November 16, 2017
Predictive analytics is poised to fundamentally transform how EHSQ programs are managed. In 2018, we’ll see more organizations in high-risk and highly regulated industries launch predictive analytics programs to reduce risk, improve employee health and safety and drive operational efficiency.
Rolling out a predictive analytics program that will meet the needs of your organization can be a big undertaking, requiring significant planning. And if you're like most forward-thinking EHSQ professionals, you're probably in that planning stage right now. To aid you in your journey to analytics excellence, below are four key factors to keep top of mind before launching your program.
Related: [WEBINAR] The Four Pillars of Analytics Excellence for EHSQ
1. Data Confidence
Right now, your organization is probably collecting data for compliance and regulatory purposes and to support day-to-day decision making – but how much confidence do you have in this data? Is it accurate and timely? Or error-prone and stale? Often, when it comes time to input data into an EHSQ management system, key components are missing and forms are only partially complete. Data entry can be tedious and time-consuming for frontline workers - but without accurate data, you can't make informed decisions or improve outcomes.
Key takeaway: make your data capture process as easy as possible so you're consistently collecting high quality data.
2. Data Security
Without question, your two greatest assets that need the most protection are your people and your data. When looking at information security, consider who in your organization should be allowed access to data and at what level. Should managers have access to data at a global level or only for their specific sites? As you’re dealing with sensitive data like employee health records, it’s critical to have the right control practices in place at both the field and record level. Then, ensure everyone in your organization adheres to these security standards.
Key takeaway: make sure you have a control mechanism in place that can grant or prohibit employee access to certain types of data.
3. Data Privacy
Data privacy is more than just protecting personal identification information or aggregating your numbers. These are essential steps, but to ensure privacy – you need to go beyond the basics so that the values of your sensitive data cannot be reverse engineered. As you collect more data and move forward with your predictive models, ensure a differential privacy mechanism is part of your predictive analytics program. This way you can significantly minimize the chances of records being identified.
Key takeaway: go a level deeper to guarantee your values cannot be back-calculated - leverage differential privacy.
4. Data Expertise
There’s a specific skill set needed to tackle data projects of this magnitude – and the people who possess this quantitative prowess are data scientists. When working with such vast quantities of data and predictive models, you need a team of specialized experts on board who can show you how your data should look and behave, point out possible limitations and areas for improvement, and can ultimately ensure your predictive analytics program is a success. Or, leave it to the experts – some solutions automate the normalization of data, then aggregate and analyze that data to reveal key insights and drive better decision-making.
Key takeaway: choose your partners wisely - if you don't have access to an in-house data science department, consider teaming up with a vendor who can build and interpret predictive models with you.
For more information on planning your predictive analytics project, watch our webinar, The Four Pillars of Analytics Excellence for EHSQ.