Research on Workforce Development ModelingYuchen YaoBlockedUnblockFollowFollowingFeb 17Recently I have been reading and researching about modeling workforce development in a data-driven way.
In this post, I’ll summarize two literatures that I reviewed and hopefully it would be useful for our future work.
The first one I read about is an introduction to build workforce models by the Healthy London Partnership Workforce Programme in 2016 (https://www.
This introduction covers several important topics: what workforce modeling can do, how is it carried out and how to build a successful workforce model.
· Workforce modeling is applied to support workforce planning and development.
It might be used to help put right people with right skills in the right place and in the right time on the right job position, with the right cost.
· It is carried out by utilizing existing data and quantitative methods to analyze the needs for an organization and help the workforce development.
The key point is to gather suitable data for developing the needed workforce models.
· A successful workforce model should identify the key stake-holders, should have well-defined customers and owners, as well as clear scope and objectives.
The data used should be reliable, the method used should be robust to support the model, etc.
· To develop a workforce model, a clear and logic process should be adopted.
For example, the planning and research, building baseline model, workshop on the model, evolving the model, etc.
Process to develop a workforce modelIn the planning and research phase, we should understand the problem and available approaches and data, which is exactly what we are doing right now.
Also, we should facilitate constructive dialogue with stakeholders and set clear goals and scope of the model that we are going to develop.
In the model building phase, we should be able to get suitable data and apply suitable methods to it, based on what we’ve researched and decided.
There are problems that we need to consider in the mean time like how to test our model, how to address the gap between the result generated by the model and the truth, etc.
After we finish the development of baseline model, a series of workshops should be held on evaluating the model and how to evolve it based on the performance and the possible new needs or change of scenarios and so on.
It’s an iterative process that keeps updating the model and improving the performance.
To summarize, this literature offers us a road map in developing workforce model, which we can refer.
And it also points out some key points and problems that we should consider and address in the process.
The second literature (Wan, Hung-da, Fengshan F.
Chen, and Glenn W.
An intelligent decision support system for workforce forecast.
TEXAS UNIV AT SAN ANTONIO, 2011.
) carries out an extensive review of the existing methodologies and technologies in workforce modeling and also provides a decision support information about when and how to use each model.
· The literature reviews a bunch of state-of-the-art techniques in workforce forecasting and summarizes them in the graph below.
I’m not going into details about every method listed below since there are detailed explanations and analysis in the original literature.
What I think is importance here is to get a brief picture of what methods are available so that we can choose from the pool when needed.
State-of-the-art techniques in workforce forcasting· The literature also offers a Scenario Specific Forecasting Technique(s) Selection Tree presented below which can serve as a guide for selecting quantitative method based on what data we have available.
Note that the decision map is built for workforce forecast task but not for other task we might be doing like recommendation, etc.
But it could still serve like a reference of some key points we should consider like what’s the scale of our data, what’s the granularity of our data, etc.
Scenario Specific Forecasting Technique(s) Selection TreeAfter reviewing the two literatures, I got a picture of workforce model and how to build it successfully.
The next step is probably to research certain models based on what our goal is and what data we are going to use.
.. More details