custom integrations on any platform

The workflows demonstrate: database queries, work item matching, regional pricing logic, and report generation. AI can analyze these to understand the complete estimation process.

Why This Matters

AI assistants can help you:

• Query the database using natural language
• Find work items by semantic search
• Calculate costs with regional pricing
• Generate reports and export data
• Build integrations with cloud services
• Understand cost calculation methodology from workflow examples

Quick Start with AI

1. Open the project in your AI-enabled IDE
2. Ask: "Show me all concrete work items with their costs"
3. The AI will use the instructions to query the data correctly

Book: For methodology details, see Data-Driven Construction Book

License

This repository is dual-licensed. See LICENSE for the full manifest, NOTICE for third-party attributions and the EU sui generis database-right notice, and DATA_DICTIONARY.md for the data schema.

What you CAN do with the DATA (CC BY 4.0)

What you MUST do (CC BY 4.0)

Attribution — required format

Every project, product, publication, or application that uses DDC CWICR data must include:

DDC CWICR by Artem Boiko / DataDrivenConstruction
https://github.com/datadrivenconstruction/OpenConstructionEstimate-DDC-CWICR
Licensed under CC BY 4.0

A machine-readable citation is provided in CITATION.cff; GitHub surfaces it as the "Cite this repository" button.

Examples of how to attribute:

Commercial licensing

CC BY 4.0 already permits commercial use subject to the attribution condition. A separate commercial licence is available if you need any of: an attribution waiver, indemnification for the data as delivered, a data-processing agreement (DPA), a service-level agreement, or regulated- industry compliance documentation.

Contact: [email protected].

Security

Do not open public GitHub issues for security vulnerabilities. Report privately via GitHub Security Advisories or by email to [email protected]. See SECURITY.md for the full coordinated-disclosure policy, response timelines, scope, supply-chain-integrity posture (Git LFS hashes), and alignment with EU Regulation 2024/2847 (Cyber Resilience Act) reporting obligations.

Support the Project

If you find this useful, please consider supporting:

Unlock the Power of Data in Construction
_{Move to full-cycle data management where only unified structured data & processes remain}

_{© 2025 Artem Boiko · datadrivenconstruction.io}

Trademarks

Autodesk®, Revit®, AutoCAD®, and DWG™ are registered trademarks or trademarks of Autodesk, Inc. OpenAI™ is a trademark of OpenAI, Inc. Qdrant is a trademark of Qdrant Solutions GmbH. All other brand names, product names, or trademarks belong to their respective holders.

This project is not affiliated with, endorsed by, or sponsored by Autodesk, OpenAI, Qdrant, or any other trademark holders mentioned above.