Exploring Job Vacancy Topics and Trends in Indonesia Using Latent Dirichlet Allocation (LDA) and Exploratory Data Analysis (EDA)

Abstract

The Indonesian labor market is rapidly evolving, creating a need for effective analysis to uncover patterns and trends in job vacancies, especially in the digital era where information is abundant and diverse. This study aims to identify dominant topics and vacancy trends to provide insights for job seekers, companies, and stakeholders in developing targeted recruitment strategies. Data were collected through web scraping from the Jobstreet website, covering company, location, position, classification, subclassification, salary, and posting time. Two methods were applied: Exploratory Data Analysis (EDA) to describe vacancy distributions, and Latent Dirichlet Allocation (LDA) to extract thematic structures from text data. EDA results show that Sales Executive is the most demanded position, Greater Jakarta is the location with the highest vacancies, and the most common salary range is IDR 4–5 million. LDA, validated using the Coherence Score, identified two distinct topics: (1) Sales, Marketing, Supply Chain, and Retail, and (2) Customer Service and Administration.. This research contributes by combining descriptive and text-mining techniques on the latest job vacancy dataset obtained directly through web scraping, producing an analysis that not only highlights vacancy distribution but also reveals latent topics reflecting Indonesian labor market trends.