ALG was appointed by the European Aviation Safety Agency (EASA) as the Data Analytics Provider (DAP) for the proof-of-concept phase of the Data4Safety (D4S) Programme, the European big data analytics platform for aviation safety.
Data4Safety is a voluntary partnership programme that aims to identify systemic risks at the EU level and their mitigations. The objective of the programme is to significantly increase intelligence capacity for safety in Europe to know where to look, and to proactively address potential safety issues as they emerge.
D4S is gathering a massive collection of data from across the aviation sector, including safety reports, flight data from airlines, traffic data from the ATM system and other exposure data that might add value to the understanding of the aviation landscape (ADS-B data, weather data, etc.). The purpose of the programme is to organise the analytical capacities of all stakeholders in European aviation safety, enabling a framework under which EU member states and industry can join forces to work together at a scale never previously achieved in Europe.
The main objectives of the programme include:
- Collecting and gathering all scattered data that may support the management of safety risks at the European level, integrating these into a big data platform.
- Taking data analysis to another level, leveraging advanced analytics to extract relevant insights from this information.
- Creating a cooperative analytical framework for all aviation stakeholders to contribute and engage in safety analysis and act proactively to mitigate identified risks.
- Building a reliable and robust governance framework, ensuring protection of the shared data, and the required protocols to frame and control all operations involving the data collected and processed by the programme.
- Integrating the programme into the European safety system, making it an enabler for the current framework to implement significant safety actions.
Objectives, goals and purposes
The project involved acting as the Data Analytics Provider for the D4S programme, responsible for providing the data analysis expertise and methods, from analysing inputs provided by data providers to supporting analysis by the D4S task team. ALG was responsible for the data science and advanced analytics capabilities from design to implementation and validation, supporting the development of directed studies and dashboarding capabilities.
Study methodology and activities
The scope of work includes contribution to the definition and coordination of data analytics needs, design and deployment of data analytics capabilities and development of data-driven directed studies.
Contributing to the definition and coordination of data analytics needs involves:
- Supporting the task teams of aviation experts from various organisations in the definition of metrics and KPIs.
- Contributing to the task teams through experts or data providers to include aviation expertise through data analysis in the context of the performance of directed studies.
- Contributing to discussions related to strategic choices for data-related components.
- Contributing to the ideation and selection of projects within the D4S roadmap, contributing to the assessment of project opportunities, e.g. feasibility, complexity and criteria for success, by providing analytics insights on data relevance and quality.
Designing and deploying of Data Analytics capabilities:
- Defining the framework, determining the analytical goals and contributing to the planning definition of analytical capabilities.
- Selecting, exploring and describing relevant data from the data collected, prepared and made available in the platform, e.g. quality checks, distribution and the correlation/interaction of the available variables.
- Transforming, cleansing and enriching selections of relevant data and integrating them to build appropriate datasets, including implementation of data-fusion capabilities.
- Designing and implementing algorithms on the operational platform.
- Designing and implementing all required business intelligence capabilities, from blind-benchmark dashboards to reports.
- Selecting modelling techniques to build models and assess their performance.
Development of data-driven directed studies:
- Supporting the definition of the best data approach to meet the objectives of the use case, structuring the analysis and mapping the objectives within the framework of the available data and their limitations.
- Supporting the task team in defining the analysis required, bringing expertise in aviation data into discussions.
- Implementing the required data analytics capabilities and supporting the development of analysis, as guided by the task team.
- Supporting D4S task teams in building the analysis and developing use cases, including reporting and dissemination.
- Coordination of planning to ensure the delivery of studies in the allocated time.
Key findings and recommendations
As the project reaches its final phase, ALG has drawn key lessons from its five years of involvement. The main takeaways include:
- Break the silos and enrich the data for analysis: the aviation industry is an intertwined ecosystem. Analysis of isolated data sources provides a limited and biased view of reality. Therefore, it is crucial to break silos and integrate data to give a comprehensive vision of operations and processes, irrespective of where the data originates (different departments within a stakeholder, multiple stakeholders, etc.).
- Building trust in data is a must: we cannot rely on digital assets to support our decision-making if we do not trust the quality of the source data. Data quality is the first step in any digital initiative. If we do not trust the source data, we cannot trust the outcome. This is often a time-consuming task that requires the involvement of technical and business teams to work together towards improving quality. It is also important to ensure data consistency and security, especially when working in environments that integrate data sources from various stakeholders. The programme has built a robust governance framework to ensure the protection, security, integrity and traceability of the data shared in the big data platform.
- Cross-functional teams are essential: by their nature, digital projects apply cutting-edge analytical techniques to untapped and, most likely, previously unexplored data. Defining relevant digital assets requires the involvement of all stakeholders, investing effort in engaging them in the overall technical development process to build data-based use cases under a cooperative framework. Involvement of a cross-functional team combining data and aviation expertise has been key to the success of these initiatives, being able to engage and coordinate efforts on the technical and business sides of any solution implemented.
Success and outcomes
The D4S project began with a feasibility study in 2015, moving into proof of concept from 2017 to 2022. The proof-of-concept phase was a success in terms of both the technical demonstration and the governance framework, proving the maturity of the programme in ensuring a protected and reliable environment for data analysis for all of the organisations involved. The main results of D4S are available on its website.
Materialising this success, the project is now in the development phase, which will run from mid-2022 to 2025. This will expand the membership of the project. More member states and airlines will join first, along with the first ANSPs, followed by airports, gradually embracing the complete European aviation sector by the end of the Development phase.