fitbit手表中文说明书

by Yash Soni

由Yash Soni

我如何分析FitBit中的数据以改善整体健康状况 (How I analyzed the data from my FitBit to improve my overall health)

事实证明，数据可以使您保持健康 (It turns out that data can keep you healthy)

Physical activity trackers have become a multimillion-dollar product category. I have had my fair share of fancy trackers, starting early from Nike Fuelband and then MI activity band. Personally, I couldn’t adopt any of them very well and eventually, they all ended up being a fancy digital watch which needed a re-charge every few days in order for them to tell the correct time.

体育活动追踪器已成为价值数百万美元的产品类别。 我有很多不错的追踪器，从耐克Fuelband开始 ，然后是MI活动乐队 。 就我个人而言，我不能很好地采用它们，最终它们最终都变成了精美的电子表，需要每隔几天重新充电才能告诉他们正确的时间。

And then, a few months back, I was gifted a Fitbit Versa by a friend of mine. Something clicked. Why?

然后，几个月前，我的一个朋友给了我一个Fitbit Versa 。 有人点击了。 为什么？

A lot had happened since I gave up the idea of health bands being effective. When I received this gift, I was at a stage where I was glued to a chair and my laptop. As a result, I had a severe problem of back pain and postural imbalance. The wrongdoing of many years could not be fixed in just days or weeks. Fighting an arduous battle against sitting, which is now considered by many as the new ‘smoking’, I became mindful about my posture and the idea of maintaining an active lifestyle.

自从我放弃了健康乐队有效的想法以来，发生了很多事情。 当我收到这份礼物时，我正处于将自己粘在椅子和笔记本电脑上的阶段。 结果，我出现了背部疼痛和姿势不平衡的严重问题。 许多年的错误行为无法在几天或几周内得到解决。 在艰苦的对抗坐姿的斗争中，现在被许多人视为新的“吸烟者” ，我开始意识到自己的姿势和保持积极生活方式的想法。

So, I thought of giving Fitbit a clean slate and tried it for a week. The Fitbit app did a pretty good job of capturing and showcasing the data in a consumable format.

因此，我想到给Fitbit一个干净的选择，并尝试了一个星期。 Fitbit应用程序在以可消耗格式捕获和展示数据方面做得很好。

It had been two weeks since my break up with my mechanical watch. Intrigued by the sheer variety of data points Fitbit could capture, an urge began to see what lies hidden behind the data.

自从我放弃机械表已经过去了两个星期。 Fitbit可以捕获各种各样的数据点，这引起了人们的兴趣，这促使人们开始寻找隐藏在数据背后的东西。

I soon began an experiment to see how well I could keep up with the goals that I had defined. I also wanted to know if there were any extraneous factors which influenced these goals. Finally, I wanted to uncover any hidden and interesting findings along the way. I was particularly interested to discover:

我很快就开始进行一项实验，以了解我能很好地实现我所定义的目标。 我还想知道是否有任何其他因素影响了这些目标。 最后，我想一路发现任何隐藏和有趣的发现。 我特别有兴趣发现：

• How active are my days? Do I spend a considerable amount of time being sedentary?我的日子有多活跃？ 我是否会花费大量时间久坐？
• How does this data vary on weekdays vs weekends?这些数据在工作日与周末之间如何变化？
• What factors contribute to the highest calorie burn?哪些因素导致最高的卡路里消耗？
• Which exercises are the best and easiest way to achieve my daily goals?哪些练习是实现我的日常目标的最佳和最简单的方法？
• Have I been following a steady sleep schedule? What factors influence it?我一直在遵守稳定的睡眠时间表吗？ 哪些因素会影响它？
• Understand the sleep stages and find out what it takes to get a better deep sleep.了解睡眠阶段，并了解如何获得更好的深度睡眠。
• What is the impact of a Netflix binge on weekend sleep?Netflix狂欢对周末睡眠有何影响？
• Train a simple Machine Learning model to see if there is a hidden pattern to attain better sleep.训练一个简单的机器学习模型，看看是否存在隐藏的模式来获得更好的睡眠。

I couldn’t find the answers to these from the standard Fitbit app. I needed the raw data.

