The data stream, fuel for your digital projects

In local authorities, digital projects most often start with a preliminary phase of capturing various information through sensors (Internet of Things, IoT), followed by storing it in a "data lake" of more or less great depth.

This approach, quite natural, runs into reality. On one hand, data is a perishable good. On the other, once it's plunged into the depths of the lake, it decays there, and it's no longer possible to recover its energy to feed business scenarios. It's not for nothing that mills have historically been installed upstream, before a river flows into a lake, to harvest all its value in real time and convert it into energy.

Bad news: data is perishable

More precisely, it's the practical usefulness of data that fades quickly over time. Take the simple example of an electric vehicle charging station: similar use cases abound, parking spaces, metering water, energy or footfall, air quality, temperature.

Over the medium term, information about our charging station (in service or out of order, free or occupied, energy consumed) and its context (location, identifiers, timestamp) is useful for deducing trends: usage rate, day/night profile, seasonality, service availability, average probability of finding a free station. This processed information helps size investments and measure the impact of the service delivered to users.

Detailed, minute-by-minute information, on the other hand, loses all usefulness over time, unless it's used on the spot, for instance to tell electric vehicle owners where a free charging station is right when they need one.

For those who enjoy a rough back-of-the-envelope calculation: suppose the different states of our charging station are read every minute, and each reading is a 1 kilobyte record. Over a year, that's 365 × 24 × 60 × 1 kB, or around 500 MB per station per year. Counting duplication and archiving, the final volume will exceed 1 GB per station per year. With the 2030 target well beyond a million stations for France alone, we'll be talking in petabytes.

Good news: data in motion has intrinsic energy

Like light, data has no mass, but that doesn't stop it having its own energy, quickly dissipated over time. In the example above, instant knowledge of a station's status has value for whoever needs to charge their vehicle at that exact moment.

In industry, immediate information about a machine's settings helps avoid producing a defective batch. In another context, the instant profile of a home's water consumption gives a useful signal about the activity of an isolated person. For a more topical case, direct knowledge of the CO2 level in a public place (school, auditorium, restaurant) lets the right person, a teacher, a technical director, a venue manager, be warned, and why not directly control a ventilation system, provided that information is available in real time, someone knows what to do with it, and the means to act exist.

Data's energy: a source of value for digital projects

Turning data into useful action within digital projects takes a few prerequisites:

• obtain the data, whatever its source: connected objects, applications, web services (weather, traffic, air quality), not forgetting people themselves;
• correlate it in real time, before its energy dissipates, and draw practical decisions from it, applicable immediately: control ventilation, adjust a building management system, open or close vents, change lighting, through scenarios that are bespoke and simple for the responsible staff to define;
• communicate with and involve teams, contractors, partners and residents, ideally without forcing new tools on them.

Converting the kinetic energy of a stream, before it dissipates, into useful work: that's the mill's job. Lakes offer rest and eternal oblivion. Don't local authorities need the former more than the latter?