---
license: cc-by-4.0
language:
  - en
task_categories:
  - text-generation
tags:
  - data-engineering
  - sql
  - apache-spark
  - apache-kafka
  - apache-airflow
  - dbt
size_categories:
  - n<1K
---

# Data Engineering Mentor Q&A Dataset

An original, source-linked starter dataset of practical data engineering questions and answers. It covers architecture, SQL modeling, Spark, Kafka, orchestration, quality, lakehouse storage, and production operations.

## Load with Python

```bash
pip install datasets
```

```python
from datasets import load_dataset

data = load_dataset("LoNebula/data-engineering-mentor-dataset")
print(data)
print(data["train"][0]["question"])
```

The repository has 50 examples split into 42 training and 8 test records. Each row includes an ID, topic, question, answer, official-documentation references, license, and chat messages for instruction tuning.

## Provenance and license

All Q&A explanations were originally composed for this project; no third-party Q&A corpus or documentation passages were copied. Each answer links to relevant primary documentation for readers to check product-specific details. The dataset is released under **CC BY 4.0**. You may reuse and adapt it with attribution to `LoNebula` and a link to this dataset.

References include the official documentation for Apache Airflow, Apache Kafka, Apache Spark, dbt, Apache Iceberg, PostgreSQL, and OpenLineage. Links are references, not claims that those projects reviewed or endorsed this dataset.

## Scope and limitations

This is a small learning corpus focused on durable concepts and safe operational habits. It is not a complete data engineering curriculum, a benchmark, or a substitute for current product manuals and organization-specific standards. Some answers may omit context; verify version-sensitive behavior in the linked official documentation.

## Citation

`LoNebula/data-engineering-mentor-dataset`, version 1.0, CC BY 4.0. Original Q&A explanations and dataset packaging by the dataset maintainers; official documentation links are listed per record.