Really good Android apps
 

My own research is giving a lot of choices for android apps, but I can't tell if they really can work magic like they promise and do what I'm looking to do. And I don't want to download a bunch of apps to try out and eat my phone memory to just turn around and delete them. So, throwing the door open, what is everyone's favorite.

First thing I want is a reliable pedometer that not only counts your steps but can automatically tell if you are walking, running or bicycling and track distance of those activities. Right now I'm using LG Health. I ride the same route every day, I've measured the distance on Google maps, yet LG Health has my riding distance different every day, up to a mile wrong on some days.

I also want an accurate speedometer. Waze used to be good but now it's crazy inaccurate. Yesterday is a good example. I'm riding through a parking lot and Waze says I'm doing a piddling 6mph, which felt inaccurately slow. As soon as I stopped pedaling and started coasting it jumped to 12mph. The rest of the trip was the same. Pedaling at 7mph, coasting jumped it to 15mph. Pedaling at a ridiculous 1mph, coasting jumped it to 11mph. Pedaling steadily, it jumped between 3mph to 6mph back to 3mph then to 12mph.
(Before laughing at my pitiful speeds, when my bike is going downhill and being pulled by gravity the fastest it has gotten is 20mph. Anything 17mph or higher is faster than the pedaling mechanism can handle and moving the pedals causes a fast no-resistance spin of the pedals.)

Automatic tracking + high accuracy = battery killer, which is why app developers aren't interested. I don't have specific suggestions, but try looking for an app that has adjustable sensitivity so you can dial it way up if you don't care about battery life.

Cell GPS is accurate within 8 meters. Now add that possible 8 meter inaccuracy to every time your phone takes a GPS reading. If you're on a weaving bike path rather a street, that inaccuracy will add up even more.
I doubt it's possible for a phone app to be more accurate.

For any of a number of technical reasons, the ONLY way to accurately and consistently measure speed and distance ridden is with a wheel mounted sensor.

GPS simply can't duplicate that because it it calculates what might be called "connect the dots" distance vs. Actual distance ridden. It's actually pretty good for most purposes, but will never satisfy data obsessed people who look too closely.

As long as you have a good GPS lock phone apps work fine for 95% of the population. Those dots that are being connected can be 60 seconds apart or 0.5 seconds apart (or any other number if you have a configurable app). The closer together in time the GPS refreshes, the more accurate your readings for speed and distance. (But the harder the app will be on your battery)

Now if you're phone has a crummy GPS antenna and you lose a lock every few minutes your speed and distance data is going to be haywire. (Once I lost GPS for about a mile and when it relocked it thought I had gone that mile in about 1 second so my average speed for the trip was about 100mph. (I ate my Wheaties that morning)

GPS is accurate to within 3 feet or less, even on a phone, if the phone's decent enough. On my LG G5 my GPS can tell which side of my living room I'm in, from inside the house, with the GPS signal going right through the roof. I can bring up my location on Google Maps and look at the satellite feed and I can watch my phone move around my house. My bike phone that I use for mileage and what not is an off brand cheaper Android and it's not nearly as good. It's fine when I'm outside but it can't get a lock on GPS without a direct line of sight (tree branches and stuff not withstanding)

Now it probably wouldn't notice if I was weaving back and forth across a bike path 2 or 3 feet at a time, thus adding distance traveled. But I also don't do that much. I pretty much ride in a straight line. It can easily see me doing circles in a parking lot or something like that though. I can look at the GPS tracking data and see where I did a loop around a parking.

So for the purposes of most people, a phone as a GPS works just fine.

If you're a competitive racer and need hyper accurate Swiss Timing Olympic style accuracy, the phone won't do. If you want to know if you biked 15 miles or 15.1 miles, a half decent phone works fine.

As for apps. Urban Biker. I love it. I don't use it for tracking much. I like it because it has a good dashboard display that tells me thins like distance and speed and can overlay on a map so it's right there if I'm in an unfamliar place.

I want fairly accurate distance so I can know calories burnt. Calorie counting ended up being more realistic if I tossed time spent doing what activity out the window and instead focused on miles, since cycling 20 minutes at 12 miles an hour were the same calories as walking 1 hour at 3 miles an hour.

Speed is important to me since I've been non active for a month and wearing layers now makes me feel slower. I want to make sure I'm not unreasonably slow when part of traffic.

