LiamOSM

I think the easiest way to accomplish what you're describing would be to replace the Arduino Nano that I used with an ESP32 or ESP8266. The ESP microcontrollers are compatible with the Arduino IDE so they're pretty straightforward to work with. They also have a built in WiFi adapter. You could have the ESP send an email when the water quota is exceeded, or you could have them constantly logging the water usage to a database (SQL if you have a domain already). The email approach would be simple, but wouldn't let you access as much data. The database approach is more complex but would let you perform analysis on the raw data.

I actually live fairly close to you then, in Burlington. I didn't factor in the flat rate monthly service charges for metering; I only used the cost per cubic metre. I obtained the values from here: https://www.halton.ca/getmedia/0827acbc-dce9-4933-8bdb-7a5de3ae3199/PW-water-rate-card-2019.pdf.aspx?ext=.pdf

The charging jack (which is normally unplugged during use) is on the bottom, so water will not flow up into it. The flow sensor jack is on the side, but when the flow sensor is plugged in, the plug blocks water from entering the jack. Also, I've found that these connectors are sealed pretty well so even if water did enter the jack, it would make its way into the enclosure.

If you wanted more precision, you could use a different flow meter. However, the one you linked to appears to use the same 4.5 factor, which suggests it's just the same flow meter that I used, in a different enclosure.

The 3 control lines are used to send data to the shift register. Then the shift register actually does the multiplexing using its 8 outputs.

A hall-effect sensor would be a big improvement over a reed switch. Once you overcame the limit of having a mechanical switch, it could theoretically accelerate until the air resistance (drag) equals the force applied by the electromagnet, causing a net force of zero and therefore no acceleration. That is, assuming there wasn't any other bottleneck component like the speed of the transistor.

I wonder if this would be able to erase the RFID tags found on clothing in stores. If the tag was erased, I wonder if it would still trigger an alarm when passing through the sensor gates at the doors.Just a theoretical question of course.

This Instructable is great! The first gears I modelled used the typical circle-shaped teeth and the meshing was absolutely horrific. I also love the resources you linked to. Great work!