The business firm conducted 360-degree feedback annually to understand and evaluate the performance of the employees. The feedback is usually a hard copy of the electronically available review form manually filled by the employees or sometimes filled online. In either case, much of the information is in unstructured form (text). The manager does evaluation of performance intuitively. This process is cumbersome, and comparative insights are seldom used. Client wanted us to draw insights from the data and setup an infrastructure for the same so that they may be able to do it on their own in the future.
We collated data from multiple 360-degree review feedback reports (PDFs). Performed data loading, designed a multi-dimensional model and loaded the data into a SQL database. For descriptive analysis, we designed an interactive dashboard that summarizes the information, and provides drill-down and roll-up operations to get insights from the data. The analysis was done across five different dimensions - people who reviewed it, people who are being reviewed, features of review, scoring and textual comments. To analyze textual comments, we applied text analytics. For predictive analytics, we built models to find clusters of different features along which all managers had similar feedbacks.These clusters indicated that the features are either not distinctive or that reviewers really did not comprehend the subtleties of the features. Linked the online review form to the data base and connected the database to our recently designed interactive dashboard with built in analytical models for future use.