We need information to ‘inform’ us, specifically when making decisions. Visualised data becomes information. We are visual beings and through the eons of time are adapted to quickly converting images into decisions. “That looks dangerous, let’s run”…..
So, we decided to exploit the vast amounts of data our clients are collecting, combine it with sophisticated algorithms, analytics and visualisations in order to support practical workplace decisions such as:
- ‘What happens when John leaves, what is our risk?’
- ‘If we create a new position who do we currently have who are candidates?’
- ‘What jobs have the lowest percentage match to the workforce?’
- ‘Do I have someone who we can second to cover for long service leave?’
- ‘What job has been selected the most as a future job?’
Our clients have a LOT of data. Our result table contains over 4m data points and even for a single person or a single job description (or business need) we have a huge 703,722,570 possible skill combinations (based on 4 possible values for each of the 362 skills/levels).
Of course, we could manually compare persons to job descriptions or write clever ‘lookups’ in Excel but this can be time consuming and gets complicated if we introduce weightings. What if a person has a skill/level identified as a previous1 skill and this matches a desirable skill/level in a job description? Should this score the same value as a skill/level identified as mostly2 matching a mandatory skill/level in a Job? We think not.
So, to use our data to help make decisions within workforce planning and, maybe more importantly, help the workforce plan their future employment journey we have to create methods of visualising data in way that can be easily understood, absorbed and acted upon.
So, here’s what we did
We needed to build an algorithm to calculate a percentage match between persons and jobs. This takes account of all the various answer combinations and presents it as a very recognisable percentage. So, if a person scores 100% they have every mandatory and desirable skill/level required by the job description and they have indicated all the skill/levels as mostly2. This information we have been presenting in corporate reports from day 1. It provides good information but requires searching and sorting to isolate specific results. We wanted to go further.
Our next task was to see things from an employee perspective. Skill/levels are NOT static, as a person progresses through their career they acquire new skills and others become rusty. To empower the workforce we created a ‘Planner’ console allowing individuals to visualise their skill/levels and associate themselves with the Job descriptions that their organisation has created and made visible in the Planner. Every employee can see their skill profile, select a current and future job and see their percentage match. If they change their skill profile the percentage match recalculates immediately. This is a very useful enablement tool for career planning but we went further. Using visualisation as our guide we allow for a job description to be ‘overlaid’ with the persons’ own skill profile so they can immediately and easily see gaps. With the gaps visible planning is now informed and practical.
By implementing this functionality we have acquired another layer of data. That is, we now know relationships between the workforce and the jobs in addition to the percentage match. We know what job a person currently performs and maybe more importantly we know the jobs they would like to perform in the future. As previously mentioned we already produced reports listing jobs and persons and percentage matching, we think it worked well but had limitations (being a table in a report) and so to add another layer of data would compound its usefulness. This sent us in search of a visualisation technique that would reveal information from complex data AND be searchable and dynamic. Our panacea came in the form of Force Graphs. A way to relate ‘things’ together and indicate the ‘force’ of a relationship. In our case the ‘things’ are people and jobs and the relationship force is the percentage match. And as an added bonus we can also visually represent the type of relationship with colour.
Force graphs can get messy. Within any organisation the relationships between the workforce and the jobs is often more complex than you would imagine. This means that for force graphs to support decisions we need clarity, looking deeper but narrower into the graph. Fortunately, this is relatively easy if we implement page filters within Microsoft Power BI®. Because we know the attributes of data behind the force graph we can start to apply filters:
We believe we are making a difference. While all the answers from our clients questions may not immediately present themselves from via force graphs, they will usually provide the information needed to use our other Analytic pages to get to the detail required.
Oh and finally, Power BI® force graphs animate and bounce, which let’s face it is pretty cool.
Macanta is a gold reseller of SkillsTX