Full Machine Learning Project — Coding a Fitness Tracker with Python (Part 1)

In this video series, we are going to build a fitness tracker in Python that can classify various barbell exercises based on accelerometer and gyroscope data. This will be a full machine learning project and new videos will be released weekly, so subscribe to stay tuned!

⏱️ Timestamps
00:00 Introduction
01:16 Project objective
01:58 Project background
07:46 The quantified self
10:40 Project overview (what you will learn)
17:37 Action items (complete these now)

Project overview (what you will learn)
Part 1 — Introduction, goal, quantified self, MetaMotion sensor, dataset
Part 2 — Converting raw data, reading CSV files, splitting data, cleaning
Part 3 — Visualizing data, plotting time series data
Part 4 — Outlier detection, Chauvenet’s criterion, local outlier factor
Part 5 — Feature engineering, frequency, low pass filter, PCA, clustering
Part 6 — Predictive modelling, Naive Bayes, SVMs, random forest, neural network
Part 7 — Counting repetitions, creating a custom algorithm

If you find these videos helpful, consider subscribing at @daveebbelaar

[MarioCeballos-jspy]

I learned that keeping track of your habits is key for improvement. I study biomedical engineering, I want to create an app that can process EMG data focus in pro players. Sports are my passion, combined with this machine learning course is a good start point in my data science career. Thank you for all this knowledge, I am really excited to follow this course.

[nethervvoid]

I've been in data 18 years now (Almost done it all: DBA, engineer, modeler, BI, Analytics - last 7 in management), have been feeling the push into data science, and now after returning from Next 24 I am more than certain everyone manipulating data needs to push into data science. Everything is merging (and admin is somewhat vanishing honestly). I really appreciate your channel. Going to be tapping your helpful vids a ton. Already installed and configured vs code, git, github, extensions, etc. Thanks so much.

[ahmeterdonmez]

Very recently I picked this video tab and moved left for future watching. Dave had 89k subscribers. Just now I refreshed the page and saw that he has 89.5k subscribers. I'm so glad you got the reward for your hard work Dave.👍👍👍

[vipuljadhav]

I am new to machine learning. I am a quantified self (I would mostly measure cardio i.e walking and running). Once I finish this video series I will explore and extract data of all the wearable devices I could and apply the knowledge I am gaining through this series into my models. Thank you Dave

[seanfeng]

This is very interesting, I will likely binge your whole playlist. Thank you for putting this together- I can tell you put a ton of work into this and to share such high level material for free is really appreciated.

[newellfitness]

I am a quantified self and I am going to be working as EPM with ML and Signal Engineers. This will be helpful! Thanks Dave

[SamuelKlettNavarro]

Hello Dave, I use a Garmin Solar to track my workouts, cardio activity and other daily parameters. I don't use during my sleep, for that, I have been tracking my nights with Sleep Cycle, for almost a decade now! Thank you for this

[duncanprins]

Hi Dave, great videos. As an old-timer (50+), I'm really fascinated by today's possibilities. I started this week to get a grip on Jupyter, but now I'm switching to VS Code as mentioned in your video. This quantified self-project seems like a nice trail to get familiar with this topic. I've been collecting data for more than 5 years with my Apple Watch, but I wasn't really able to do anything sensible with the data. Maybe this gives some ideas. It's great to see a fellow Dutchman succeeding in this field. Looking forward to the next videos.

[joshuapinto]

I love this . As a beginner I will try this out . As u said some things are advanced but could learn from you .

[hasanrants]

I'm excited to follow the project, today is 3rd August 2024.
I consider myself as a quantified self because I also hit the gym and measure my lifts, apart from that I try and measure my progress for learning Machine Learning and making progress as a skilled Data Scientist.

[noPwRon]

Excited to dive into this series. I am a dancer and have been struggling to find ways to track my improvement with specific techniques and movements.

Hi Dave, i am using garmin fenix 6 to track my running, bike, heart rate, vo2max, sleeping etc. Looking forward to see all the videos. And cheers for the today's discussion on data science.

[dusanzdravkovic]

I am a quantified self, measuring everything you said as well on the apple watch. great vid!

[jpclark]

Dave, your videos are a great help to me. Thank you! I am not routinely a quantified self. I go through phases though where I record time and cardiac parameters during workouts. You don’t mention it here, but there is a quantifiable self analog around personal finance, if you track ins and outs over a day as well as interactions with personal health, $/calorie.

[Beverage]

I started watching your videos a while back and honestly, I was not interested in machine learning before, but I think would like to learn more or maybe go into the field of data science. As I am just starting out, I'm still not sure what to do or how to begin, there are just too many stuffs out there, but your videos are helping out a lot. Many thanks!
btw I'm not sure about me being the quantified self, I just do random exercise.

[UdaiBhati-lckb]

I'll be doing it fully excited for the project man...

[littleKingSolomon]

Thank you very much for these videos. I do not consider myself a quantified self, yet!

[brians]

sup Dave, i consider myself a quantified self because i measure the number of steps I take everyday and sleep hours

[nuwayir]

waiting for more series like this, it's awesome 🤩

[jaimesanchezclaros]

I used to qunatified my sleep quality, but the idea of tracking my workouts sounds amazing! The more data to have the better to track how the progress go up and up :)