Turkish-EV-Charging-Intent-Dataset

A compact Turkish-language text-classification dataset for routing electric-vehicle and charging questions. It contains 192 synthetic-editorial queries balanced across eight intent classes, with fixed train, validation, and test splits. It is not search-log or customer data and contains no personal data. The default prediction target is the intent column. Official resources: - [Permanent Zenodo record - DOI 10.5281/zenodo.22062688](https://doi.org/10.5281/zenodo.22062688) - [Canonical GitHub repository](https://github.com/gokimedia/turkish-ev-charging-intents) - [TekPedal open-data documentation](https://tekpedal.com/acik-veri) - [Kaggle dataset mirror](https://www.kaggle.com/datasets/morrispoint/turkish-ev-charging-intent-dataset) - [Hugging Face dataset mirror](https://huggingface.co/datasets/FoodSecuriry/turkish-ev-charging-intents) See the canonical repository and DOI record for full provenance, limitations, validation details, and responsible-use guidance.