doublecringe123/parsed-vacancies-from-headhunter-tashkent

Dataset Card: Parsed Vacancies from HeadHunter Tashkent

Dataset Summary

This dataset contains parsed job vacancies collected from the HeadHunter Tashkent website. The dataset includes fields such as job title, required experience, employment type, company name (anonymized), salary, job description, and required skills. The data was collected using Selenium WebDriver and BeautifulSoup for scraping the website. To protect the privacy of the companies, all company names have been anonymized and replaced with the generic label 'company' using the NER model Davlan/distilbert-base-multilingual-cased-ner-hrl.

The dataset has been filtered by language, containing only entries in Russian and English. The filtering was done using the langdetect library to ensure accurate language classification.

Dataset Structure

Features:

• title: Job title (string)
• experience: Required experience (string)
• employment: Employment type (string, e.g., full-time, part-time)
• company_name: Anonymized as 'company' (string)
• salary: Offered salary (string, may contain ranges or be absent)
• description: Job description (string, anonymized to remove specific company details)
• skills: Required skills (string, comma-separated)
• mode: Mode of employment (e.g., remote, on-site) (string)

Example Row:

{
    "title": "Software Engineer",
    "experience": "3-5 years",
    "employment": "Full-time",
    "company_name": "company",
    "salary": "1000-1500 USD",
    "description": "We are looking for an experienced software engineer to join our dynamic team. Proficiency in Python and Django required.",
    "skills": "Python, Django, REST APIs, Git",
    "mode": "Remote"
}

Data Collection Process

The data was scraped using Selenium WebDriver and parsed with BeautifulSoup. To ensure privacy and data protection, all company names were identified using the Davlan/distilbert-base-multilingual-cased-ner-hrl NER model and replaced with the generic term 'company' to prevent the exposure of private or sensitive information.

Only job listings in Russian and English were retained, as detected by the langdetect library.

Intended Uses

This dataset can be used for research and development of natural language processing (NLP) models focused on job-related data, such as:

• Job title extraction
• Skills extraction and matching
• Salary prediction
• Multilingual job description analysis (Russian, English)

Dataset Limitations

• Language: Only Russian and English job listings are included.
• Anonymization: Company names have been anonymized, which may affect models that require company-specific information for accurate prediction.
• Incomplete Salary Data: Some job postings may not include salary information, which may affect salary-based models.

Licensing

Please ensure that you comply with the terms and conditions of the HeadHunter Tashkent website regarding data usage.

Contact Information

• GitHub: goin2crazy
• Hugging Face: doublecringe123
• Email: yanyap666@gmail.com

Citation

If you use this dataset, please cite it as:

@dataset{doublecringe123_2024_headhunter_tashkent,
  author    = {goin2crazy},
  title     = {Parsed Vacancies from HeadHunter Tashkent},
  year      = {2024},
  url       = {https://huggingface.co/datasets/doublecringe123/parsed-vacancies-from-headhunter-tashkent},
  email     = {yanyap666@gmail.com}
}