Dirty Dozen of Data in Aircraft Maintenance

Data created by aircraft and made available in systems is increasing steadily and aircraft maintenance engineering processes face new challenges such as data integrity. Simultaneously, these processes become more dependable on data; Airworthiness compliance, EFB, ETL, analytical software tools, fuel efficiency tools, you name it. As mentioned in our blog about ‘Human Interaction with IT systems in Aircraft Maintenance’ due to continued digitization in the aviation industry, airlines and in specific Engineering & Maintenance departments are faced with an even more complex and diversified digital landscape than ever before and data is just part of almost every action taken.

Everyone working in aircraft maintenance is encountering the ‘Dirty Dozen of human factors’. They are an important part of the training for everyone involved in aircraft maintenance processes. So, as aircraft maintenance processes become more digitized and the amounts of data processed increase almost daily, it is necessary to create more awareness for digital and data related needs and introduce an additional line of dirty dozen – The dirty dozen of data in aircraft maintenance:

1.Lack of data:

Whether you conducting a data analysis, migrating data, or anything else: It of utmost importance to have as much data available as possible. A lack in data means that the output is not reliable as you do not have all the means together to conduct the analysis, or migration.

Utilize normalized database systems as much as possible to support and manage your business processes.

2.Ignoring data quality:

Having large amounts of data available but the quality is lacking, the output is flawed. I mean everyone knows the world-famous rubbish in is rubbish out, that’s especially true for data.

Configure the usage of your data sources systems in such a way that users can only enter the right information and ensure as less human input is required as possible.

3.Not educating the business on the value of data:

Ignoring what added value of data driven decision making and importance of data: The statement: ‘we have done it always this way’ is probably not an uncommon one, but also the maintenance and engineering department must adjust to the more digitized aviation world to stay compatible.

Creating awareness and providing real-life evidence business cases is the way to convince people, data = facts; and facts cannot be ignored.

4.Insufficient skills to work on data:

For example, data analytics, making a pie chart in excel is not difficult, but managing a data warehouse and presenting information in a logical cognitive matter is a field of its own.

To get the right output which is of added value for the company make sure to work with skilled staff who know how to deal with complex amounts of data, identify the right data sets that need to be analyzed and how to best present this.

5.Expecting a single person to be the perfect data scientist:

Next to the above given phenomenon it is also happening that one person should deal with everything. As mentioned in the introduction, the amount of data is increasing steadily and only having one person available who must deal with all of it is tricky as a regular working day has around 8h, but data is created 24/7.

To ensure all data is handled properly ensure that key-users are engaged, properly trained and empowered to work with data. Knowing the processes in detail and can work with data makes for a killer combination. It needs to become part of day-to-day work of all people in the company.

6.No testing:

Not validating models and predictions, statistics can be skewed in any way to make a theory seem plausible.

Always have a counter sample group and test models on multiple different sets of data prior to assuming your analytics shows you the truth.

7.Not separating the signal from the noise: 

One situation can trigger many indications and it should be identified which of the indications are relevant to the source issue.

A common used methodology is noise reduction which allows to identify patterns between occurrences in data (signals) and determine if they are related. This becomes very relevant once analyzing Aircraft Health Monitoring Data and other on-board collectible data of aircraft.

8.Ignoring the importance of IT support:

Using different data software tools means that not only the skills of the people using the software tools are of importance but the IT support needs to be splendid as well. Nothing worse than a server down in a digitized operation.

Make sure that if you are buying a software tool that you check the support criteria of the seller and involve your internal IT department in the implementation of the software to avoid unpleasant surprises and an interruption of the software during operational usage. Or even better, leave it to the experts to run your system.

9.Lack of integration:

Having all kinds of software tools in place does not mean that automatically benefits will be reaped. It’s the collective of all, or the sum of the individual being greater that triggers the benefits of a digitized operation. This means integration of your system.

Use integration layers between your different software solutions and tools and make sure they are open to a controlled flow of data being imported and exported out of their databases. Make sure to clearly define master data systems and systems drawing their information from the master systems to prevent lack of integrity.

10.Trying to solve everything at once:

Having the right people in place and using the right tools does not mean that everything can be solved at once, does not matter which data related project you are conducting.

As for every other project, set goals and objectives, asking what is of importance for the department to focus on the most significant issues first. In that way, everyone has the same goal in mind and work is done accordingly and no one gets lost in the huge amounts of data.

11.Time pressure:

Also in the dirty dozen of data, time is an important factor. Today, still most data entries are still done mainly by us, so the human error is still in place. Combining this with time pressure increased the risks of entering wrong data into the systems.

Continuously strive to reduce human input into your systems. Automate what can be automated.

12.Not adapting current processes:

It is not enough to buy e.g. a data analytics tool or new MRO software. These new software solutions also have an impact on existing processes and procedures, they need to be adapted so that the maintenance and engineering department get most value out of it.  

It is of utmost importance that the current process area analysed and compared with the ‘needs’ of the new software and adapt the organization accordingly.