Human Resource Data Analysis

Implemented residual sum of squares to detect change points for identifying a level shift in the time series data (in the HR domain) for one world’s leading professional services firms. Developed a KNN causal estimation algorithm for determining the causal strength of KPIs after using partial correlation to define independence between the variables. Also developed ingestion pipelines and prepared de-normalized data formats for the data lake. The use case involved had 22 KPIs along with a monthly data refresh of roughly 25GB. The analysis improved the accuracy and time taken in attrition and hiring analysis by 23% and 35%. It also helped in indirectly improving their financial budget by 8%.