我在标准Fitbit应用程序中找不到这些答案。 我需要原始数据。

获取数据 (Getting the Data)

The first task was to figure out ways to extract the data from my device. Scanning through the developer pages, I found that they have a provision of Web API to access the user data. Examining these APIs, I was shocked to see the sheer amount of data that is being captured and saved every minute. Steps covered, calories burned, sleep stages and even heart rate/minute for any-given-day is being recorded!

第一项任务是找出从我的设备中提取数据的方法。 扫描开发人员页面时，我发现它们提供了Web API来访问用户数据。 检查这些API后，我震惊地看到每分钟都会捕获并保存大量的数据。 记录任何步骤的步数，燃烧的卡路里，睡眠阶段，甚至心率/分钟 ！

Sometimes, the seductive lure of knowing about our general wellness makes us forget what personal information we end up sharing. Reading through their privacy policies, I found that Fitbit has put in additional checks to keep the data safe. Anyways this requires a separate post of its own, so without digressing from our main goal, let’s continue.

有时，了解我们的整体健康的诱人诱惑使我们忘记了最终共享的个人信息。 仔细阅读他们的隐私政策，我发现Fitbit进行了其他检查，以确保数据安全。 无论如何，这需要一个单独的职位，因此在不偏离我们主要目标的情况下，让我们继续。

I registered my app and got the necessary client side credentials to begin the data grabbing. After going through the necessary authorization steps, I collected and merged my daily Activity, Sleep and Heart rate data and dumped it in an Excel file. After some data cleansing, the dataset was ready!

我注册了我的应用程序，并获得了必要的客户端凭据才能开始获取数据。 经过必要的授权步骤后，我收集并合并了我的日常“活动”，“睡眠”和“心律”数据，并将其转储到Excel文件中。 经过一些数据清理后，数据集已准备就绪！

PS. The entire code can be found here along with the Jupyter notebook.

PS。 完整的代码可在此处找到，以及Jupyter笔记本 。

PPS. Disclaimer: This data analysis is based on a very limited set of data points and will be difficult to generalise to the masses. Please consider it as a fun read!

PPS。 免责声明：此数据分析基于一组非常有限的数据点，因此很难推广到大众。 请认为这是一本有趣的书！

活动分析 (Activity analysis)

Fitbit has a wide array of data points to measure daily activity levels. Steps, Calories, and floors are some of the standard measures. It also tracks how many minutes I spend daily being moderately, slightly and very active.

Fitbit有各种各样的数据点可以衡量日常活动水平。 台阶，卡路里和地板是一些标准措施。 它还跟踪我每天花多少分钟保持适度，轻微和非常活跃。

Not fussing on the daily calorie burn, I had kept a goal of attaining 8000 steps each day on my Fitbit device. The graphs below suggested I am averaging about ~7800 steps per day which is quite close to my goal. There are some studies which suggest hitting 10000 steps per day is ideal and that shall be the next target.

我不必为每天的卡路里消耗而烦恼，我一直保持着每天在Fitbit设备上达到8000步的目标。 下图显示我平均每天大约要走7800步，这非常接近我的目标。 有一些研究表明，每天达到10000 步是理想的 ，这将是下一个目标。

Tuesdays to Saturdays were the days where I averaged about 40 minutes of very active minutes — which simply translates to active exercise. Fewer minutes on Sundays were due to purely due to laziness/recovery time. The drop in active minutes on Mondays proves that I am falling for the Monday Blues and guess it’s time to fix that. ??

在星期二至星期六，我平均每天要进行大约40分钟的非常活跃的活动-这简直就是积极运动 。 周日更少的时间完全是由于懒惰/恢复时间。 星期一的活动时间减少证明了我对星期一的忧郁症很感兴趣，并且认为是时候解决这个问题。 ??

