Swiss German Scripted Monologue Speech Dataset for Delivery & Logistics Domain

The audio dataset comprises scripted monologue speech data in the Delivery & Logistics domain, featuring native German speakers from Switzerland. It includes speech data, detailed metadata, and accurate transcriptions.

About this Off-the-shelf Speech Dataset

Introduction

The Swiss German Scripted Monologue Speech Dataset for the Delivery & Logistics Domain is a meticulously curated resource developed to support German language speech recognition technologies, with a focus on real-world delivery and logistics applications.

Speech Data

This dataset includes 6,000+ high-quality scripted monologue recordings in Swiss German, crafted to simulate practical scenarios in the delivery and logistics industry. These prompts are ideal for building robust, domain-specific conversational AI and customer support systems.

•Participant Diversity

•

Speakers: 60 native German speakers

•

Regional Representation: Covers diverse dialects and accents from multiple regions of Switzerland

•

Demographics: Participants aged 18–70, with a 60:40 male-to-female ratio

•Recording Specifications

•

Nature of Recordings: Scripted prompts and monologues

•

Average Duration: 5–30 seconds per clip

•

Format: WAV files, mono channel, 16-bit depth, 8 kHz and 16 kHz sample rates

•

Environment: Noise-free, echo-free, quiet recording settings

Topic & Scenario Coverage

The dataset captures a wide variety of realistic delivery and logistics situations, including:

•Customer service dialogues

•Order processing and status inquiries

•Shipping, delivery, and tracking updates

•Returns, refunds, and complaint handling

•Technical assistance for delivery issues

•Regulatory questions and operational policies

•General advisory and domain-specific statements

Linguistic Features

To simulate authentic conversations, prompts include:

•

Names: Regional male and female names in natural formats

•

Addresses: Diverse location references including street names and regions

•

Dates & Times: Common references for delivery slots, pickups, and ETA

•

Order Numbers: Tracking IDs, invoice numbers, and order references

•

Quantities & Weights: Units related to shipments and packaging

•

Logistics Providers: Mentions of real or fictional courier and logistics services

Transcription

Each audio file is paired with a verbatim transcription, enhancing usability for training and validation:

•

Content: Exact match of the audio prompt

•

Format: Plain text (.TXT) with filenames aligned to audio files

•

Quality Assurance: All transcripts are reviewed by native German linguists for precision and consistency

Metadata

Comprehensive metadata accompanies every audio file and participant profile, supporting flexible filtering and model adaptation:

•

Participant Metadata: Unique speaker ID, age, gender, region, and dialect

•

Audio Metadata: Prompt transcript, recording environment, device used, sample rate, bit depth, and file format

Use Cases & Applications

This dataset is a versatile asset for developing various AI-powered voice and language solutions in the Delivery & Logistics sector:

•

Speech Recognition Training: Build ASR systems tailored for logistics workflows

•

Voice Synthesis: Train TTS models for courier updates and customer support bots

•

Voice Assistants: Enable natural-sounding, logistics-specific voice assistants

•

Chatbots: Build dialogue models to automate tracking, complaint handling, and more

•

NER Models: Extract structured data like names, order IDs, dates, and prices

•

Language Understanding: Develop models for sentiment analysis, topic detection, and intent classification in delivery-related contexts

Ethical & Secure Data Collection

All data was collected using Yugo, FutureBeeAI’s proprietary and secure data collection and transcription platform

•Data privacy and participant consent were ensured through strict ethical practices

•No personally identifiable information (PII) is present, making the dataset secure for commercial and research use

License

This dataset is published by FutureBeeAI and is available for commercial licensing. It serves as a high-quality resource for enterprises and researchers building German speech solutions for the Delivery & Logistics sector.