HR Analytics & Machine learning case study

Hiring 'A' players on the team is one of the most important decisions any company. Companies generally sit on wealth of data related to prospective candidates, interview decisions, performance of selected candidates over the years, candidates that have left the organization, and other important elements of candidate selection decision making.

A fast growing staffing firm was diligent in maintaining information about the candidates that it had submitted, decision made by the client for each of the submitted candidate, success of each candidate at the client site, candidates that were not the right fit, and more. The staffing firm wanted to improve its closing ratio performance without incurring additional cost for recruiters.

Data & Analysis
We reviewed the data related to candidates, decision making process, success / failures of candidates, decision making style of each client and other related data elements. Exploratory analysis was conducted to determine which elements have correlation?, how is the data structured and what is the format?, are there outliars? what is the range, mean, deviation of the datapoints? are they categorical or discrete or continuous variables?

After careful data analysis, the team prepared data for model development by splitting the dataset into training and test dataset. The data was checked for multicollinearity and several supervised and unsupervised models were considered. The final model consisted of limited number of data elements that were strong predictors. The model was able to provide the likelihood a candidate will be rejected by the client with a fairly high degree of accuracy and sensitivity.

Once the internal screening process has been completed, the firm uses the model to evaluate if the candidate will accepted or rejected by the client. This has helped the firm limit the number of candidates it is submitting and in-turn improved their closing ratio.