Abstract:This study investigates role of social media user engagement metrics in predicting career success likelihoods using supervised machine learning techniques. With platforms like LinkedIn and VKontakte becoming pivotal for networking and advancement, user statistics have emerged as potential indicators of professional capability. However, research questions metric reliability considering impression management tactics and biases. While prior studies examined limited activity features, this analysis adopts a robust CatBoost model to gauge career success prediction from multifaceted social data combinations. The study utilizes user profiles of over 17,000 on a major Russian platform. Individuals are categorized by an algorithm accounting for factors like salaries, experience, and employment status. User statistics spanning engagement, content sharing, popularity, and profile completeness provide model inputs. Following comparative evaluation, CatBoost achieved superior performance in classification accuracy, precision, recall and ROC AUC score. Analysis of SHapley Additive exPlanations values provides explanatory modeling insights into influential metrics, thresholds, and patterns. Results reveal subscribers, reposts and interest pages as highly impactful, suggesting that influence and content resonance predict success better than sheer visibility indicators like multimedia volumes. Findings also point to optimal engagement ranges beyond which career prediction gains diminish. Additionally, profile completeness and regular posting are positive to a limit, while likes to have negligible effects. The study contributes more holistic, data-driven visibility into effective social media conduct for career advancement. It advocates prioritizing network cultivation, tactical self-presentation, shareable narratives and reciprocal relationships over metrics gaming. Findings largely validate strategic communication theory around impression management and relationship-building.

How Recommendation Letters Affect Career Mobility：Evidence from a Social Networking Sites LinkedIn

The implicit network inferred from users’ residences and workplaces enhancing collaborative recommendation on smartphones

Social Recommender Systems Based on Coupling Network Structure Analysis

Talent Recommendation on LinkedIn User Profiles

Influence Of Reciprocal Links In Social Networks

Understanding and predicting future research impact at different career stages—A social network perspective

Social Recommendation With Learning Personal and Social Latent Factors

Where to Go: an Effective Point-of-interest Recommendation Framework for Heterogeneous Social Networks

A Hierarchical Career-Path-Aware Neural Network for Job Mobility Prediction

A Survey of Link Recommendation for Social Networks: Methods, Theoretical Foundations, and Future Research Directions

Exploiting User'S Social Network: A Novel Method to Recommend Most Attractive and Targeted Service

A Personalized Researcher Recommendation Approach in Academic Contexts: Combining Social Networks and Semantic Concepts Analysis.

What Is My Next Job: Predicting The Company Size And Position In Career Changes

Link Prediction and Recommendation Across Heterogeneous Social Networks

The role of strategic online engagement and content curation in professional branding and career advancement on social media platforms

User Recommendations in Reciprocal and Bipartite Social Networks--An Online Dating Case Study

Mechanism of Social Capital on Chinese Employees' Career Success

User recommendation in reciprocal and bipartite social networks -- a case study of online dating

Community-Based Scholar Recommendation Modeling in Academic Social Network Sites

Social Networks, Job Satisfaction and Job Searching Behavior in the Chinese Labor Market

DLRRS: A New Recommendation System Based on Double Linear Regression Models