Data science in a nutshell
It is important to note the difference between traditional Business Intelligence (BI) and data science. BI is concerned with aggregating and analysing past data and visualising it. Data science is more focused on modelling past data in order to predict future trends. BI is therefore focused on descriptive analytics whereas data science is focused on predictive or prescriptive analytics.
There are also tools in the data science toolbox for advanced descriptive analytics. This comes to light in how a root cause analysis is performed when using BI or data science. The BI analyst will for example have to hypothesize about the root cause, and then plot or aggregate the applicable variables to determine whether they are in fact the correct indicators. For example, when trying to determine why orders are late, the BI analyst would plot indicators against the label of the order being late or on time. The trends may then show which indicators are positively correlated with the order status.
A data scientist would rather categorise orders by their status, compile a model that describes the data, and then look at the feature importance of the model variables to try and determine the leading indicators that show what causes an order to be late.
Value for airport operations
IoT, big data and data science will provide increased situational awareness and improved flow for airport operations. Imagine if you can update flight arrival and departure times in near real time by calculating the knock-on effect of early morning delays due to, for example, the impact of severe weather patterns and delays in other feeding airports. The benefits include improved traveller satisfaction, enhanced planning accuracy for apron operations, enhanced planning accuracy for ground and air crew scheduling and many more.
Predicted delays can be communicated to airlines, travellers, airport shuttle services and other airport logistics providers such as air catering, baggage handling, fuel, fire management and any other appropriate stake holders.
Safety and security within the airport building will also be enhanced due to reduced traveller congestions when flights are delayed, and these delays could have been predicted and communicated proactively.
Other benefits include:
- Predictability of billing for airport services, leading to improved revenue assurance.
- Enhanced visibility and traceability of regulatory compliance.
- Prescriptive maintenance of airport building, infrastructure and equipment.
- Improved planning accuracy for procurement for airport retail and apron stock like fuel.
- Improved planning accuracy for special apron services such as de-icing and tying down of small aircraft.
- Improved planning accuracy for apron logistics including equipment and crew.
- Improved planning accuracy for staff logistics
The success of any big data and data science endeavour depends heavily on data quality. It is critically important to ensure that data quality has strategic focus and support. An important reason Data Science, ERP, Robotic Process Automation (RPA) and Digital Transformation projects fail to deliver value is the lack of high-quality data. The sheer granularity of the digital world means that any errors in data are amplified exponentially and they lead to vast deviations from the truth. Think of the effect of a two-degree error in direction if you are travelling 100 metres vs. 10 kilometres vs.10,000 kilometres vs. 10 million kilometres.