You are currently viewing Before artificial intelligence, data can predict situations

Before artificial intelligence, data can predict situations

I have been hearing for several months about artificial intelligence in all directions. Supporting my imagination, images from the films Blade Runner, I-Robot or IA go to my head and I tell myself that in 2019, when I go to the Post Office to pick up my parcels, I will come across a bionic entity created by man and which follows the laws of Asimov but which, nevertheless out of politeness, would welcome me.

To get there, there is still a long way to go and even with Moore’s law, it won’t be tomorrow that my robot will give me my package. In the meantime, we are educating bots, we are trying to set up semi-autonomous systems, personal assistants who know how to speak or search for answers on the internet, and many more. In short, before intelligence, we are not even at the stage of artificial consciousness….

For my part, not being passionate about this aspect of artificial intelligence, I have been working for some time on business intelligence at the heart of tourism systems. The tourist world answers 80% of the same questions every day…. For the tourist offices, what is the weather like today (looking at the sky, you can tell….), where are the nearest toilets? For socio-professionals, are there people on the territory? should I plan to take out the umbrellas and cutlery in addition? should I open more or less tracks? For public services, how full will my car parks be today? What will be the density of the access roads to my destination?

So many questions that the data of a territory can easily answer once logged and archived within a data lake.

Imagine a screen connected to the heart of a densely populated area. Displaying the nearest toilets on demand is simply a query between the base of the public toilets on their GPS coordinates and the point where the screen is located. The weather ditto, why display a 7-day weather forecast when 95% of the demand concerns the day’s weather?

Even further, the famous “What to do today, it’s raining?” comes down to a crossing of current meteorological data and the base of tourist activities tagged with a “rain” criterion on its activities. all that remains is to display it.

But what is predictive in all this? Nothing at all.

Predictive systems are based on the history of the ecosystem in which they evolve. The greater the history, the greater the chance of finding points of convergence in the situations sought. I’ll give you a concrete example… In the context of a ski resort (of course I know the subject quite deeply), the access controls to the ski lifts give a very precise number of daily passages, at the device near. Let’s cross this data with the meteorological readings over the same period and add an additional criterion which is the rhythm of school holidays in Europe. By analyzing the current day with the criteria mentioned above, Business Intelligence should be able to predict the expected audience on the slopes and….their staff to be provided at the cash desk to welcome them.

If we are lucky enough to have an audience calculation for the entire territory, this request can be similar at the scale of the destination, apart from the event criterion injecting a one-time population surplus that is impossible to predict in this ecosystem.

In summary, the basic predictive system can be summed up in a simple recipe:


With all of this historical data, the ecosystem can evolve on its own. Everyone wants to add personalized advice following the daily analysis as mentioned above.

With the proliferation of access controls at activity providers, they can be inserted into the ecosystem to also define a typical customer journey according to seasonal and meteorological criteria, to be specified according to the calendar and adjuvant events.

In summary, before saying that the future is in artificial intelligence, we should go step by step. Not everyone is Tony Stark and Jarvis only exists in the movies for now. Let’s use our data to optimize our businesses and to allow us to estimate more precisely what our customers want to do. It remains for us to adapt to provide for it.

Leave a Reply