Hackathon: Writing a Microsoft Band App In One Weekend - Update 4

Update 4: Hear My Heart Beat

Can you see the 20 strokes in here? Paddle Mate found 17.

It was a dark and storming night. Taylor Swift sang her heart out in downtown Seattle while I coded my heart (beat) out writing Swift in Sammamish.

But enough of that, onto the progress report. Last night I:

• Added Heart Rate Monitoring into the app.
• I said what the heck, let's grab that calorie count from the Microsoft Band.
• Then I lost my mind and decided to add a "Reset" into the app to make collecting data easier.

Oh ya, then I hooked the stroke counter. It was about time I added the feature I started this app for.

Heart Rate

I now collect the Heart Rate samples while I update the current Heart Rate label in the app. Once the workout is complete you can now view the Heart Rate in the in-app graph using the heart rate/stroke toggle icons along the right of the graph.

Calorie Counts

Just displaying the calories burnt during the workout, as reported by the Microsoft Band.

Reset

During data collection on Saturday morning I had to force close the app to clear it. I had to do this between each data collection.

We aren't animals here. Force closing apps? What will be next? Side loading apps you got from a friend off of 10x3.5" floppy disks which are really not floppy? No, we won't.

Behold the Holy Reset of Antioch!

Yes, the reset was that good.

Wet And Wild 2: Get Me Some Samples

With only one Sprint Pace sample I decided I needed more samples. Here are the results of today's collection with the Stroke Detector I worked on yesterday.

Pace	Strokes	Detected Strokes	Accuracy
Easy	30	27	83%
Sprint	10	10	100%
Easy	25	23	92%
Sprint	20	17	85%
Easy	50	47	94%
Mixed	120	96	80%
Easy	100	84	84%

Overall I am pretty happy with the new stroke detector. With an 80%+ accuracy I think that is pretty good for a one day of crash course Linear Regression, Moving Average, Classification, and Standard Deviation coding session.

My next phase of stroke detection I think will involve some machine learning to see how more accurate I can get the algorithm. At the moment I see I need to do a couple things based on the samples:

• Classify strokes and trailing strokes by pace (Slow, Medium, Fast) and then use that classification to adjust the stroke detection algorithm.
• Better rejection of garbage samples (non-strokes).

About Class Structure

You know that joke about MVC? Massive View Controller?

That's this app at the moment.

I know, I should hang my head in shame. But screw it, I will put making this an architecture marvel to knock your socks off on my "Backlog".

To be truthful it is not all a MVC. I have the following classes:

• PaddleCounter class to handle stroke analysis.
• Ring class to handle the ring buffers for moving average calculations for the PaddleCounter.

The rest of the logic at the moment is in the ViewController but will change once I extend the app.

Notes From Wet Tests

During testing on the lake I made some notes about the app:

• As we know,trying to read an iPhone in the sun sucks. I am going to need a high contrast UI with very large components if the user has a hope in hell of seeing the strokes, heart rate, etc.
• The waterproof plastic case I have is soft and as such touching one point on the screen may cause a different location to register the tap. This means that during a workout I will need to add a dialog to prompt a user "Do you really want to stop you workout?"

Hackathon: Writing a Microsoft Band App In One Weekend - Update 3

Update 3: Wet and Wild

Data! We have Data! We have tons and tons and tons of Data!

Samples for 40 Strokes at Sprint Pace

Those same 40 strokes charted and tweaked for algorithm building

Sample Gathering

This morning I made a first pass of the Stroke Counter algorithm. I then went down to the lake and collected 10 sample groups to run against the Stroke Counter.

The sample groups included:

• 20 Strokes at an Easy Pace
• 5 Strokes at an Easy Pace
• 40 Strokes at a Sprint Pace
• 60 Strokes at an Easy Pace

Turn Simulation

• 20 Strokes on Left Side Only
• 10 Strokes on Left Side Only
• 20 Strokes on Right Side Only
• 10 Strokes on Right Side Only

Odd But Why Not

• 20 Strokes Backwards
• 20 Strokes Backwards

Algorithm Test Run 1

The first test run of the sample data against the initial Stroke Counter sucked...I mean it was very inaccurate. There goes trying to "guess" my way through this.

Ok, so more work was needed on the algorithm and that was my day. About 9+ hours of tweaking later and I am passing all Easy Pace tests. Tomorrow I will run a new sample session with this algorithm and see how it stands up during use in the kayak.