Analysing the amount of calories burned per minute for various Activities shows some interesting findings. Though there is a lot of similar data available on internet, it is very difficult to generalise these numbers for everyone. Since a lot of this depends on the fitness levels, demographics, skillset and most importantly how much I enjoy doing some specific exercises.

分析各种活动每分钟消耗的卡路里数量，得出一些有趣的发现。 尽管互联网上有很多类似的数据，但是很难为每个人概括这些数字。 由于这很大程度上取决于健身水平，人口统计，技能和最重要的是我喜欢进行一些特定锻炼的程度。

It is interesting to see that running helps me burn almost 12 calories per minute. The math is simple: to compensate for a beer, a 10 minute run is what I need. ??‍+ ? = ?

有趣的是，跑步能帮助我每分钟燃烧近12卡路里。 数学很简单 ：要补偿啤酒，我需要跑10分钟。 ?? +？ =？

Tennis ? — the favourite activity from the lot — takes the second spot. That’s again a win-win scenario! It will be interesting to see if this number changes as I improve my skills.

网球？ -最受欢迎的活动-位居第二。 这又是双赢的局面！ 看看这个数字是否随着我提高自己的技能而改变会很有趣。

Swimming numbers were not shocking to me because I am still struggling to keep up with my continuous lap count. And after spending some time in pool, the exercise turns to a leisure activity.

游泳人数并没有让我感到震惊，因为我仍在努力保持连续的圈数。 在游泳池里玩了一段时间之后，锻炼变成了休闲活动。

Point to note here is that calories burned should not be the only metric on which these activities can be graded. But, this happens to be the only metric which I can currently measure via Fitbit.

这里要注意的是，燃烧的卡路里不应成为对这些活动进行评分的唯一指标。 但是，这恰好是我目前可以通过Fitbit测量的唯一指标。

Lastly, it is useful to see how the various data points correlate with each other. Plotting a correlation heat-map helps uncover some findings.

最后，查看各个数据点如何相互关联是很有用的。 绘制相关热图有助于发现一些发现。

The calories burned is strongly related to amount of steps and active minutes. Minutes sedentary has a negative correlation with Weekdays which implies that I spend more time slacking off on weekends.

燃烧的卡路里与步数和活动分钟数密切相关。 久坐不动的分钟数与工作日负相关，这意味着我在周末花费更多的时间来放松。

睡眠分析 (Sleep Analysis)

Sleep is essential to helping maintain mood, memory, and cognitive performance and there is no running from it. We spend about a third of our life in sleeping. That’s a staggering 26 years spent sleeping in bed! While metabolism generally slows down, all major organs and regulatory systems continue to function. Hence, it becomes important to get the most out of our sleep.

睡眠对于维持情绪，记忆和认知能力至关重要，因此并不能避免睡眠。 我们在睡眠中度过了大约三分之一的时间。 在床上睡觉真是花了26年 ！ 虽然新陈代谢通常会减慢，但所有主要器官和调节系统仍会继续起作用。 因此，充分利用睡眠变得很重要。

Reading more on this, I found that there are some standard ways which can help achieve a good night sleep.

通过内容，我发现有一些标准方法可以帮助您睡个好觉。

• Following a good sleep schedule

  遵循良好的睡眠时间表

• Avoid bright/blue light at night before hitting the bed

  晚上在睡觉前避免明亮/蓝色的灯光

• Avoid caffeine later in the day当天晚些时候避免咖啡因
• Sleep in a cool and dark room在阴凉的房间里睡觉

• Getting at-least 7–9 hours of sleep. There are some studies which say that even in 5 hours you can achieve the most out of your sleep.

  至少睡眠7–9个小时 。 有研究表明，即使在5小时内，您也可以充分利用睡眠。

Over the course of this experiment, I tried to follow the above steps to bind myself to a strict sleep schedule. It was time to validate them.

