Simple AI exercise timer step-by-step video

To create your simple AI exercise timer, 
you’ll train a machine learning (or ML) model. 

This model will recognise when you’re 
exercising and when you’re not exercising.

You’ll then combine the model with some 
ready-made code for an exercise timer...

before downloading it to your 
micro:bit and using it in real life.

Click ‘Open in micro:bit CreateAI’ to launch the project. 

The project comes with 3 samples 
of movement data for exercising

and 3 samples of movement 
data for not exercising.

You’ll need to add more samples by 
recording your own movement data.

micro:bit CreateAI collects movement 
data samples using the accelerometer  

(or movement sensor) on the micro:bit.
You will wear a micro:bit and battery  

pack on your wrist or ankle, so that you can move 
freely to record your own movement data samples. 

To get started, you need to set 
up the data collection micro:bit. 

Connect the wrist-worn micro:bit to CreateAI.
If your computer has Bluetooth enabled then you  

will just need 1 micro:bit and a USB data lead.
If you don’t have a Bluetooth connection,  

you’ll be prompted to use 2 micro:bits.
The second micro:bit will remain connected  

to the USB cable and act as a radio 
link to the data collection micro:bit.

Follow the instructions on screen to connect.

Once your data collection micro:bit is connected

you’ll see the lines on the live graph 
change as you move your micro:bit about.

You’re now ready to add your 
own movement data samples.

As this project already includes some 
data samples, we suggest you just add  

1 more sample for each action for now, and spend 
more time collecting and analysing data later.

Decide what ‘exercising’ 
action you are going to do. 

This could be running, walking briskly, 
jumping, boxing, dancing, or any other exercise. 

Make sure the micro:bit is attached to 
the wrist or ankle that will be moving.

To add data to a specific action, 
select it by clicking on it. 

You will get a 3 second countdown 
before a 1 second recording starts. 

Click record and start moving right away 
to ensure you get a clean data sample. 

A clean sample is one where you 
are moving for the entire sample,  

you don’t start late or finish moving early.
Next try adding an extra data sample  

to the ‘not exercising’ data set.
Select it by clicking on the action,  

then stay still, or only move very 
slightly as you record the sample. 

You’ll notice that the x,y,z 
lines change places depending  

on the angle at which you hold your micro:bit.

The project doesn’t have a lot of data right now, 

but we have enough to train our own 
machine learning model using CreateAI. 

So click ‘Train’ to use the 
current data to build an ML model.

The tool now builds a mathematical 
model that should recognise different  

actions when you move your micro:bit.
As soon as the model has been trained,  

you’ll see the Testing model page.
Your data collection micro:bit can now  

be used to test how well the model is working.
It should still be connected to the tool,  

and you’ll see that as you move it, CreateAI 
is estimating what action you are doing.

Try out different levels of exercising or 
not exercising to see both the estimated  

action and the certainty bar graph change.
The % on the certainty bar graph shows how  

confident the model is that 
you are doing each action.

You may notice your model is not estimating 
some actions accurately, or maybe it is  

working well for one action but not the other,
so after exploring how it is currently working,  

it’s a good idea to click on ‘Edit 
data samples’ and improve your model.

Machine learning models usually work best with 
more data, so record some extra samples for each  

of the actions, or focus on collecting more data 
for the action that was problematic in testing.

You can record 1 sample at a time or 
you can record 10 samples in sequence.

Clean data samples also 
help an ML model work better 

so interrogate your data set and identify any 
data samples that could confuse the model. 

You can delete these by pressing x.
Once you’ve added more data and checked  

your data set, click Train model 
again to use your amended data set. 

Then test the model again on 
the ‘Testing model’ page.

Once you’re happy with how 
the ML model is behaving,  

you can use it with the ready-made project code.

Click on 'Edit in MakeCode' to see the code blocks 
in a special version of Microsoft MakeCode.

You can always return to CreateAI using 
the arrow in the top left of the screen.

These code blocks use the model you 
have created within an exercise timer.

The code uses two variables to keep track 
of how long you've been exercising and how  

long you've not been exercising.
When the program first runs it sets  

these timer variables to 0.
The 'on ML start' blocks are  

triggered when the ML model decides you have 
started either exercising or not exercising. 

They show different icons on the 
micro:bit's LED display depending  

on the action it has estimated you are doing.
The 'on ML stop' blocks are triggered when the  

ML model decides you have finished an action, 
in this case exercising or not exercising. 

Code inside each block clears the screen and 
adds the duration of the action that has just  

finished to the variable storing 
the total times for each action.

The ML model works with the code to allow you 
to view the total time spent on each action. 

Press button A to see the total time you have 
been exercising and press button B to see the  

total time you have been inactive.
The timer counts in milliseconds,  

thousandths of a second, so the number shown 
is divided by 1000 to show a time in seconds.

To make your simple AI exercise 
timer run on your micro:bit, you  

just need to download this code to a micro:bit.
If you don’t have another micro:bit available,  

simply replace the code currently on the data 
collection micro:bit with the project code.

Now test the project out in real life.
Do the correct icons display  

when you are exercising or not?
You can test if the timer code is working  

well with the model in 3 easy steps:
Press the reset button.

Exercise for 30 seconds.
Then press button A.

You should see the number 30 
scroll across your display.

You’re now ready to connect to CreateAI, 
collect your own data, use it to train,  

test & improve a machine learning model.
And you can then combine this model with  

the ready-made code and try 
it out on your own micro:bit.