The amount of data produced by businesses doubles roughly every two years and aviation is no exception. In fact, the aviation sector has always been at the forefront of digitalisation and produces petabytes of data that is mined to create new opportunities for innovation – these innovations normally produce even more data to unlock even more opportunities!

As a senior consultant at Egis, I often come across clients who face operational problems that can be solved simply by looking inwards at the data that is already available within their organisation. Other times it’s a case of data collection and management processes not being established or indeed missing completely. Despite companies paying significant sums for state-of-the-art data collection/processing systems, the people running them are often unaware of the full capabilities of these systems and the data/outputs available at a mouse click. In practice, the data required to solve the client’s issue is almost always readily available at the client’s site and the question is not “do we have the data?” but rather “how to interpret the data we have to solve the problem?”. And that’s usually the point where we step in and propose a suitable solution. This could be based on modification to existing systems (for example design of new database queries) or development of a new database or dashboard to facilitate data extraction and analysis.

For example, we helped a major UK airport automate a process for the development of regular airspace performance reports that helped improve their relations with neighbouring communities and noise action groups. Previously, the data was manually collected and processed, which used to take several days. Our solution used a set of dedicated queries to export the required data from the airport’s reporting systems into a business intelligence tool that could autonomously perform the required analysis. The outputs were linked directly into the reporting template so that the only thing left for the client team to do was write the commentary on the results according to their strategic objectives. Using this new process, the airport was able to save significant effort on preparation of these reports.  

In another project for a European air navigation service provider, we were asked to identify and quantify which factors affect the horizontal flight efficiency of flights within their airspace. At the beginning I was worried that the data to quantify the impacts may not be readily available. However, as in most cases, it turned out the client already had the answer within their data eco-system but was unable to unlock its value to find the answers. We helped them connect the dots and automated the analysis process by creating a dashboard linked to their own data collection systems. The client now runs periodical flight efficiency analysis on their own.

Finally, availability of data is only half the story. In order to interpret the data correctly we need to know what exactly it is that the client needs, as well as who else (other than the client) might have an interest in the final outputs. Aligning the requirements raised by different parties can often be more difficult than the question of data availability itself. For example, in one early project, we agreed a set of requirements with a client’s operational team and as the project neared its completion, the Board of Directors became involved and suddenly expressed requirements different to those agreed with the operational team. That situation no longer arises since we improved our own requirements capture processes.

Looking back at these and other ‘data mining’ problems that we’ve helped solve there were things that all of these projects had in common:

  • Availability and volume of data.
    There were many times when we worried whether the data to deliver the project was available. But looking back over the more than 10 years I’ve been doing this work, I cannot remember a single project or a client where we needed to purchase external data or carry out an onsite data collection exercise. Sometimes the data was buried down in the depths of client archives, other times it just needed minor updates but, I think it is safe to say that in general there is enough data in the aviation industry to answer (almost) all sorts of questions.

  • Uncertainty with regards to requirements.
    Only a handful of projects had a clear set of requirements defined from the beginning. Typically, it takes a series of discussions with a number of people in the client organization to understand all potential points of views and (often contradicting) requirements. Utilizing our in-house process for development and capture of client requirements we have always been able to meet or exceed expectations.

  • Drive for greater efficiency.
    Across aviation the focus has been on finding ways to improve working processes and data analysis. For clients in Eastern Europe, this was usually about sticking to existing tools/platforms but improving the working methods and/or developing new functionalities for these platforms (ie. in form of plug-ins, scripts or macros). However, with clients in Western Europe, we have witnessed a drive to move away from the typical Microsoft Excel/Access approach in pursue of dashboards and cloud-based solutions.

With the volume of available aviation data growing exponentially, artificial intelligence systems that are already being deployed in aviation will have ever increasing amount of data to train on. In turn, we can expect a gradual decrease in human involvement in data processing/analysis and increase in accuracy of ex-post analyses and ex-ante predictions carried out by artificial intelligence. It’s likely that over time, my current role as a data analyst will be fully replaced by computers. But before this happens, I’m sure I’ll work on many more interesting assignments – maybe even for you?

Share on