Stennis Space Center’s Safety and Mission Assurance (SMA) Directorate combined its safety data from seven separate databases to create a more streamlined collection of data. The center hopes the system, known as the SMA Crystal Ball, will eventually predict future safety incidents based on past data to help Stennis management better allocate center resources.
Safety incidents — including injuries, property damage and mission failures — cost NASA and contractors thousands of dollars in direct and indirect costs. The leading indicators to these incidents are located in the thousands of data points in agency data systems. By combining the leading indicator information from the NASA Mishap Information System; Close Call Reporting System; Audit Tracking and Information System; Decision Data Management System; Integrated Risk Management Application; Maximo; and the Safety, Health and Environmental Tracking system, Stennis set the stage to then use that collection of data to develop and test an algorithm to predict when and where there is a high probability of a safety incident occurring. (Note: These systems were hosted by various centers including Stennis, Marshall Space Flight Center and the NASA Safety Center.)
Although the algorithm needed for prediction is still in its infancy, the SMA Directorate plans to release the database in early Fiscal Year 2019 (FY19), as the combined data alone will be a timesaver when trying to locate data and it provides the added convenience of all data points being presented in the same format (rather than those of the various systems).
“We immediately are going to be able to save our people time and give them the types of tools to do relational trending and be able to do parts of their jobs better,” said Kamili Shaw, lead; Safety, Quality and Management Systems Division, SMA Directorate, Stennis. “This will allow us to put our resources where they can be of most benefit to reduce incidents and mishaps and implement facility safety improvements throughout the life cycle of a project.”
Although the concept of the SMA Crystal Ball started in 2014, Stennis kicked off the effort in 2017. The first step was identifying INSIGHT as the proper platform to house the new database, a system already developed by the Office of the Chief Information Office and used on programs like Orion. INSIGHT allowed Stennis to integrate several data sources with a search and export capability.
“This OCIO platform really enabled us to jump forward on this work,” said Shaw.
After importing the data from the various systems, an intern started looking for various ways to categorize the data, as well as various methods for using it to make predictions. Although initial predictions didn’t achieve high enough levels of accuracy to implement straight away, the work done to date shows the pros and cons of the various methods, and Stennis is now able to map its next steps to improve the prediction capability by improving the data.
As Shaw explained, Stennis isn’t seeing as many safety incidents in recent years, which makes machine learning difficult as it requires a lot of data points; while fewer incidents is a positive for the center and the agency as a whole, it requires the team to narrow-in on the most useful data for identifying relationships and therefore predictions.
As the SMA Crystal Ball moves into this new phase in FY19, Shaw and others in the SMA Directorate are interested in sharing their story to-date with other centers who may be interested in implementing a similar system.
“We could conceivably use any agency data — a smaller or larger set of data — the platform is expandable,” said Shaw. “We’re really showing what can be done with data everyone has sitting around. We’re showing what we can do with tools that already exist at NASA — we didn’t buy anything new. [And we’re] showing how we can make this data more valuable, and that should be transferable to a lot of different areas.