How does Flow work? Follow
Flow: a handheld revolution, and a milestone in the evolution of air quality monitoring.
What’s so special about it? It’s the first environmental sensor of its kind to be this small, this portable, and this personal. Air quality monitoring usually involves many devices, many times the size of Flow, never before designed for lives on the move. After 4 years of research and development, we’ve achieved a true leap forward in miniaturisation. By shrinking a wide range of scientific equipment inside a sleek sleeve of polished steel small enough to hold in your hand, we’ve designed a device that will elegantly follow you wherever you need it to.
Want to learn more about what’s whirring and purring inside? So glad you asked: here’s a post to run you through how all the pieces come together.
Before we start, one notion to have in mind is that Flow is the combination of three things:
- Hardware: the device itself, and all the components tucked away inside its stylish Chamonix Grey casing.
- Firmware: programs loaded onto the hardware that power the sensors and churn out measures – the brains of your Flow, if you will.
- Software: the app installed on your phone, and gateway to all your data!
Step 1: air enters your Flow. You’ll notice the holes carved into the sleeve go all the way around – we call it the “360 air intake”. To make sure the air inside your Flow is exactly the same as outside, we fit in a fan to optimize air flow through your, well… Flow. It may be tiny at 5mm diameter, but it’ll hit 15,000 RPM like it’s nothing – more than an Italian sports motorcycle will ever do! And, if you listen closely, you’ll soothe your eardrum with its gentle whirr.
Step 2: the Particulate Matter sensor lights up. It shoots a laser beam at the air brought in by the fan; every time a particle is hit, light is dispersed – disco-ball style. This micro light show is detected by a photovoltaic cell that translates the laser’s deflected beams into electrical current we can measure. A word of caution: if the idea crosses your mind of opening up your Flow and pointing the laser into your eye, please don’t.
Step 3: the NO2 and VOC sensor does its thing. Think of it like a toaster. A tiny membrane is heated up to 350 degrees (!), and mercilessly disintegrates any NO2 or VOC molecules passing through. This lets us measure the variations in energy required to maintain the membrane’s temperature stable as it’s happily toasting away. Note that 350°C is a lot, so it requires some tricky battery life management on our end to make sure your Flow remains active all day. It also means that when Flow turns on, you can expect a 15-30min warmup before hitting toasting temperature.
Step 4: all sensors are calibrated. Over time, hardware components “drift”, a bit like musical instruments. If you’ve ever played guitar, you’ll know that as temperature and humidity vary, the wood expands and contracts, changing string tension by a tiny amount – effectively detuning your instrument by that much. Likewise if you live somewhere, say, windier, warmer, or drier than average, your Flow’s components will be changed in small ways. So, over time, each Flow is unique! That’s why we created calibrating algorithms to make sure all of these differences are accounted for, giving you spot-on readings all day long.
Step 5: the Neural Networks are activated. A fancy way of saying we’ve trained programs to detect patterns in the data they’re being fed. These patterns help determine levels of PM, NO2 and VOCs in the air pulled into your Flow. In case you were wondering, they’re called neural networks because they’re designed to learn in a way that mimics the human brain. Google started using them to detect cats in YouTube videos, but we’ve put them to work to detect what’s in the air you breathe!
Step 6: patterns are converted into measures. Particles Per Billion (PPB) for VOCs and NO2, micrograms per cubic meter for PM. And yes, that does mean your Flow analyses billions of particles every day. Luckily, counting particles isn’t nearly as sedative as counting sheep.
Step 7: Flow applies filtering for good measure. You’ll likely be carrying Flow around with you everywhere you go. Within a day, a week, a month, you’ll be exploring all kinds of places and environmental conditions. As a result, there may be unexpected and sporadic spikes in your Flow’s measures because of patterns it doesn’t yet recognise. Solution to the problem: we’ve built in a virtual data scientist to review your readings, real-time, and send misleading measures home faster than you can say “air quality index”.
Step 8: filtered measures are turned into an AQI. Particle counts and micrograms are all well and good, but what really matters is understanding their health impact. That’s what we’ve designed the Plume AQI for: an immediate overview of pollution levels, just as the temperature might give you a first indication of the weather outside.
Step 9: AQI receives the visual treatment on a screen near you. Open your Flow app, and… ta-da! There it all is, finally! Behind the scenes, your Flow will beam all the AQI data that was being stored on its flash drive to your smartphone as soon as it detects an active Bluetooth connection. The result is neatly displayed in a beautiful Montserrat font, and because your phone contains a GPS we can pair AQI data with your location to render it all inside colourful maps. The first step to the Revolution is a colourful map, or so we’ve been told.
Step 10: to the cloud! It’s the story we’ve been pitching from day 1: air pollution data is always better when it’s shared. So when your phone is connected to a WiFi or 3G/4G network, we’ll be safely storing it in our servers. This is how we’ll be able to start layering our users’ data on top of all the maps we’ve already built from public data. And that, my friends, will truly be the next leap forward in air quality monitoring!