在这个实验过程中，我尝试按照上述步骤将自己绑定到严格的睡眠时间表。 是时候验证它们了。

From the graphs below, I found that I was averaging a sleep of 7 hours without much deviation in the numbers. Though I was able to hit bed before 11, the wakeup timings still ranged from 5:30–7:00 AM.

从下面的图表中，我发现我平均睡眠7个小时，而数字没有太大的偏差。 尽管我可以在11点之前上床睡觉，但唤醒时间仍在5：30–7：00 AM之间。

Though the average duration was somewhat similar, the overall sleep quality was not the same. On some days, I was very active even by attaining 6 hours of sleep, while there have been many instances when even after sleeping late, I wasn’t feeling fresh. I found the answer by analysing the mystic sleep cycles.

尽管平均持续时间有些相似，但总体睡眠质量却不同。 有时候，即使达到6个小时的睡眠，我仍然非常活跃，而在很多情况下，即使睡到很晚，我也不感到新鲜。 我通过分析神秘的睡眠周期找到了答案。

While we are asleep, our body typically goes through several sleep cycles, alternating between the following stages:

当我们入睡时，我们的身体通常经历几个睡眠周期，在以下阶段之间交替进行：

LIGHT SLEEP: This stage typically begins within minutes of falling asleep. Breathing and heart rate typically decrease slightly during this stage. Light sleep promotes mental and physical recovery.

轻睡眠：此阶段通常在入睡的几分钟内开始。 在此阶段，呼吸和心率通常会略有下降。 轻度睡眠可促进身心恢复。

DEEP SLEEP: Deep sleep typically occurs in the first few hours of sleep. Breathing becomes slower and muscles relax while heart rate usually becomes more regular. When we wake up feeling refreshed in the morning, it is likely that we have experienced solid periods of deep sleep. Deep sleep promotes physical recovery and aspects of memory and learning.

深度睡眠：深度睡眠通常发生在睡眠的最初几个小时。 呼吸变慢，肌肉放松，而心率通常变得更规律。 当我们早上醒来时感到神清气爽，这很可能是我们经历了深沉的睡眠。 深度睡眠促进身体恢复以及记忆和学习的各个方面。

REM SLEEP: REM sleep is an active period of sleep marked by intense brain activity. The first phase of REM sleep typically occurs after an initial stage of deep sleep. Breathing is more rapid, irregular and shallow. The eyes move rapidly in various directions, hence the name Rapid Eye Movement — REM Sleep. This is the stage where we generally see dreams in our sleep. REM sleep has been shown to play an important role in mood regulation, learning, and memory.

REM睡眠： REM睡眠是活跃的睡眠时间，以剧烈的脑部活动为标志。 REM睡眠的第一阶段通常发生在深度睡眠的初始阶段之后。 呼吸较快，不规则且浅。 眼睛向各个方向快速移动，因此名称为快速眼动-REM睡眠。 这是我们通常在梦中看到梦境的阶段。 REM睡眠已被证明在情绪调节，学习和记忆中起重要作用。

The plot below shows that on an average, my body spends just about 17% in Deep sleep, 19% REM and the rest in either light or being slightly awake. The date-time plot of Light and Deep sleep shows that these numbers vary a lot.

下图显示，平均而言，我的身体在深度睡眠上的花费约为17％，在快速眼动方面的花费为19％，其余部分则在光线充足或略微醒着的时候花费。 浅睡眠和深度睡眠的日期时间图显示，这些数字相差很大。

If we plot the correlation of different sleep stages, we see that time spent in bed is highly correlated with Light sleep but there is no strong correlation with Deep sleep.

如果我们绘制不同睡眠阶段的相关性图，我们会发现在床上度过的时间与轻度睡眠高度相关，而与深度睡眠却没有强相关性。

This essentially means that just by sleeping more doesn’t always guarantee a good deep sleep. I guess this helps validate the important learning about sleep:

从本质上讲，这意味着仅睡得更多并不总是可以保证深度睡眠。 我想这有助于验证关于睡眠的重要学习：

It is the quality of sleep that matters and not the quantity.

