At Think we spend a lot of time working with, analysing and building insights from airport and ATM data sets. However, one of the challenges we are frequently faced with is getting hold of the data from airports in the first instance. Despite spending vast sums on new airfield and terminal systems when it comes to making use of the data these systems produce beyond the scope of each system itself, we commonly see a lack of joined-up thinking. This has prompted us to ask ”do you have a Big Data strategy for your airport?”
It sounds complicated. You may have a multitude of systems, built by multiple suppliers and which have each evolved over time. This sounds like it is going to take a lot of time to un-pick.
Big Data Strategy Starter-for-Ten
Well, why don’t we try a top-down approach and start simple? As a starter-for-ten, how about:
- Record everything
- Store everything
- Share it with everyone
Ok, maybe a bit too Silicon Valley? This needs a bit of refinement.
Big Data Practically
Are you really going to record and store everything? Every update, every instruction, every change? Practically, maybe not, but this is a strategy so let’s stick with it for now. What data are you sure you don’t want – now or in the future? Can you second guess what data will be useful to you in ten years’ time? If there were no cost at all to storing this data would it matter? So I’m going to leave that one as is.
Where and how to store everything is trickier. Where does this data go? Raw text files copied to a hard disc somewhere? It’s going to be easier if this all goes to one place which is easy to access and, of course, secure. Somewhere where, in principle, anyone (we’ll get to this later) with web access could get to it if they need to.
Once it’s there, is the data consistent? Each system you have might speak its own language, especially if no one told the developers what to record. How do you know that vehicle “x” in one output is the same thing as vehicle “a” in another, or that passenger “y” was onboard? Why did one system record time as a text field, one as a number and does anyone even know what that field is meant to be…? There needs to be some common schema that links the data together. There are technologies for this as you would expect, though that’s something for another day. This gives us:
- Record everything consistently
- Store it securely (and somewhere easy to access)
- Share it with everyone
Big Data for Everyone?
Everyone? Well, no, that’s unlikely to work either. Some data will be too sensitive to share and should be locked away for reasons of commercial confidence, legality or security. After that, who is going to benefit from accessing your data? To some extent, maybe you don’t know until you give it to them. All your staff? Seems a good place to start. Your partners. Yes, why not? Your customers? Getting harder, they’ll have their own angle but can you both improve your operations as a result? Do they always ask you the same questions? Why not give them the capability to answer it themselves? Your competitors? Maybe if they pay enough!
Whatever you choose, if you have done “1” and “2” then it’s just a matter of (secure) access. The data is easy to get at, it is consistent and it will be cheap to manage and change the access. This leaves us with:
- Record everything consistently
- Store it securely (and somewhere easy to access)
- Share what you can with everyone you want
Think Big!
Strategies are easy to write of course. Implementation is a different question, and ultimately there’s no way to avoid some hard work. I just wanted to look at this from a different perspective. The toolkits to achieve this are there, often open source, and continually increase in power. The skillsets to help you analyse and visualise the data are a core capability of our team at Think. Soon enough you’ll only be limited by the vision of what you want to achieve, so why not Think Big?
Author: Michael Cowham, ATM Consultant
Recent Comments