Today’s technological world is highly driven by topics such as big data, AI, blockchain and machine learning. However, and interesting enough, when working with airlines all over the world we recognise quite often the challenges airlines are facing – especially in an airline operations context – are much less buzzword-driven and more down to earth.
More precise: Within the context of airline operations, many carriers (probably except the large ones) are still trapped in an outdated technological world. Old-fashioned databases, Microsoft Excel and complex IT landscapes are still very common among airlines. That’s why implementing AI and the other mentioned technologies feels like proceeding to the second step before taking the first.
I share that opinion. From my perspective you should learn to walk before you learn to fly. Subsequently, the first step in this direction is getting rid of outdated databases and creating a proper data warehouse — which can later serve as a base for the introduction of new technologies.
To get that straight: After having worked with several airlines all over the world, one thing I can clearly state is that airlines which rely on an operational data warehouse definitely move at higher pace when it comes to data analytics, performance reporting and development of new IT solutions.
Actually it’s not surprising. When doing a bit of research one will quickly find the main advantages of a data warehouse. For example, Data Warehouse Information Center lists the following benefits:
- Delivers enhanced business intelligence
- Saves times through standardisation
- Enhances data quality and consistency
- Provides competitive advantage
- Improves the decision-making process
- Enables organizations to forecast with confidence
- Streamlines the flow of information
That brings me to the point of what is essential when planning to introduce a data warehouse for airline operations data? Or in other words, what are the must-haves a data warehouse should offer. Some of them apply to other industries too, but some of them are tightly linked to the specific needs and requirements of airlines.
And by the way: As you’ll certainly assume, this is not a complete a list of relevant aspects (neither technically nor business-wise) which have to be considered when starting the journey towards a data warehouse 🙂
Essential 1: Real-time counts
Airline operations is a real-time business and is mission-critical. And for that reason a data warehouse should not only offer the possibility to access historical data, but also act as a real-time data source. Certainly, this is not important when it comes to analytics — but it is of eminent importance when it comes to real-time reporting, real-time KPIs and real-time alerting. And those topics receive an increasing attention. So you better be prepared.
Essential 2: It’s all about the quality
The success of a data warehouse stands and falls on the data quality. Therefore, a data warehouse has to be much more than a lake all data is dumped into. Quite contrary, a professional data warehouse has to offer specific technical layers for improving the quality and cleansing of incoming data.
Especially at airlines – where data often comes from different sources (e. g. timestamps are received from aircraft, ground handler or disposition systems) – a data warehouse must provide mechanisms to identify the “best” data. Those cleansing layers can be based on so called master matrixes. Another approach is to establish an XML-based rule-driven layer, where specific cleansing rules are implemented to control incoming data.
Essential 3: Reach for the clouds
Despite having the credo that business drives technology, there’s one technological aspect which is essential to consider. By all means, the chosen warehouse technology has to be “cloud-ready”. That offers two huge advantages: 1) You will be able to apply cloud technology to your warehouse. Especially when it comes to re-calculating values or loading additional historical data, cloud technologies can massively speed up the processes. 2) Only by applying cloud technologies you’ll be able to make use of all those fancy stuff like AI, machine learning, etc. in a later stage.
Essential 4: Sharing means caring
Since it is all about integration a warehouse should offer possibilities to easily and swiftly connect to external systems. This, on the one side, is important in order to enable your analytic teams to access all contained data with state-of-the-art tools, like Tableau or PowerBI. And on the other side, warehouses should also provide the possibility to other system to access the contained data via a standardised API. Never ever go for a solution that doesn’t offer those possibilities.
It is NOT about technology — in a first step
Don’t get me wrong: Of course, technology matters. You’ll have to make decisions whether to go for mariaDB or Oracle, where to use Hadoop cluster and identify how tools like Kafka, Drools, Nifi or Talend can help you. But those aspects should never be answered in a first step.
Collect the business requirements. Analyse your current system landscape. Assess the goals you want to achieve. Pay attention to the limitations you have. And once that has all been done, you can start and assess which technology fits best to your results. Sounds trivial I know — but believe me: I’ve seen it to often going the other way around.
A data warehouse is essential and is key to generating competitive advantages. When implementing a data warehouse for your airline operations you should specifically pay attention to these four aspects 1) real-time, 2) data cleansing, 3) cloud-readiness and 4) integration.
Benjamin Walther, CEO, Information Design.
Get in touch with Benjamin directly: firstname.lastname@example.org or connect with him on LinkedIn