Moving away from paper-based tracking system to automated way of managing it via ERP software is one of the major achievement for many organizations. Mere using the technology to capture the incident information is a normal business as usual practice, effectively harnessing the technology to support risk management is the greatest opportunity for most organizations. Risk Management not only enable the performance but also protect the business. It is found that majority of drivers for operational risk management are technological in nature like Industry 4.0 principles, competitive marketspace and cost-effective data automation.
Understanding how the risk is managed within the organization and how employees perceive risk management and what are the factors which are causing risk is a pre-requisite for effective management. The factors affecting safety incidents, contractor / employee productivity and employee perception about safety are important factors in understanding and preventing the risk.
With the availability of SAP EHSM Incident Management and SAP Predictive Analytics capability which has ability to connect with “R”, it is possible to analyse Risk, predict risk and Prescribe the risk. In this topic I am going to give an overview of SAP Predictive Analytics use case on how it can be effectively used for operational risk management for managing the incidents.
To give a brief introduction to SAP EHSM incident Management – it is a tool which can be used to capture Incidents, Near Misses and safety observations within an organization. It can be used to capture incident information and to carry out the investigation analysis. The EHSM data structure allows for capturing additional information covering incidents as required.
SAP Predictive Analytics is built for Data scientists and business analysts for making predictive analysis. SAP Predictive Analysis offers in two modes Expert Mode and Automated mode and it is integrated with SAP HANA database. There are many data types which are relevant for data science consisting of the essential “V”s . Inside SAP data can be demarcated into Transactional data, unstructured data, Real Time data, Location or Geo data, or machine data. The data sources in SAP include HANA data so that data analysis can be done within the database and or it can be extracted and merged with other non SAP data and read from Hadoop via data sources and analysed further.
Once conditioned you can use them into PAL or Predictive Analytics Library, Text search, Spatial analysis, Graph analysis so on. The algorithms in PAL and the R integration for HANA can be invoked from SAP Predictive analysis, a client tool for the definition, exploration, execution, visualization of predictive analyses. The PAL is essentially table based in the sense that for each algorithm you have three tables, an input table, a parameter table and an output tables. You can find details about PAL in product literature. Once analysed the data can be fed to SAP Predictive Analytics, SAP Lumira, AFM, other tools like Analytics and BI Tools.
All the predictive analytics which are used in Operational Risk Management can be classified into one of four categories: descriptive, diagnostic, predictive, and prescriptive analysis. Descriptive analysis typically helps to describe a situation and can help to answer questions like What happened? What is injury rate, how many injured by department and other similar statistics?, etc. Diagnostic analysis helps you understand why things happened and can answer questions like Why did it incidents happen? Predictive analysis is forward-looking and can answer questions such as What will happen in the future given the current and past trend in incident rate? As the name suggests, prescriptive analysis is much more prescriptive and helps answer questions like What should we do? What is the probability that an equipment fail resulting in incident?, or How should manage my maintenance investments?
Descriptive analytics of Incidents:
Based on the injury / illness data, details about the safety pyramid analysis can be carried out using R Extension using pyramid.R plot function and SAP Predictive analytics also shows the Staked bar chart. Text analysis can be carried out to analyse the incident nature and details of the occurrence. Computation of various statistics namely LTIFR, Accident Severity Rates, Accident Frequency Rate distribution by country / region can be analysed. Time series analysis will help in understanding the seasonality. The common analysis parameters namely Incidents distribution analysis, drill down of Accident types, technical factors of incident occurrence and correlation between incidents and other SHE factors can be performed by built in analytical queries.
Diagnostic analytics on safety perception:
Once details about Incident / Injuries are assessed, employee sent analysis can be performed using natural language processing is used to cluster the sentiment based on the words. Based on the detailed interviews captured in injury / illness investigation sheet or periodic surveys data, it can be analysed to understand the sentiment. The behaviour-based information or employee perception about safety culture is very useful in identifying the safety sentiment and injury occurring scope in an organization. The employee comments on safety over the years can be used to capture sentiment. A detailed assessment can be performed on various job roles and nature of the sentiment.
The details on the Sentiment score matrices will help in understanding the employee engagement in managing the injury / illness within the organization and help in chalking down measures and provide corrective measures.
Predictive analysis of incidents:
Incident / Injury data based on the event type, occupation, nature of event, nature of injury can be used to predict the future events. Data which is captured in the system can be used to predict the future trend. Time series analysis using “R” extensions in SAP Predictive Analytics can be used to identify the trend. Similarly, topic mining and clustering techniques can be used to extract insights about the incidents. With SAP S4HANA spatial analytics detailed geo mapping of injury information can be analysed and with built in Analytics statistics like month wise injury information, different body part and injury information and distribution based on department can be performed.
Predicting equipment failure leading to incidents / near misses
Using a streaming sensor data, it is possible to predict the equipment failure. In a workplace common equipment’s sensor data is analysed for target variables namely say vibration, noise, electricity / fuel consumption pattern. Based on the historical analysis it is possible to identify / predict the equipment failure which might result in injuries or near miss.
Finally to conclude some of the challenges in deploying data analytics for operational risk management include good data availability, data analytics scalability, how best to make use of advanced predictive analytics tools, domain expertise to analyse the data models. The nature of data being used depend on the type of the analysis you are trying to carry out for some types of risks you can use the in-house SAP EHSM captured data and the input is straight forward, whereas for some types of risk the data / indictors are not visible, and it needs careful planning along with external data to come out with a solution. The practice of Data sciences seems like a lot of math’s and formulas, but beyond this very boundary lies the real business benefit. No doubt the area of data science is hot topic and the more we prepare ourselves for it, the better we will place our self for managing the future. The better our insight into the future, the greater the potential for competitive advantage
Be prepared for this new and exciting field by knowing more about SAP Predictive Analytics, “R”, PAL, and other Machine Learning offerings.
Thanks – Jak