I’ve thought from the start that the Withings had a wonky way of measuring elevation, but I pretty much ignored it.

But today I decided to take the GPS on our walk into town, and compare the GPS elevation rise to what the pedometers were giving me.

Now, from everything I’ve read, the FitBit gives you ten feet of elevation for every floor climbed. That’s easy enough.

But the more I looked at the Withings data, the less sense it made. So this evening I decided to sit down and try and figure out just what the hell the Withings data was trying to tell me.

For my walk, I wrote down the starting elevation and the ending elevation.

My numbers were 1619 ft and 2026 ft, which gave me a change in elevation of 407 ft.

Then I asked Michael to see if he could give me his elevation gain from the website.

His number was 95 ft.

Wha?

So, I go to the website, and get the elevation gain for each 30 minute time period.

Can you read that number? The time is from 6:30 to 7:00 and it gave me 92 ft and 30 floors. Adding the numbers for the three 30 minute periods (16/6, 0/0, 92/30) I get an elevation gain over 90 minutes of 110 ft.

100 != 407

So, I note flights and elevation for every 30 minute time period [(14, 2, 13, 2, 6, 7, 1, 8, 35, 24, 4, 20, 9, 3, 3, 2, 6, 30, 5, 1, 8, 4, 1 flights) and (44, 7, 40, 8, 19, 22, 4, 25, 105, 74, 14, 60, 28, 9, 10, 7, 18, 92, 17, 3, 25, 12, 3 feet)] and I get a total of 646 feet and 208 flights of stairs.

Look back at the image. It’s giving my total elevation gain for the day: 2190 ft. None of those numbers seem to have any relationship to each other.

I fully admit that simple math is not my strong point, but I don’t see how my daily elevation gain total can be three and a half times higher than the sum of the individual time periods.

I’ll also note that the 646 ft measurement is much closer to what the FitBits gave me, (assuming 10 ft elevation rise per flight of stairs) an elevation gain of 640 or 670 feet (64 and 67 flights of stairs) for the day.

This… this makes no sense. And even assuming the sum of the individual time periods give me the correct elevation gain, the flights of stairs it gives are ridiculous, unless the Pulse things that a flight of stairs is only three feet high.

So, now I truly, deeply doubt the information the Pulse has been giving me–at least anything beyond basic step count.

Yeah, I continue to find this fascinating.

First:

I carried the GPS to work today, so I could determine precisely how long the walk to work is.

The walk to work was 1.0 miles, the walk from work was 1.1 miles.

But what amused me was the elevation graph.

It’s ALWAYS uphill.

Second:

Over the past year and a half I managed to put on enough weight that I stepped on the scale and read: 155.* That is more than I feel is healthy for me, so I need to lose at least 15 pounds. (Hence the pedometer testing.) Because of its integration with MyFitnessPal, I decided to get a FitBit One, not to track steps as much as to keep track of my calorie intake / outtake.

The FitBit One was waiting for me when I got home, so of course I immediately set it up.

Because I am a dork, this amused me:

And of course I’m adding the FitBit One to my pedometer tracking.

I can’t wait to find out of the One overcounts steps to the same extent as the Ultra.

(rubs hands together in glee)

I’M SO EXCITED!

* While I was clawing my way out of the pit of depression, I essentially stopped weighing myself, because I didn’t need one more thing to feel bad about. I believe that was the correct decision. Unfortunately, I was also been binge eating, so now my pants are too tight and it’s time to pay attention to what I’m putting in my body again.

Here’s a peek at one of the charts I made from Saturday’s data.

It looks at the count differences during different periods of time. You can see that biking was problematic for the pedometers–which I knew it would be. I think I’d have to put the pedometers on my shoes to make them count accurately during biking.

But you can still see the tendency of the FitBit to overcount steps, regardless the activity.

I tested the Accupedo pedometer phone app yesterday, but the step count was so far off, it skewed my calculations. Badly.

I turned down the sensitivity (the better phone apps I looked at had a setting for changing the sensitivity), and today’s counts were much more in line with the other pedometers.

The first part of today’s activities I knew would be problematic for the pedometers: we went biking at Coopers Rock. It’s amusing how far off each device is from every other device.

But, I had my GPS, so I got all that information.

Then I realized that when we walked to dinner, I should take the GPS! Further proof the Withings Pulse can’t calculate elevation for shit!

