By: Greta Roberts, Conference Chair, Predictive Analytics World for Workforce 2016
In anticipation of his upcoming Predictive Analytics World for Workforce conference co-presentation, Open Sourced Workforce Analytics: An Overview of 3 Algorithms for 
Common Predictive Modeling Situations, we interviewed Jason Noriega, Diversity Analytics Team Lead at Chevron. View the Q-and-A below to see how Jason has incorporated predictive analytics into the workforce of Chevron. Also, glimpse what’s in store for the new PAW Workforce conference.
Q: How is a specific line of business / business unit using your predictive decisions? How is your product deployed into operations?
A: Attrition of top performers and other employees with critical skills is a major concern. At many of my prior companies and roles, many of the predictive modeling products I have developed for business units involve:
- Identifying important variables that impact turnover;
- Visualizing patterns of high risk for turnover;
- Utilizing the understanding of those patterns to improve retention.
Business units have used my predictive models to
- Improve the interviewing/hiring process;
- Improve hiring of candidates who are most likely to stay;
- Improve the effectiveness of people managers.
Q: If HR were 100% ready and the data were available, what would your boldest data science creations do?
A: In that dream scenario, my boldest data science creations would
- Predict the performance of job candidates before they are hired, and thus improve hiring decisions;
- Create an employee attrition risk application with controllable variables that can be manipulated to see the impact on attrition risk scores for individual employees;
- Predict how long it will take to fill a position, given the specific characteristics of that position;
- Use web scraping techniques to pull public data of employees to combine with internal company data and improve predictive models of attrition;
- Develop an interactive university hiring simulation model to predict and optimize the diversity of hires.
Q: When do you think businesses will be ready for "black box" workforce predictive methods, such as Random Forests or Neural Networks?
A: Although they may not know it yet, most businesses are ready for black box methods, as long as they are visualized in an intuitive display for business leaders with all levels of technical backgrounds to understand.
For example, in March 2015, I participated in an employee attrition predictive modeling competition on CrowdANALYTIX.com, in collaboration with my teammate, Nery Castillo-McIntyre, and won 1st place out of a pool of 330 data scientists. We used a black box method in order to improve predictive accuracy, visualized the predictions using Tableau, and then presented clear visualizations in the form of an interactive dashboard.
These interactive visualizations helped the client see clearly the critical variables identified by the black box method in an easy to understand format that could be quickly deployed to target individual employees for retention.
Q: Do you have suggestions for data scientists trying to explain the complexity of their work, to those solving workforce challenges?
A: In order to take an analysis from theory to practice, decision makers need to clearly understand the data scientists’ work. For that reason, data scientists must be mindful of their audience, keep the complexity of their work to themselves but be ready to show it upon request, and explain their work in a simple way. Effective visualization is the essential key to success in this regard.
Q: What is one specific way in which predictive analytics actively is driving decisions?
A: Many companies have started using publicly available employee career history data and deriving key variables to use for predictive models.
For instance, the data used for the employee attrition competition my colleague, Nery Castillo-McIntyre, and I won was scraped from the web. The variables we extracted from it were highly predictive, and the client could use the patterns we found to hire prospective employees most likely to stay, and to develop targeted retention efforts for employees already onboard.
Q: How does business culture, including HR, need to evolve to accept the full promise of predictive workforce?
A: In order for the business culture to accept the full promise of predictive analytics, there must be change in three areas: people, processes, and technology.
People: The right leaders must be able, ready and willing to utilize predictive analytics to make more informed decisions. This includes analytically-minded business partners, managers who want more than mere insights into reports, and executives who drive culture change.
Processes: Driving culture change requires a deep understanding of how HR and the business at large carry out their functions, and where in the process key predictive information is needed to make better decisions. Crucial data elements must be readily stored and accessible to be used to generate value.
Technology: An important tool to shift the culture, this may include off-the-shelf technology, systems integration and data warehouse construction, as well as open source technology for advanced analytics.
Q: Do you have specific business results you can report?
A: For many of the companies that I have worked for in the past, examples of some of the business results that have been gained from my predictive models included:
- Making significant improvements to the diversity of university hires;
- Improving short tenure attrition by hiring candidates who are most likely to stay;
- Reduced employee attrition following maternity leave.
——————————-
Don't miss Jason’s conference co-presentation, Open Sourced Workforce Analytics: An Overview of 3 Algorithms for Common Predictive Modeling Situations, at PAW Workforce, on Monday, April 4, 2016, from 10:40 to 11:25 am. Click here to register for attendance.
By: Greta Roberts, CEO, Talent Analytics, Corp. @gretaroberts and Conference Chair of Predictive Analytics World for Workforce.