Safety Data Management


Gaining central access to flight safety related information via the avialytics Safety Data Management solution.


Key Features

• Single isolated safety-related data store

• Discretization and abstraction process

• Application to visualize the data of specific flights

• Multilevel scorecard to put the safety events into a wider context

• Data mining to suggest influential factors

• Agile expandability

Evolving Flight Safety data

into actionable insights

In order to achieve the desired improvement in flight safety performance the following areas have been identified. They are centered on the people involved and responsible for quality and safety and positioned as decision support tools to enable, enhance and focus safety relevant knowledge.

Data Capture:

Safety data hub
You can only manage what you can measure and a common ground is needed to draw conclusions from various sources of information. Regarding flight operations a humongous amount of potentially safety relevant information exists in a large variety of ways and on different levels of aggregation and structure.
Following up on the idea of a “Single version of the truth” in a data warehouse environment, a relational data mart is the core component of the avialytics flight safety analyzer. Together with a multidimensional data store for high performance analytic processing it is building a big data enabled repository for all kind of flight safety information.

Data that is transferred into the repository passes through an optional and individually configurable discretization layer. The purpose of this layer is to concentrate raw information into meaningful clusters and to protect the personal rights e.g. of the flight crews. The level of discretization can be adjusted and applied to future as well as historic information as long as the original data is still accessible.

Even though many information that can be found in an airlines’ data warehouse will be included in the Flight Safety data store, this redundancy is by design. The flight safety repository is explicitly isolated from all other data stores in a company and information can only be transferred going through the abstraction process, configured by the customer and integrated into the data retrieval & capture agents. Access to the flight safety data store is limited to authorized users in the quality and safety environment.

Once data is available in the repository, information provider agents make this information available to analysis and reporting tools as well as downstream operational systems.


Unified Data Access:

Flight Safety Box
The primary way to access the information in the flight safety repository is the avialytics Flight Safety Box.

The flight list is a common entry point which presents an overview of all flights with safety specific indicators and various and flexible filter options.

The safety indicators displayed in the flight list can be dynamically configured and customized based on available and meaningful information.

Examples for flight safety specific indicators are e.g. “FDM event triggered”, “Occurrence Report filed”, “Hard Landing indicated”, “Fatigue”, Weather, …

An important aspect of this consolidated view is the combination of events from various sources. E.g. the FDM reports stick shaker activation and there is a corresponding occurrence report generated by the pilots via the EFB. In the same scenario the absence of such occurrence report would also be valid and potentially very important information.

The default filter of the indicators relates to the classification of the safety relevant information and consists of five stages:

Color Meaning Risk Example
transparent No information available No FDM event triggered
grey Information available, but not significant Metar WX Info
green Minor significant low
yellow Major significant medium
red Critical significant high

These stages can be adjusted based on the needs of the customer and the classification is either transferred from the originating system (e.g. FDM) or can be configured within the discretization layer.

From the flight list a drill down into the details of the flight can be initiated. The different safety aspects of each flight are represented by separate tabs.

The content and format of the tab can also be customized to the needs of the user. Furthermore it is possible to add additional tabs for new areas of interest or in case a certain area grew too complex to present all meaningful details on a single page.

Find below some short examples of tabs that have already been created for the application and will be included in the basic setup.

Flight Safety Box: General

The purpose of the “General”- Tab is to present background and statistical information about a specific flight, in order to put a safety event into a wider context.

Which information is displayed and the granularity depends on the requirements of the flight safety department and upon which data is deemed useful for further generation of knowledge.

Flight Safety Box: FDM

As the name suggests, the FDM Tab is dedicated to the FOQA information. It is possible to integrate only information that triggered a flight safety event or a wider range of parameters (like usage of auto throttle), that can be retrieved from the FDM data and might be of interest even or especially when no safety event was triggered.

A FDM event can be linked to a safety or occurrence report from a different source. It can also be invalidated to be excluded from further statistical processing, e.g. if a long landing was the result of a training session.

Flight Safety Box: Occurrence
The occurrence tab is used to display and structure mandatory and voluntary occurrence reports e.g. entered in an EFB solution. Like the other tabs it can be designed to show all kinds of information on different aggregation levels with the option to include attached documents or photos etc.

