Human Capital Analytics Laboratory

Understanding the past, predicting the future


Mission

The mission of the New York University Human Capital Analytics Laboratory is to advance thinking and practice of human capital management for the betterment of of workers and organizations. The members of HCA Labs seek to advance artificial intelligence, machine learning, and natural language processing through research and innovation with human capital data. HCA Labs members explore human capital data to understand the past and predict the future.

Who We Are:

Leadership Team

Harold Kaufman, PhD – Management of Technology, NYU Tandon School of Engineering

Julian Togelius, PhD – Computer Science, NYU Tandon School of Engineering

Paul Squires, PhD – Psychology Department, NYU Graduate School of Arts & Science

Catalina Jaramillo, PhD Candidate – Computer Science, Tandon School of Engineering

Student Researchers

Abhi Dasari (Sri) – CS, NYU Tandon School of Engineering

Ansh Desai – CS, NYU Tandon School of Engineering

Anyada Assavabhokhin  – IO Psychology, NYU Graduate School of Arts & Science

Eli Jaffe – IO Psychology, NYU Graduate School of Arts & Science

Kate Merritt Paulson – IO Psychology, NYU Graduate School of Arts & Science

Kevin Patel – CS, NYU Tandon School of Engineering

Wenjie Xie -IO Psychology, NYU Graduate School of Arts & Science

Sarah Chen 

HCA Labs Projects

Using ONET work activities to predict job classification.

ONET is a product of the US Department of Labor which provides extensive information about jobs in the US labor market. ONET jobs are grouped into nearly 1,000 job titles each of which is described by a set of work activities. Using only the work activities, the research seeks to use NLP and ML to create predictive models that predict job titles. If this is feasible, then the team will use job ads to cull work activities, build and manage an ontology of work activities which can be applied to many human capital problems such as predicting the future of work, career exploration, and workforce planning.

Predicting personality profile scores from interview data

The research used NLP embedding techniques including word2vec, GloVe, and BERT to form word embeddings from job candidate interviews. The embeddings were then used as input into supervised machine learning models that predicted the candidates’ personality profile scores.  Results indicate that BERT embeddings combined with a bidirectional LSTM provided the most accurate prediction of personality (r-square = .66).

HCA Labs Current Projects

Hirevue

The purpose of this research project is to build a valid and reliable new Personality measure powered by a next-generation Machine Learning sustained video-based approach. To accomplish that, the relationship between traditional personality measures, video based questions, and employee outcome measures will be explored using cutting-edge NLP models to analyze the video transcript. The data comes from 700+ working adults, who were asked to complete survey questions and open-ended video questions via the HireVue platform. The survey measured Big 5 personality dimensions (Openness, Conscientiousness, Extraversion, Agreeableness, Neuroticism) as well as grit, flow, work performance, demographics, and social desirability. Using the survey data, we aim to first predict inter-correlations between the Big 5 personality dimensions, Grit, and Flow and analyze the relationships between the personality and measures of Work Performance. With the video questions, we hope to train HireVue’s AI algorithm to predict the previous measures using candidates’ interview responses.

Candogram

Job data & Language Generation

Develop a dynamic public source database containing the accurate skills and requirements needed for all jobs. More specifically, identifying how changes in taxonomy of jobs and changes within the skills needed for job success have changed over time and will trend in the future by fine-tuning language models and running conditional text generation of job ads. The data consists of multiple datasets of job ads and job descriptions, including LinkedIn, NASWA, & O*NET. The primary objective is to provide queues to the model such as “looking for digital sales executives, the required skills are:” to inform recruitment processes, training for up-skilling, and strategic business planning. 

Publications and White Papers

Jaramillo, C.M., Squires, P., Kaufman, H.G. & Togelius, J. (May 21, 2021). Overcoming Challenges to Understanding and Predicting the Evolution of Work in the Post-Covid Era, 10th Annual Conference on Human Capital Innovation in Technology & Analytics. NYU Tandon School of Engineering, Brooklyn, New York.

Jaramillo, C.M., Squires, P., Kaufman, H.G., Mendes da Silva, A., Togelius, J., (2020, December), Word embedding for job market spatial representation: tracking changes and predicting skills demand, in 2020 IEEE International Conference on Big Data (Big Data), Atlanta, GA, USA, 2020, pp. 5713-5715, doi: 10.1109/BigData50022.2020.9377850. https://ieeexplore.ieee.org/document/9377850

Squires, P., Jaramillo, C.M., & Kaufman, H.G. (May 21, 2020). Predicting the emergence and obsolescence of future jobs; AI in the Workplace: Future Directions in People Analytics, 9th Annual Conference on Human Capital Innovation in Technology & Analytics. NYU Tandon School of Engineering, Brooklyn, New York.

Squires, P., Mendes da Silva, A., Jaramillo, C.M., Togelius, J., & Kaufman, H.G. (2019, November). Temporal Projected Word Embeddings for Predicting Future Job Requirements. Poster session presented at the Natural Language, Dialog and Speech (NDS) Symposium of The New York Academy of Sciences (http://www.nyas.org/NDS2019). New York, NY.

Squires, P., Jaramillo, C.M., & Kaufman, H.G. (May 30, 2019). Understanding the Future of Work: Applying AI to Predict Job Classifications; Predicting the Future of Work: The Impact of AI, 8th Annual Conference on Human Capital Innovation in Technology & Analytics. NYU Tandon School of Engineering, Brooklyn, New York.

Squires, P., Kaufman, H.G., Togelius, J, Jaramillo, C.M., (2017), A comparative sequence analysis of career paths among knowledge workers in a multinational bank, 2017 IEEE International Conference on Big Data (Big Data), Big Data (Big Data), 2017 IEEE International Conference on, 3604. doi:10.1109/BigData.2017.8258354. https://ieeexplore.ieee.org/document/8258354

Squires, P., Kaufman, H.G., Togelius, J, Bruna, J, Jaramillo, C.M., Fang, M., (2017), SAP Leonardo ML Research Retreat, SAP sponsored research projects, 19th of October 2017.

Jaffe, E., Mo, F., Jaramillo, C., Squires, P., & Togelius, J. (2022, July 22). Understanding the Great Resignation; An Analysis of Job Ads and Attrition Data. View Here

Sponsors and Supporters

We wish to thank and recognize organizations that have provided us with data and support the mission and activities of HCA Labs.

Hirevue 

Candogram

NASWA