Puck.js to GCP BigQuery & Data Studio
This tutorial shows you how use Puck.js and a Raspberry Pi to stream data to GCP.
The GCP Tutorial (Temperature data to GCP BigQuery Table & Data Studio Graph):
- The puck.js app advertises the temperature (and other info) to the Raspberry Pi EspruinoHub MQTT broker
- NodeRed ingests the MQTT message & cleans it up with a function
- NodeRed sends temperature data to a GCP Pub/Sub topic via node-red-contrib-google-cloud pub/sub out node with credentials
- A GCP Pub/Sub topic accepts the message & relays it to the subscribers
- A GCP Pub/sub subscription stores the data to a BigQuery Table
- A Data Studio dashboard displays the data from BigQuery
Prerequesites
- A Puck.js device (or other Espruino board if you don't need Bluetooth LE
- Raspberry Pi with Bluettoth/Wifi Capabilities and NodeRed & EspruinoHub installed on it
- Install Espruino Hub & NodeRed on the Raspberry Pi
- Set EspruinoHub & NodeRed to auto start on a reboot
- A GCP Account (This is free to set up & this project can be done in a free tier)
Raspberry Pi Setup
Note: This tutorial assumes you have EspruinoHub & NodeRed installed on your Raspberry Pi already!
Edit the Raspberry Pi so EspruinoHub will ingest the MQTT messages and NodeRed will publish to the right GCP project
- SSH into the Raspberry Pi and edit the attribute.js file in the EspruinoHub. This code will create a filterable channel on the MQTT broker.
Go into the attributes.js file
cd
cd EspruinoHub/lib
nano attributes.js
- Edit the exports.names dictionary in the attributes.js file, adding the code between the "Add These" Comments.
- This adds an MQTT channel for Voltage, Light, Acceleration, Gyro, Movement, Magnetic Field info, Button Presses, LED states, NFC info and bluetooth info. The Puck will advertise info to the Raspberry Pi on these channels. Only one of these channels will be used for this project, but adding these in allows us to advertise the other sensor states in a later builds.
exports.names = {
// https://www.bluetooth.com/specifications/gatt/services/
"1801": "Generic Attribute",
"1809": "Temperature",
"180a": "Device Information",
"180f": "Battery Service",
//////////////////////////////
// Add These /////////////////
//////////////////////////////
"182a": "VoltageSensor",
"182b": "LightSensor",
"182c": "AccelerationSensor",
"182d": "GyroscopicSensor",
"182e": "MovementSensor",
"182f": "TiltDetection",
"183a": "MagneticField",
"183b": "MagneticFieldData",
"183c": "ButtonPress",
"183d": "LEDState",
"183e": "NFCDetection",
"183f": "NFCData",
"184a": "BluetoothScanData",
"184b": "BluetoothScanState",
//////////////////////////////
//////////////////////////////
//////////////////////////////
// https://github.com/atc1441/ATC_MiThermometer#advertising-format-of-the-custom-fir>
"181a": "ATC_MiThermometer",
"181b": "Body Composition",
"181c": "User Data",
"181d": "Weight Scale",
// https://www.bluetooth.com/specifications/gatt/characteristics/
"2a2b": "Current Time",
"2a6d": "Pressure",
"2a6e": "Temperature",
"2a6f": "Humidity",
"2af2": "Energy",
// https://www.bluetooth.com/specifications/assigned-numbers/16-bit-uuids-for-member>
"fe0f": "Philips",
"fe95": "Xiaomi",
"fe9f": "Google",
"feaa": "Google Eddystone",
"6e400001b5a3f393e0a9e50e24dcca9e": "nus",
"6e400002b5a3f393e0a9e50e24dcca9e": "nus_tx",
"6e400003b5a3f393e0a9e50e24dcca9e": "nus_rx"
};
- Open the NodeRed service file
cd
cd /.
cd lib/systemd/system
sudo nano nodered.service
- Add a GOOGLE_CLOUD_PROJECT name as an environment variable to the NodeRed Service file.
- This will allow you to publish to the GCP topic you're about to set up.
Environment="GOOGLE_CLOUD_PROJECT=mygooglecloudproject-123456"
- Reload the systemctl daemon
systemctl daemon-reload
Flash the Code to Puck.js
- What the Code Does
- As long as the button state remains 1, every minute, your Puck.js will advertise the temperature (f), battery level, voltage, light, and LED states. It also will advertise when it senses movement, a change in magnetic field, NFC field, or a button press. For this tutorial, we'll just be using the temperature advertisements. Upload the code to the Puck.js's flash, so it'll still be there if you swap out the battery.
- Testing the Code in Espruino Web IDE
- Check that the code is running by reading the console log output. Make surethat temperature data is coming across. Sometimes temperature data doesn't advertise if the temperature hasn't changed. If you want to test the temperature advertisement, warm up or cool down the puck temp sensor by transferring heat with your hand.
- Start Advertising!
