Predicting the Probability and Salary to Get Data Science Job in Top Companies

摘要

Purpose: Predict the probability to get data science job in Fortune 500 companies through predictive analysis. Methodology/approach: Following an introduction of career-based social websites and Human Resource Analytics, the authors processed the common features from LinkedIn and Glassdoor which are necessary to connect two different data sources with features as company name and job title, applied the methodology of Data Mining - Cross Industry Standard Process. The major machine learning algorithms include gradient boosting decision tree and logistic regression. Findings: Predict the probability to get the job in different categories of companies with expected salary mean. Originality/value: Instead of traditional employment survey this research base on web analytics, data mining and predictive modeling which enabled low cost, high efficiency, short lead-time analytics. The methodology could be widely used to discover all kinds of career based insights for various research purposes.