重要的是睡眠质量，而不是数量。

Following a strict sleep schedule on weekdays is easy, but weekends are a different ballgame altogether.

在工作日遵循严格的睡眠时间表很容易，但是周末则完全不同。

The boxplot above shows Saturdays are the most affected, where the time in bed ranges from 5–9 hours. The Netflix Binge and weekend parties are some of the vices which have affected this routine. On the contrary, the smaller boxplot on Sundays depicts me gearing up for the Monday mornings. It’s interesting to see how these subconscious body behaviours are clearly exposed in these plots.

上面的箱线图显示，周六受影响最严重，躺在床上的时间为5–9小时。 Netflix狂欢和周末派对是影响此套路的一些弊端。 相反，周日较小的箱线图描绘了我为周一早晨做准备。 有趣的是，在这些图中如何清晰地暴露了这些潜意识的身体行为。

Finally, I wanted to see if my daily activities had any effect on my sleep. Though I didn’t have much data for the Machine Learning model,an initial run showed some interesting results. It predicted that being active in the day and getting to bed before 11PM has some positive contributions to the final deep sleep minutes.

最后，我想看看我的日常活动是否对我的睡眠有影响。 尽管我没有太多有关机器学习模型的数据，但初步运行显示了一些有趣的结果。 它预测白天活跃并在晚上11点之前入睡对最终的深度睡眠时间有积极的作用。

Though it is too soon to validate it, I will be repeating this again once I have some more sleep data and additional features to improve the accuracy of the model. The details can be found in this Jupyter Notebook.

尽管现在进行验证尚为时过早，但是一旦获得更多睡眠数据和其他功能以提高模型的准确性，我将再次重复此过程。 可以在此Jupyter Notebook中找到详细信息。

值得吗？ (Was it all worth it?)

This experiment has been a rewarding experience. I identified some interesting ways in which my body responds to external stimuli. It is like a machine where tweaking certain knobs can help achieve different results.

这个实验是有益的经历。 我发现了一些有趣的方式，可以让我的身体对外界刺激做出React。 就像调整某些旋钮可以帮助获得不同结果的机器一样。

Next up in line I plan to set up some improved activity goals. I will also be trying out some biohacks to see if these have any positive effects on my sleep quality. I am also considering developing a Fitbit alarm app which only wakes me up once I have acquired sufficient amount of quality sleep (not sure if this already exists?).

接下来，我计划设置一些改进的活动目标。 我还将尝试一些 生物黑客技术 ，看看它们是否会对我的睡眠质量产生积极影响。 我也在考虑开发一个Fitbit警报应用程序，该应用程序只会在我获得足够的优质睡眠后才能唤醒我(不确定是否已经存在？)。

Finally, I do not intend to label this as an experiment anymore. The daily routine which felt forceful in the beginning has become a habit now. In the past, I’d come across many articles that were studies of the importance of waking up early as well, and finally I embraced it first-hand. This Medium article, which is in fact my first one, is one of the many byproducts of this new found habit.

最后，我不想再将其标记为实验了。 一开始就觉得有力的日常活动现在已经成为一种习惯。 过去，我遇到过许多文章，这些文章也研究过早起床的重要性，最后我亲身接受了它。 这篇中篇文章，实际上是我的第一篇文章，是这种新发现习惯的许多副产品之一。

Thanks for taking the time to go through my analyses. I will be more than thankful to know if you liked it and have suggestion about any improvements in general! :)

感谢您抽出宝贵的时间进行分析。 如果您喜欢它，我将不胜感激，并希望就总体上的改进提出建议！ :)

翻译自: https://www.freecodecamp.org/news/how-i-analyzed-the-data-from-my-fitbit-to-improve-my-overall-health-a2e36426d8f9/

fitbit手表中文说明书