I use Strava on my phone but keep it in my pocket with the screen off so it doesn't use much battery. I use old timey Cateye Strada wireless computers to look at while riding, download the Strava after. I think Strava will figure calories for you, no guess as to how accurate. It's fun to see the maps and elevation.

GPS can only deliver an approximation of real time speed. Those apps are more accurate a few seconds or minutes after digesting the data, and after a ride is closed and the data is finalized.

I've tried most of the popular apps on iPhone and Android and there's always a lag of several seconds before changes in speed register on the display. This occurs throughout the roller coaster routes I ride where I know I'm blasting downhill at 20-30 mph but the app still shows my slower speed from several seconds earlier while climbing to the crest before the next downhill; and vice versa. After several trial runs with each app I don't even look at the display while riding. The phone is either in my pocket or the display is turned off when mounted on the handlebar.

For accurate real time speed a speedometer would be better. More accurate for distance too.

GPS apps do a lot of guesstimation after an activity session, including calculating distance and route. I've run three apps simultaneously on different devices and each shows slightly different data. Just the nature of the beast. It can get even weirder when the device loses GPS sync and the app tries to guesstimate our probable route. Often Cyclemeter would guess pretty accurately and trace my probable route, while Strava's trace would show straight line leaps across barbed wire fences and through pastures between where sync was dropped and regained.

If you do a lot of moving throughout the day -- standing, walking, sitting, walking, standing, etc. -- which would be common with retail and warehouse jobs, you might consider disabling auto-pause/resume on any app that offers this option. Some cycling apps automatically pause at slow speeds, assuming we don't want to include our slowdowns for traffic lights and stop signs, etc. Unfortunately that means the apps will ignore our time off the bike walking, which might matter if we're doing lots of errands. I get huge differences in distance, route, speed, etc., between various apps when I'm combining casual cycling with errands.

Wahoo Fitness for iPhone has user customizable auto-pause/resume threshold and activity profiles. So I created a "cycling/errands" activity profile and set the auto-pause/resume to a 1 mph threshold. This pauses when I'm actually stationary and resumes when I walk or ride. For my "cycling" activity profile I set the auto-pause/resume threshold to 4 mph.

Unfortunately the Android version of Wahoo Fitness seems to lack this user customizable auto-pause/resume threshold. So I'll continue using my older iPhone for most rides, and it works fine without a data plan just running off GPS during activities and wifi at home.

If I'm recalling correctly Cyclemeter and Strava lack this level of user customization in easily setting up activity profiles that match our real world activities.

We agree 100%. For most people, and most purposes GPS data is fine.

But, read the OP, which brought out my response. GPS is not accurate enough for folks who'll obsessively compare data ride to ride and fret over the variance.

A 1 mile variance would not be a big deal on a long ride. A 1 mile variance on my 3 mile commute is a big deal. And if it's getting that wrong by that much, then how inaccurate is the distance walked? Is my 4 to 7 miles during work actually only 1?

Read my earlier post. If you want reliably accurate data, but a $15.00 cycle computer with a wheel sensor. These can be calibrated to small fractions of 1%, and are extremely reliable and consistent.

That said, a 30% variance for GPS distance is way beyond the pale. Something else must be going on.

However, the nature of GPS errors is such that they tend to average out over time, so you'll generally get much more reliable data on longer rides than shorter. It's like tossing a coin, you might get 7 or 8 heads out of 10 tosses, but you'll rarely be much above or below 50 out of 100.

A wheel sensor would be great for only cycling, but my big concern is that if the miles I cycled is so off, then how badly off is the steps/distance I'm walking? I don't have any way to measure the walking so I have to find something that accurately measures cycling to ensure walking is accurate too.

I use Strava and it seems to be reasonably accurate.

The only time I notice inaccuracies is when I walk in and out of buildings at lunch. If I walk out to the Botanical Gardens and back, it's fine ... but in and out of shops it tends to get a bit confused.

When Rowan and I cycle together, our numbers do vary a little but they're really close ... nothing to worry about.

I don't know that you will be able to find anything that will automatically know if you're walking, running, or riding. The easiest would be a high-end wearable like an Apple Watch or Garmin Fenix, where it's always tracking you, and you can change modes with a couple of taps.