First up, what Withings recording for me elevation today, all day:

From the time period of 3:30 to 5:30, it gives me an elevation gain of 119 feet.

And now, what the Withings recording for Michael. Keep in mind that we spend the entire day together.

Now here’s what the FitBit recorded for Saturday:

Then numbers aren’t directly compatible, as the FitBit displays the data is flights of stairs. But you can see the trend.

Here’s the route recorded by GPS:

According to my GPS, my elevation gain for the entire trip was 310 ft. So the Withings gave me about a third of the elevation I actually had.

So, compared both to the FitBit and to actual GPS measurements, the Withings Pulse has been giving me only about a third of the elevation it should if I’m climbing hills (versus climbing stairs–it seems to get the stairs fine).

BAH HUMBUG.

Oh, in case you were interested, here was our elevation from our biking at Coopers Rock:

And here’s the route:

Here’s a quick peek at my data collection file.There will be charts.

I’d noted previously that the Pulse had dissed me on elevation.

Well, today I got further proof that the Pulse is screwing me on elevation.

Michelle’s Fitbit:

Michelle’s Pulse:

Michael’s Pulse:

At noon, I climbed stairs during my lunch walk. At ~5PM, Michael walked home from work, which is mostly uphill.

At ~7 we walked home from dinner, which is also mostly uphill (Essentially, we live at the top of a hill, so it’s always uphill). The FitBit shows that as the greatest elevation of the day. The Pulse seems to give me almost no elevation, but Michael’s Pulse gave him credit for walking up University.

Needless to say, this irks me.

More on pedometers! Yay! (The first post is here.)

Because I am a TREMENDOUS dork who loves data, the ability to view my data over time is very important to me.

It’s also what drove me away from the FitBit universe–the inability to view, manipulate, and download my own data in any useful manner.

FitBit Dashboard (web)

FitBit Dashboard (Android)

I’ll admit I haven’t tried in almost a year, but I was never able to get my step data downloaded from the FitBit website in any useful manner.

Withings Dashboard (web)

Withings Dashboard (Android)

Sadly, I got no useful downloaded data from the Withings website. Boo!

Omron Dashboard (web)

The Omron site gives me my data in a pdf. It’s not a cvs file I can dump straight into Excel, but it’s something I can put into Excel with relative ease.

I’ll be curious as to whether they come out with a phone app. I’m betting someone will.

Omron Health Management Software (PC install)

In addition to the many views here, I’ve been able to export the data into a cvs file that I can open in Excel. This gives me EVERYTHING. It’s pretty awesome.

But of course, if you have a web presence, then you can set up partnerships with other companies and web apps. Some of these can be incredibly useful, such as a partnership between your activity monitor and an app that lets you track what you’ve been eating. Combining the two tells you how many calories you’re burning through movement, and lets you know how many calories you can eat that day for your goal (maintain weight, lose weight, gain weight).

I really like MyFitnessPal, even though I haven’t used it for months. It is very good at making you aware how many calories are in what you’re eating, and you can also track salt and other things you might be watching for health reasons. In fact, MyFitnessPal was what convinced Michael that Fettuccine Alfredo was not the best choice at an Italian restaurant. (Oh look! You’re going to eat all your calories for the day! Too bad you already ate breakfast and lunch!) Although I don’t have high blood pressure, I found seeing the sodium content of many foods to be shocking.

And the ability to scan the bar codes of foods is extremely useful. And if you cook from scratch, you can create recipes and calculate from there, though I found it more useful to use MasterCook for the calorie calculations.

I’ve tried lots of other different apps, most of which–for a variety of reasons–didn’t work for me, but that doesn’t mean they’re bad. FREX, I prefer to walk indoors, so RunKeeper didn’t work for me at all, since it relies upon your phones GPS.

FitBit and Withings are partnered with lots of sites and apps. Omron almost none, although when I got the Withings Pulse, there weren’t that many apps that used the Pulse data (in fact, most of the connections are more for their blood pressure monitors and scales).

Here are a few of the more popular apps for each tracker:

FitBit
Lose It
MyFitnessPal
SparkProple
DigiFit
MS Health Vault
MapMyRun
Earndit

Withings
RunKepper
MyFitnessPal
BodyMedia
FitBit
LoseIt
HealthVault

Omron
Runkeeper
EarndIt

So, last year I bought I got a Withings Pulse and was very optimistic / hopeful about how it would work out. It was a new product, the website was still being built, and there were regular software updates to the device.