Improvements For Sprinting

For Sprinting or fast and strong strokes I am going to tackle the problem by looking at peak samples in trailing sample ranges. If a peak reading goes above a known Sprint High Water Mark (HWM) that is outside the normal HWM then I will apply a modifier to my HWM and LWM (Low Water Mark). Then as long as the trailing peak samples above the "Sprint HWM" I will apply a modifier to my stroke classification.

I can see that one of the trickiest problems will be handling the transition from an Easy pace to a Sprint pace. This will be interesting to solve.

Tweaking The Algorithm

So, what did I do to tweak the algorithm today?

The exporter I built into the app exports the sample data into a file with Comma Separated Values, with one reading per line. I am using the Share Sheet from my app so I can easily save my samples to Dropbox after each sample collection.

Once I got back to the house. I imported the sample groups into Numbers and created charts for each sample set.

For each chart I left the chart zoomed out on top and then I created a duplicate chart below and zoomed it in. I applied the following adjustments to my chart:

• Series
• Turned on Trendlines with Moving Average and tweaked the Period to find a sweet spot for the algorithm high and low water marks.
• Turned on Error Bars with Standard Deviation to help isolate the Trendline while reviewing between the 10 sample sets.
• Changed the Trendlines to Red and increased them to 4 pt. This helped them stand out in the charts.
• Axis
• I added Minor Ticks and tweaked each chart so I could better isolate the peaks and valleys of the sample.

Beginning of the Easy Pace 60 Strokes Group

Cool Charts! Now What?

I got several suggestions from friends to do machine learning, use Azure, or many other tools. For many of these, they are overkill for the initial app's stroke counter. As well, I am trying to create the initial app over this weekend so I don't have enough cycles to get the app working and learn machine learning. I did dive into machine learning a bit today and picked up some pointers to help me design the stroke counter.

I plan to take time to do further refining of the stroke counter later and look at how machine learning could help with that. For now, let's get the basics working.

So, I did the tweaking the hard way. I wrote test cases that consume my sample files and run them against the stroke counter algorithm. This has been very helpful and allowed me to tweak the algorithm to the current state (mentioned earlier) where I am correctly detecting strokes for an Easy Pace. I am also 75% correct for Sprint Pace but I want to get this closer to 100% accuracy.

Time for Pizza

It's not a Hackathon without Pizza. Calzones are on the way and then I get back to coding.

Hackathon: Writing a Microsoft Band App In One Weekend - Update 2

Update 2: Stroke Counter First Pass

I imported a sample set I collected of a simulated kayak sitting on the deck. Then I wrote a PaddleCounter class that I can feed accelerometer into. The class calculates the stroke count as it processes the incoming accelerometer data.

This data then needs to be displayed in the awesome UI I cooked up on a budget. Ok, it never cost me anything and it shows, but you get the picture.

As we can see a couple new features are slipping in. I figure I should track Heart Rate and average Heart Rate for the session. I am also going to collect calories burned and compare that later to calorie burn estimates to see if it is reliable. But I will hook those up later this weekend.

Collecting Samples During a Session

I decided to go with NSNotifications to send out stroke updates from the PaddleCounter class. So, when the PaddleCounter detects a stroke it will send a notification that the main ViewController can then use to display the stroke change.

Testing the Stroke Counter

I want to make testing the stroke counter as easy as possible. I plan to collect different samples and track the number of strokes. Then algorithm development will go like this:

1. Feed my samples into the stroke counter algorithm.
2. Compare what it dumps out with the expected number of strokes.
3. Tweak algorithm, rinse, repeat.

My first stroke counter at about 1 AM sucked and came out with terrible results. What to do?

Charting the Samples

Ok, I was trying to eyeball the samples and hence it was hard to track the trends. So, I need to chart these samples to better understand the data I am getting.

Before I try to rewrite the stroke counter I am puttingthe samples on a chart. For this I am going to use ios-charts by Daniel Cohen Gindi to draw my sample set so I can look for trends.

Here's the new UI with the raw data in the chart and then a zoom in on section:

Samples with Default Zoom

Samples Zoomed In

Ok, now it's time to go hit the water and collect some real samples.

Hackathon: Writing a Microsoft Band App In One Weekend - Update 1

2 Hours in and I have:

• Basic UI for Data Collection
• Accelerometer Data Collection to CSV file using NSOutputStream.
• Export of CSV file using the UIActivityViewController.

Next up:

• Write a stroke counter first pass using some simulated kayaking samples.

Here is the UI for the first set of data collection.

I know, its amazing. Where did I get this UI skills? What can I say, I an old school green on black terminal kinda guy.