I have a Garmin VivoFit, which is like their version of a Fitbit. It does a pretty good job of discerning walking from running, and since it's inertial, rather than GPS, it works indoors and out. Syncs with the phone to get miles/calories in to the Health Apps.
I don't believe that any of the wearables will do cycling on their own, unless they have an integrated GPS, or are synced with a phone.

I use an app ( currently RideWithGPS) to record mileage on my rides, but I mount a basic computer on the bike as a speedometer, even a $10 one works fine, if you take the time to set it up right.
It's on as long as the wheels are turning, and the batteries last forever. On long rides, of several hours, it saves tons of battery life on the phone, since you don't need the screen on at all times.

As far as accuracy, when you're using phone apps, it all depends on your phone. My old HTC wasn't bad, but the iPhone SE that replaced it had a horrible time with GPS dropouts, so much that I stopped tracking with the phone, and just went with the wristband. The 5 got broke, so I got an old iPhone 4S out of the drawer, and found it had great signal. RWGPS can even tell which side of the street i'm on. YMMV

This three feet claim is way better than what you could get with GPS last time I was working on it. Half a dozen years ago GPS could only give you +/- 30 feet. If you've got some references to that kind of improvement, I'd like to see them.


Are you sure (a) your phone isn't doing some sort of differential GPS, perhaps with multiple nearby cell towers? Do most cell phones now also use GLONASS, and is that sufficient to improve the accuracy ten-fold?

6 years ago is an awfully long time.

The satellite system is HYPER accurate and has been for 25 years. It's the receivers that are getting better, both with the quality of the receiver and the software that runs it and the cost to make them accessible to the average consumer.

When I say the sat system is capable of HYPER accuracy I'm not messing around. The military has used tech that's several grades higher than consumer GPS since the first Gulf War. The military uses the same sat system that the public does. But the military has had access to higher accuracy receiver systems for a long time. Remember those Tomahawk cruise missiles that hit a specific house from 300 miles away while traveling a 700 mph? At those speeds you can't be off by 30 feet at any time during the flight or there's no time to correct the course. A 30 ft course error 10 miles from the target could be a 3000 foot error at impact. At that point you may as well be lobbing old Nazi V2 rockets from WWII.

From personal experience in the 90's I can tell you military GPS tech was insanely accurate. They were accurate to within 2 feet, and that was with secondary systems for land nav, not with the pricey receivers the Air Force and Navy put on missiles.

That GPS signal is used for other stuff too other than positioning. It sends a time index that is used to coordinate communication system encryption. (Technically the time index is how the positioning system works. Communications equipment just uses the clock to coordinate) A network of radios all have to be set to the same GPS time to the constantly changing encryption system is coordinated, otherwise radio one is transmitting on frequency 1 with encryption sequence A while radio 2 is listening on frequency 2 and decoding with decryption sequence B. The system hops between several frequencies and encryption patterns multiple times per second. And the radios have to be in PERFECT harmony to work. It uses GPS time to sync. The GPS time has to be perfect to within a fraction of a second. And that was 20 years ago when I used it, before everyone had a Garmin in their car and a GPS tracker in their pocket.

As far as tracking accuracy, my phone is just as accurate without a SIM card and in airplane mode, meaning it's not pinging cell towers to improve accuracy. And even if it was, the cell towers can't tell if I'm the left or right side of a room. They are not that accurate. All the cell towers know is that I'm pinging 3 different towers so I must be in the general area that is serviced by all 3 of those towers.

Sigh. I found some setting on my phone that uses a lot more battery power in order to get a better location fix, and I turned it on for today's ride. Somehow I managed a 39 minute 8.5 mile ride in the 15-20 minutes it took me to get from home to work.

I'm going to try Strava. It can't be any worse than that.

Some phones are better than others for this function. It might not be the apps.

Strava may not do everything you want but I predict you will enjoy it. Fun and easy. Also easy on the phone battery.

It might be your phone. The tracking app you're using is only as good as the phone's GPS antenna, and whatever other locating services it uses.

my experience, using Runkeeper (for running) and RideWithGPS (ride tracking/navigation) My mid-level Android HTC was decent, except in high wind condition. My iPhone SE (5) was horribile. Constant loss of signal, or it would hop two or three blocks away for a data point, then back on course, adding hundreds of yards in seconds, and rendering average speed/pace worthless.