Flight Safety Box: Weather
The information displayed in the weather tab is populated via a data retrieval agent from the Aviation Weather Center at NOAA’s National Weather Service.

It displays the METAR information received throughout a two hour time period around the Offblock/TakeOff Events at the departure airport and the Landing/OnBlock events at the destination airport.

Flight Safety Box: Multileg

On the Multileg tab a specific flight is brought into the context of the crew duty period it is part of.

If e.g. the duty period of a flight consists of 5 legs, all of these legs will be visible on the Multileg page, along with an indicator of all safety relevant events that occurred throughout that duty period. It is possible to differentiate the affected flight phases and to navigate to the corresponding tabs with detail information about the events.

Flight Safety Box: Gate Keeper
In case the available data is not sufficient to explain a certain scenario, a gatekeeper process can be activated to retrieve additional information from the people involved.

Flight Safety Box: Metric

The metric tab shows a selectable measure of the current flight compared to all flights operated in the past for a selected metric. Metrics are predefined properties of the flight and are organized in metric groups. Base metrics predefined are e.g. departure and arrival airport, aircraft, window of circadian low, crew complement, flight and duty time.

Exceptions can be visualized graphically and quickly and with different granularities to reduce fuzziness.

Flight Safety Box: Assessment
Conclusions drawn from the analysis can be entered and formalized in the assessment tab. This tab is the predecessor of the future avialytics quality and risk assessment feature.


SPI Definition:

KPI Designer
Depending on their rights users are able to define their own Safety Performance Indicator (SPI) to visualize exceptions and to track safety and process relevant measures. Limits and color coding help to identify spikes. These limits can be used to subscribe to data driven reports to get a notification if the limits are exceeded. SPIs can be shared between users and may be used in individual analysis using Flight Safety Cube or reporting as well as in the user’s flight safety dashboards.


Performance Monitoring:

Flight Safety Dashboard

The avialytics flight safety dashboard is the entry point to access current and historic safety information in the form of a multilevel scorecard including user defined safety performance indicators, moving averages and trends. It also presents an overview of the selected safety metrics (factors identified to potentially have an influence on safety relevant occurrences like weather conditions, crew experiences, load factors …)

Whenever one of these metrics diverts from the statistically determined expected range, the metric is highlighted for historic and further investigations.

Using the Flight Safety Box or the Flight Safety Dashboard further investigation can be started by drilling down into the details and context of a specific occurrence. Whereas the Flight Safety Box gives the users access to the individual details and the surrounding attributes of the specific flight the dashboard puts the occurrence statistically into a wider context and allows for a slice and dice of all potentially influential elements.


Data Analytics:

Flight Safety Cube
The flight safety cube ensures decision makers and analysts are always provided with appropriate, accurate and easy to consume information. Through the automation of analysis and data aggregation, the multidimensional analysis accelerate critical decision making, eliminate dependency on analysts, and empower leaders with the insight to drive your safety management strategic imperatives. Multiple frontend applications can be used to build workbooks, dashboards and reports easily. Multidimensional analysis enables the user to analyze flight safety data totally flexible and at best performance. Thus empowers to follow the train of thought without being interrupted for gathering additional information.


Automated Knowledge Discovery:

Statistics & Data Mining
The avialytics Flight Safety Box links and generates millions of directly or indirectly flight related data and sets them into relation with safety relevant events. While its primary objective is to allow a user to access the information in various and flexible ways to gain insights, a data mining approach is used to suggest influential factors based on automated algorithms. The results again are presented to the user as an additional starting point for investigations.


Double Loop Learning
Lessons learned and process improvements are central elements of the avialytics Flight Safety Box. While analyzing data, reviewing information and gaining insights, additional ideas and information domains will inevitably be identified. To cope with that challenge data retrieval processes underlying the Flight Safety Box can be extended for new data sources by providing a new or extending existing data retrieval agent. On the user part, new metrics and SPIs can be defined and integrated into the existing configuration with the option to apply the new structures to the existing historic data, refining previous analytic results.