Unfortunately, the Pulse has failed to live up to my expectations.

1. The pulse rate measurement portion of the device stopped working several months ago. I didn’t use it frequently, but it was annoying that it completely stopped working.

2. The website never developed a way for me to download my data. As that was what made me want to stop using the FitBit, that was a huge problem for me.

3. You cannot edit your data. This means that if you forget to turn off the sleep timer, it says you slept for 12 or 18 or whatever hours. I have obsessive compulsive disorder, so this bothered me IMMENSELY. Immensely, as in, I stopped using the sleep timer completely, because I couldn’t bear to see the incorrect sleep times.

4. Although you are supposed to be able to compare data with other Pulse users, Michael and I haven’t yet gotten that feature to work. So I’m collecting data in a void. Which would be okay if the data was accessible to me to download. But it’s not.

5. The altimeter is off. Michael and I spent a day doing the exact same things, and his altimeter recorded twice the height mine did, when, if anything I should have had a few more flights of stairs than he did. Plus, it gives the distance in feet, instead of flights of stairs. I can grok flights of stairs, but a change in elevation measured in feet means next to nothing to me.

So, I need another pedometer.

After a lot of research, I purchased two: The Omron HJ-32 and the Ozeri 4×3 Sport Digital Pocket 3D Pedometer.

The Ozeri was immediately sent back. First, there was no way to download the data, but more importantly, it did NOT (as stated in the description) count flights of stairs climbed. If I can’t download the data, then I’d have to type it all into a spread sheet, and… just… no. Which is too bad, because it was a nice pedometer.

So now, I’m carrying around four pedometers:
1. My original Omron HJ-720
2. The Withings Pulse
3. My old Fitbit Ultra
4. The Omron HJ-32

These are four very different Pedometers, with four different ways of moving the data from the Pedometer to a computer.

1. The Omron HJ-720 is a dual axis pedometer. This means that it measures most accurately in an upright position (such as sitting in a belt clip). It does not have an altimeter, so it only measures steps, not stairs. Data transfer is done by plugging the pedometer into a mini USB cable, you then open the Omron program on your computer, and tell the Omron Health Management Software to grab all the data from the pedometer.

2. The Withings Pulse uses a 3D motion sensor and has an altimeter, and can be placed anywhere on your body in any direction, and measures steps and altitude. The Pulse uses bluetooth to transmit data from the pedometer to your cell phone, and then your cell phone transfers that data to the Withings website. Which means that you must have a cell phone with a bluetooth connection, and you have to leave your bluetooth on all the time if you want it to sync automatically. I had problems with the pedometer automatically syncing data with the website.

3. The Fitbit Ultra has a 3D motion sensor and an altimeter, so it can be placed anywhere on your body in any direction and measures steps and stairs. The device transmits the data wirelessly to the base station plugged into your computer and from there to the FitBit website. The data is never actually on your computer, it just passes through your computer to the FitBit website. Unless you have multiple base stations, the data is transmitted only when you are near your computer.

4. The Omron HJ-32 (I love Omron, but they do NOT have catchy pedometer names) is a tri-axis pedometer, which means it can be placed anywhere on your body in any direction, but does not have an altimeter, so it only measures steps, not stairs or altitude. It is essentially a USB stick with an attached pedometer. You pull off the cap, plug it into your computer on which you have installed the software, and your data is pulled from the pedometer and put on the Omron website. From the website you can then download you data.

All four pedometers come with belt clips, but I have them stuck in my pocket (because I generally don’t wear a belt).

The FitBit has the easiest data transfer–it happens without you having to do anything.

The Omron is best for giving you your data, in a format you can manipulate as you please.

The FitBit is best if you want to compete and compare with friends and family. There are tons of people using the FitBit. (I don’t know anyone else using an Omron.)

The Omron (at least the HJ-720–we’ll see about the HJ-32) is the sturdiest pedometer I’ve ever seen. I bought mine in 2009 and it still works. I’ve put it through the washer at least twice, dropped it countless times, and it still works perfectly. The FitBit Ultra had problems charging and with the outside case cracking. Not all the features of the Withings still work, although the pedometer still functions as it did when I first bought it. I am pretty sure that neither the FitBit nor the Pulse would survive a trip through the washer.