I recently went back to my 7-year-old iPhone 4, and was surprised to find out it's much better than the two newer phones, my RWGPS tracks can even show which side of the street i'm on.
My 8 is on order, but I may keep this little guy and convert it in to a stand-alone navigation/tracking unit.

I once did a gps experiment in the context of a golf device - it comes with the location of the greens, etc and tells you how far you are from those points. I just programmed in 4 points all 100-200 yards from a fixed location in my yard. I then took 3 or 4 readings (spread out across the day) for a month or so. While I had no idea how far each of these points were from me, they certainly should not change. I found the readings to vary about plus or minus 5 yards.

From what I have read atmospheric variations are the major source of error. And the WAAS system was designed to adjust for that. And note that measurements taken close together in time and space are way more repeatable than this. This was back in 2010, so maybe things have changed.

dave

When measured against a wheel sensored computer, Strava's pretty accurate on distance in my experience. Results do seem to vary by device on ride AVG speed, in my experience. I've had phones/devices that are "faster" and others that are "slower".

Calories
Calorie estimates based on GPS or even on a heart rate monitor are quite inaccurate (and often seem to estimate high). And it's likely to be even more inaccurate if the app doesn't know when you are biking or walking.

You might look into a heart rate monitor that can work with your phone. My HRM lets me maintain a hard, but sustainable effort.

~~~~
googling for numbers:

walking: something like 90 or 100 calories per mile.

biking: between 20 and 25 calories per mile.
My faster, moderately hard effort group rides are often around 30 cal/mile. I've been as low as 15 cal/mile on a 40 mile easy paced ride. An all-out 5 minute hill climb was about 90 cal/mile.

Your biking 1/3 hour at 12 mph = 4 miles = approx 80 to 100 cal. (Bikes are extremely efficient!)
walking 3 miles = approx 280 to 300 cal

Of course, these are averages for a whole session. Lots of hills or intense efforts will bump up the numbers.

Power meters on bikes measure calories quite directly, and fairly accurately. (basically: There's about 4 calories per joule, and people waste 3/4 of their calorie burn in heat, 1/4 going to useful effort. So the number of joules just about matches the number of calories burned. ) Many riders report somewhere in the low 20 calories per mile from this data.)

Averages
All GPS apps and devices have to do some data smoothing. They all use slightly different methods, so there's always variation. My ride recording will report slightly different distances, and somewhat different elevation gain and average speed if I download the same data file to 3 different apps. But they are all within 10% or 15% usually, and that's good enough.

~~~
I have a wheel sensor with my Garmin GPS, for accurate distances and speeds. It's usually quite similar to other rider's Strava phone distances. The elevations can be off a little, maybe 10-20% variance between different devices and apps.

GPS accuracy:
My older Garmin 705 records once a second. Here's a ride I did, heading uphill north outbound, then downhill southbound and turning right. You can even see which side of the road I was on. The one-per-second dots are color coded by speed.

Newer phones are faster and more accurate with GPS than my old, slow Garmin. But trees and buildings can affect the accuracy.

https://lh3.googleusercontent.com/-5...2520points.jpg

I'll have to second the accuracy of the garmin 705. It works better than expected and can upload to garmin connect (obviously) and strava. And since it's not the latest gee-whiz-bang device, you can pick them up for a small amount of $$.

But if you're sort-of a gear geek, you may prefer the more recent units. Unfortunately the price does go up, but not necessarily the accuracy. I guess you get to pick your poison, so to speak.

THREAD DRIFT ALERT: When I was first in the army in the late 90s early 00s the explanation for the greater accuracy of the military over the civilian devices was that they didn't want people to be able to make cheap guided missiles with super accurate GPS devices they could buy on the market.

Of course it was just a fellow soldier telling me this so that may very well have been army urban legend. (Like stress cards and the bullet on top of the flag pole.)

If true, evidently something changed in the early 00s because the civilian stuff suddenly became super accurate -- and much more reliable.

The stuff we had when I was deployed in the middle east was accurate enough, but SUPER SLOW and GIGANTIC. It was easier to buy handheld Garmins and have them shipped to us, sometimes, than use the various handheld and vehicle-mounted GPS units the military officially provided.

(Side note: we had delays getting repair parts for our HUMVEEs because GM found it more profitable, at the time, to manufacture parts for the civilian Hummers.)

So back to the OP: Yeah, sure, your phone is plenty accurate enough for every day purposes.