A PulseSensor kit, a compatible board with an analog input, USB cable, and either the WebSerial tool or Arduino with PulseSensor Playground.
Entering a competition? Ask first.
Ask your teacher or adult sponsor before collecting data. Fair rules change, and projects involving people—including some self-testing—may require review before experimentation. Testing other people may require prior IRB approval and consent or assent. Never collect or publish participant names or identifiable health data.
Plug it in, open WebSerial, and rest a finger on the sensor. That wave is you, live, right now.
Now find the small bump partway down the falling edge of each beat. That is your aortic valve snapping shut. You are watching a valve inside your own chest close, on a $30 sensor, at your kitchen table.
Nothing to record. Nothing to submit. Just watch it for a minute.
Watch the wave, then run your hand under warm water and watch again. The size of the beat can change a lot. Nothing to write down.
You and somebody else
Take turns on the same sensor. Everyone’s wave has a slightly different shape. Find one thing that is different between the two of you.
Noticed something you cannot explain? Good. That is a question, and the projects below show you how to chase it.
Three starter projects
Begin with a result you can explain.
Each one is a single question you can answer in an afternoon. Do as much or as little of the record-keeping as you want—unless you are entering a fair, in which case see the box at the bottom.
Beginner1 sensor30–60 minutes
1. Make a clean PPG graph
Question: Which parts of the wave come back the same way, beat after beat?
Try thisNothing yet. Just get one recording you are happy with.
Keep the sameSame finger, same posture, same attachment, same gentle pressure, same lighting.
Watch forThe peaks, the slow drift of the baseline, and any stretch that is too messy to use.
Show someoneYour wave with one good section and one messy section marked, and why you can tell them apart.
You’ll finish with: one labeled waveform and a short explanation of what repeated, what changed, and why you threw any part of it out.
Question: Pick one thing—where you put it, how hard you press, the room lighting, or moving around. How much does it change what you can actually read?
Try thisChange exactly one thing. Everything else stays put.
Keep the sameSame person, same spot, same posture, same attachment, same length of recording.
Watch forHow much of the recording is usable, how tall the beats are, and whether peaks get miscounted.
Show someoneThe two recordings side by side, and one honest thing that could have thrown it off.
You’ll finish with: a comparison of the two conditions, plus one limitation. “No clear difference” is a valid answer here.
Everything above works fine as a rainy-afternoon experiment. If you are being judged on it, this is what turns it into evidence: run at least three trials per condition, save the raw waveform rather than just the BPM or a screenshot, and label every trial.
Ask Write one question.
Baseline Record a clean comparison.
Change one thing Keep the rest constant.
Repeat and save Keep raw data and trial labels.
Report Show evidence, uncertainty, rejected data, and failure.
Copy this lab plan
What I wondered: ___ What I changed: ___ What happened: ___ What I am still not sure about: ___
Advanced · go further
4. Rebuild the experiment from a published paper.
Every project above starts with your own question. This one starts with somebody else’s—and asks whether you can get their answer. Replication is the least glamorous and most important job in science, and almost nobody does it.
Question: Can four PulseSensors recover heart rate and respiratory rate the way the published method says they should?
The paper: Stevens and colleagues (2024) used four PulseSensors to collect PPG from 15 participants, derived heart and respiratory rates using off-the-shelf hardware and open-source software, and compared the results against commercial wearables. Same sensor you have. Published method. The comparison is yours to run.
Try thisFollow the published method as closely as your setup allows—then write down every place you had to depart from it.
Keep the sameSensor sites, sample rate, recording length, and posture. Hold them constant across every trial.
Watch forWhere your numbers agree with the paper, where they drift, and which channel is responsible.
Show someoneYour result next to the published result, with the differences explained rather than hidden.
You’ll finish with: a replication attempt—including the parts that did not reproduce. A failed replication that is honestly documented is a real contribution, not a failed project.
Recording other people? The consent and review rules in the box near the top of this page apply here in full. Ask your teacher or sponsor before you collect anything.
Then pick any other entry on the PulseSensor in Published Research page and do the same thing. Each one is a method somebody already wrote down for you.
Know what this sensor measures.
PulseSensor records one green-light reflectance PPG channel. It does not measure SpO₂ or blood pressure and is not medical equipment. Breathing, timing, and variability observations remain experimental until validated against an appropriate reference.