Pedometers with data display:

Pedometers compared to a tube of lip balm:

Pedometers stacked atop my Galaxy S4 (in an Otterbox case):

Pedometers with charging cables or data transfer cables:

Never fear, I’ll have data comparisons coming up next soon.

We went to Green Bank Saturday, which is home of the National Radio Astronomy Observatory.

We amused ourselves at the Science Center.

Here’s Michael:

Some self-portraits:

Michael’s picture of me:

A month or so ago, Michael and I had an argument about what you call an object such as the one he is standing next to here.

One of us said, “trunk” and the other said, “stump.”

Who is correct (or is there a more accurate term)?

Also, one of us has a degree in Forestry.

I think that I’ve generated enough data now.

Couple things: where the pedometer is placed affects the step count. I wore the pedometers in several configurations: on my waistband, in my pocket and a combination of both. When they were on my waistband together (Friday), the Omron and the Pulse gave the closest results. When in my pocket together (Monday), the Pulse had a slightly higher count than the Omron.

The Fitbit always had a higher count, averaging 800 steps more than the Omron. (The Pulse averaged 15 steps more than the Omron.)

The Pulse and the Omron were relatively close in counting active minutes, with the Pulse being higher, due to the fact that you have to be active for ten minutes before the Omron starts that time.

A couple things I didn’t mention previously:

The Pulse has the ability to measure pulse rate. You hold the device between your fingers, cycle through the menu until you find the heart (pulse rate) icon, touch that icon and it measures your pulse through your finger. If you have the Pulse in the wrist band for sleeping, make sure the device is resting over the pulse point in your wrist and start the measurement.

It looks like the measurements are relatively accurate on average, but not very precise for any single measurement. The chart is hard to read, but I would take several measurements, one right after the other. The results are clustered around what should be the correct value, but can fluctuate wildly.

I’ll probably play with the pulse measurement more in the coming days, but for now, I’m not that impressed with it. Which is fine, because I wasn’t that interested in that option.

The battery life on the Pulse is nowhere close to as long as the battery life of the FitBit, however, on a positive note, the way it charges, through a short mini USB cable, means I’m less likely to forget to put it back on. (I had to resort to charging the FitBit at work, and taking off my shoes and putting the charging station on my shoe, to remind me to put the thing back on.) I also like the charging monitor better, and think I won’t have the same charging issues I did with the FitBit (sometimes I would put it on the charging station, and it wouldn’t charge, or the display wouldn’t update to show me that it was charging, so I generally had to fuss with it).

Another point in the favor of the Pulse is that it’s a brand new device and so the phone and web applications are still being developed, and one of the things that will be added is the ability to download your data as a CSV file. I have NEVER been happy with the fact that FitBit wanted to charge you a subscription to download YOUR OWN DATA, and that the data is “allowed” you to have was in no way the full data set you created.

I am hoping that the data set the Pulse generates will be closer to that created by the Omron. I’ll let you know how that one goes as well.

One additional thing I haven’t tested is wearing the Pulse while hiking and biking. I’ll be curious to see how accurate that data is compared to the data generated by my new Oregon.

No updates yesterday, because yesterday was all napping and reading.

The FitBit continues to significantly over-count steps. The Pulse vacillates around the total for the Omron, sometimes over, sometimes under. Interestingly, for the entire time period, the Pulse averaged out to one step more than the Omron. The FitBit averages 783 steps more than the Omron.

The FitBit is obviously delusional with regard to Active Minutes.

I’ll be very interested to see today’s totals.

Obviously, I need more data points.

WOW.

Here are the end of the day totals:

Average of Differences Wed – Fri

Steps
FitBit: 774
Pulse: -559

Active Time
FitBit: 230
Pulse: 20

A quick note about Active Time. You have to be moving for ten minutes before the Omron begins to count active time, so it won’t count any walk 9 minutes or shorter, no matter how quickly I walk. For the FitBit, I combined Fairly Active and Very Active minutes; for the Pulse, I combined moderate and intense activity, although over the three day span I had less than 5 minutes of “intense activity.”

I think that the Pulse is correct in the amount of “intense activity” I had. I was pretty lazy this week during my walks.

Also, I’m not quite done today–I still have laundry and a couple other chores that will have me up and moving around. So I’ll look at this again tomorrow.