- To get the data to stop showing up in the console & start advertising, take the battey out & put it back in. As long as the code was uploaded to the device's flash, it'll start advertising when the battery is put back in. The temperature data won't start sending over until one minute has gone by. Then it'll advertise the temp every minute.
var state =1;
NRF.setTxPower(4); // Full Power advertising
//Send MQTT Advertisements & Console Logs Every Minute while Button State === 1
setInterval(function () {
setTimeout(function(){
if(state === 1){
// Sent: 79
// - /ble/advertise/c3:5a:61:d8:02:05/temp
// - "1809": "Temperature",
// 1809 => {"temp":79}
console.log("temperature : " + [Math.round((E.getTemperature()*9/5)+32)]);
NRF.setAdvertising({
0x1809 : [Math.round((E.getTemperature()*9/5)+32)]
});
}
},1);
setTimeout(function(){
if(state === 1){
// Sent: 100
// - /ble/advertise/c3:5a:61:d8:02:05/battery
// - "180f": "Battery Service",
// 180f => {"battery":100}
console.log("battery : " + [E.getBattery()]);
NRF.setAdvertising({
0x180F : [E.getBattery()]
});
}
},5000);
setTimeout(function(){
if(state === 1){
// "182a": "VoltageSensor",
console.log("Voltage : " + [String(Math.round(analogRead(D30)*6.6 * 100) / 100)]);
NRF.setAdvertising({
0x182a : [String(Math.round(analogRead(D30)*6.6 * 100) / 100)]
});
}
},10000);
setTimeout(function(){
if(state === 1){
console.log("Light : " + [String(Puck.light())]);
NRF.setAdvertising({
// "182b": "LightSensor",
0x182b : [String(Puck.light())]
});
}
},15000);
setTimeout(function(){
if(state === 1){
console.log("LEDs: " + digitalRead(LED1).toString(16)], [digitalRead(LED2).toString(16)], [digitalRead(LED2).toString(16));
NRF.setAdvertising({
// "183d": "LEDState",
0x183d : digitalRead(LED1).toString(16)], [digitalRead(LED2).toString(16)], [digitalRead(LED2).toString(16)
});
}
},20000);
},60000);
//Button Press
//Turn Off/On MQTT Advertising
var pressCount = 0;
setWatch(function() {
pressCount++;
state = (pressCount+1)%2;
if ((pressCount+1)%2) digitalPulse(LED3,1,1500); //long flash blue light
else
digitalPulse(LED3,1,100); //short flash blue light
console.log('button_press_count : [' + pressCount + ']');
console.log('button_state : [' + (pressCount+1) + ']');
console.log('state: ' + state);
NRF.setAdvertising({
0xFFFF : [pressCount],
0x183c: [((pressCount+1)%2)],
});
}, BTN, { edge:"rising", repeat:true, debounce:50 });
//Movement Sensor
require("puckjsv2-accel-movement").on();
var idleTimeout;
Puck.on('accel',function(a) {
digitalWrite(LED1,1); //turn on red light
if (idleTimeout) clearTimeout(idleTimeout);
else
if (state === 1) {
console.log('movement : 1');
NRF.setAdvertising({
0x182e: [1],
});
}
idleTimeout = setTimeout(function() {
idleTimeout = undefined;
digitalWrite(LED1,0);//turn off red light
if (state === 1) {
console.log('movement : 0');
NRF.setAdvertising({
0x182e: [0],
});
}
},500);
});
//Magnetic Field Sensor
require("puckjsv2-mag-level").on();
Puck.on('field',function(m) {
digitalPulse(LED2, 1, 200);//flash green light
if (state === 1) {
console.log('magnetic_field : [' + m.state + ']');
NRF.setAdvertising({
0x183a: [m.state],
});
}
});
//NFC Detection
NRF.nfcURL("http://espruino.com");
NRF.on('NFCon', function() {
digitalPulse(LED2, 1, 500);//flash on green light
console.log('nfc_field : [1]');
NRF.setAdvertising({
0x183e: [1],
});
});
NRF.on('NFCoff', function() {
digitalPulse(LED2, 1, 200);//flash on green light
console.log('nfc_field : [0]');
NRF.setAdvertising({
0x183e: [0],
});
});
Setting up GCP Pub/Sub & BigQuery
- Go to your GCP Console
- Create a Pub/Sub Schema for to Ingest Temperature Data
{
"type": "record",
"name": "Avro",
"fields": [
{
"name": "temperature_fahrenheit",
"type": "int"
}
]
}
- Create a Pub/Sub Topic
- Create a Dataset & Table in BigQuery
- Create a Pub/Sub Subscription that writes to the BigQuery Table
Setting up NodeRed
- Add a MQTT In Node that listens for the temperature data
- Add a function node that processes the message payload
- Install "node-red-contrib-google-cloud" to your nodered palette
- Add a pub/sub out node that publishes to your GCP Pub/sub topic
- Edit the pub/sub out node to add GCP credentials from IAM. This will allow you to publish to that topic
BigQuery & Data Studio
- Open up the table in BigQuery, Click Export & Explore in Data Studio
- Explore the data set & make a line graph or timeseries graph displaying temperature over time
Buying
Puck.js devices can be ordered from here
This page is auto-generated from GitHub. If you see any mistakes or have suggestions, please let us know.