Harkathon: Writing a Microsoft Band App in One Weekend

My wife and kids are out of town for the weekend so what's a geek to do? Hold a Markathon of

course.

A Markathon? That's just a Hackathon with a Mark.

What type of App should I create? I've decided I want to learn Machine Learning but I want to be practical and start small so I am going to create a simple stroke counter app for kayaking. For this I will do basic calculations to create the stroke counter. For this weekend the goal to lay the groundwork to do more complex analysis of strokes. 

With this weekends Markathon my goals are:

• Build a harness to collect sensor data.
• Be able to export sensor data during a workout from the Microsoft Band to a CSV (Comma Separated Value) format.
• Collect some Kayaking sample data.
• Crunch the data and create a basic stroke counter.
• Create an app front end for the stroke counter.
• Prep the app for Test Flight.
• Find some testers who want to try out a stroke counter for the Microsoft Band.

I will post throughout the weekend as I complete stages. Follow along and feel free to tweet me if you have questions.

Launched! Selfie Band for Microsoft Band

Well, I finally completed an app (again) for the iOS AppStore and launched it. The app is called Selfie Band, a camera app for the iPhone which allows you to use a Microsoft Band as a remote control. Check out more about the app at the NewThistle Blog: Announcing Selfie App

You can also checkout the Selfie Band HomePage.

My first app was launched in 2009. My second app (with a team of 6 Cocoaheads from the Ottawa Cocoaheads) was launched in 2010.

Then I went to work for BlackBerry for a little under 4 years where the employment agreement included a non-compete clause that meant no apps in competing AppStores. Hence, I was a Mac OS X/iOS dev at BlackBerry but could not ship my own apps.

I left BlackBerry at the end of 2013 and hence that restriction was lifted. At first, my biggest problem was what to write?

When I joined the Microsoft Band team (as a contractor) there were no restrictions on releasing my own apps. In fact I have been encouraged since I work on the Public SDK and help third party developers integrate their apps with the Microsoft Band on iOS.

Selfie Band has been several months in the making, though it might not look like that to some. In that time I took on maybe too many after-hours activities. A couple include:

• Prepping and presenting 2 talks about the Microsoft Band SDK.
• Writing 3 Band Apps "almost" ready to ship.
• Starting no less than 4 other Microsoft Band Apps.
• Helped a pre-investment startup get an early prototype setup for initial integration with their Web API.

So, as time dragged on and I had not shipped anything I knew it was time to concentrate on one app, cut features and ship it. There are many things I would love Selfie Band 1.0 to have but shipping was more important than adding features users might never want/need.

About 5 weeks ago I put the final touches on Selfie Band and prepped for AppStore submission. After a month in review including one rejection, Selfie Band came out of review last Friday and it was left to me to hit Launch.

So, last night, at NSCoder I put the final pieces in place and Launched!

It feels good to be back in the AppStore with something new that I designed and wrote myself.

If you have a Microsoft Band and an iPhone then check out Selfie Band so you can take photos from your iPhone with your Band.

Creating Family Traditions

Dragging the Kayaks back up the lake's "shore"
which is very long due to the use of the
lake as a reservoir for the Yakima Valley
Agriculture industry.

Last year I started taking one of my boys for a camping trip with just them and Dad. This year was time for my youngest to start this tradition with me.

For our trip, I decided I needed to find something where we could use his new Kayak. I got him one for his birthday and it was a hit. He loves kayaking like his older brother loves mountain biking.

So, after some research I decided on Cle Elum Lake since it had lots of camping nearby on the Cle Elum River. It had an interesting lake to kayak and there are lots of opportunities to fish.

Now, why only one of the boys? Having a chance to spend solo time with each boy gives my wife and I a chance to really hear them, bond with them, and not have other demands getting in the way. My wife takes the boys on a trip to a different city and I take them camping.

I learned a lot camping with my Dad and spending many days fishing with him. He never had to deal with the challenge of having multiple kids vying for his attention. For me, taking each boy on a solo camping trip makes the trip more enjoyable for me (lower stress from sibling rivalry) and lets me spend solid quality time with my boys.

I don't have video of Finn kayaking since he decided he wanted to use my kayak so he could have the camera and since I had a new kayak which he had not used. So, I have some footage (thanks to Finn since he worked the camera for most shots) from Finn's perspective as we kayaked in Cle Elum Lake.

The trip went awesome and we had a wonderful time kayaking, fishing, and wandering the banks of the Cle Elum River. It was a beautiful spot that I will be exploring more.

Hiking the very low Cle Elum River near our campsite.