你睡眠情况怎么样？Jawbone Up 数据，R 实现

身边很多朋友都对 Jawbone Up 对睡眠监控的情况很感兴趣，一般会顺带问问：

“你睡眠情况怎么样？”。

“好吧，俺计算一下再告诉你。”

下面的文中涉及到如何用 R 画基于日历的热度图，以及如何将多个图进行合并的内容。

数据获取

需要使用的数据是每日的睡眠情况数据。

要获取这个数据有两个办法，最简单的办法是从 Jawbone.com 网站下载个人的数据文件。详情参见个人数据官方下载 。

还有一种办法是利用 非官方API 中的 user/%userXid%/trends 方法获取按日统计数据。

下面的代码采用 Python 实现，用于从服务器端获取 Trends 数据。

account = { "email" : "your@email" , "passwd" : "yourpassword" } device = DeviceJawboneUp () # login device.get_users_login(account["email"], account["passwd"]) # Trends daytrends = device.get_users_trends( endDate = None , inRange = "d" , bucketSize = "d" ) device.saveJsonData( filename = "/users_trends-day.json" , data = daytrends)

1 2 3 4 5 6 7 8 9 10 11

account = { "email" : "your@email" , "passwd" : "yourpassword" }   device = DeviceJawboneUp ( )   # login device . get_users_login ( account [ "email" ] , account [ "passwd" ] )   # Trends daytrends = device . get_users_trends ( endDate =    None , inRange = "d" , bucketSize = "d"    ) device . saveJsonData ( filename = "/users_trends-day.json" , data = daytrends )

返回的JSON数据示例如下：

{ "meta": { "code": 200, "message": "OK", "user_xid": "RGaCBFg9CsDYVvm2xxxxxx", "time": 1374161411 }, "data": { "earliest": 20130605, "data": [ [ 20130605, { "e_protein": null, "weight": 57.0, "m_distance": 270.0, "s_awakenings": null, "height": 1.58, "m_lcat": 88.0, "n_awakenings": null, "s_quality": null, "n_deep": null, "m_steps": 367.0, "e_cholesterol": null, "s_light": null, "s_deep": null, "s_bedtime": null, "n_bedtime": null, "m_workout_time": 0.0, "e_calcium": null, "e_sat_fat": null, "n_duration": null, "n_light": null, "e_calories": null, "pal": null, "m_lcit": 0.0, "s_awake_time": null, "e_unsat_fat": null, "m_calories": 12.3110000491, "bmr": 1319.38485068, "m_total_calories": 1331.69585073, "e_sugar": null, "e_sodium": null, "s_awake": null, "s_asleep_time": null, "s_duration": null, "gender": 1.0, "age": 37.0136986301, "e_carbs": null, "e_fiber": null, "n_quality": null, "n_awake_time": null, "n_awake": null, "n_asleep_time": null, "m_active_time": 201.0 } ] ], "links": { "next": "/nudge/api/v.1.34/users/RGaCBFg9CsDYVvm2kchbcw/trends/?end_date=20110724&range=d&range_duration=730&bucket_size=d" } } }

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64

{      "meta" : {          "code" : 200 ,          "message" : "OK" ,          "user_xid" : "RGaCBFg9CsDYVvm2xxxxxx" ,          "time" : 1374161411      } ,      "data" : {          "earliest" : 20130605 ,          "data" : [              [                  20130605 ,                  {                      "e_protein" : null ,                      "weight" : 57.0 ,                      "m_distance" : 270.0 ,                      "s_awakenings" : null ,                      "height" : 1.58 ,                      "m_lcat" : 88.0 ,                      "n_awakenings" : null ,                      "s_quality" : null ,                      "n_deep" : null ,                      "m_steps" : 367.0 ,                      "e_cholesterol" : null ,                      "s_light" : null ,                      "s_deep" : null ,                      "s_bedtime" : null ,                      "n_bedtime" : null ,                      "m_workout_time" : 0.0 ,                      "e_calcium" : null ,                      "e_sat_fat" : null ,                      "n_duration" : null ,                      "n_light" : null ,                      "e_calories" : null ,                      "pal" : null ,                      "m_lcit" : 0.0 ,                      "s_awake_time" : null ,                      "e_unsat_fat" : null ,                      "m_calories" : 12.3110000491 ,                      "bmr" : 1319.38485068 ,                      "m_total_calories" : 1331.69585073 ,                      "e_sugar" : null ,                      "e_sodium" : null ,                      "s_awake" : null ,                      "s_asleep_time" : null ,                      "s_duration" : null ,                      "gender" : 1.0 ,                      "age" : 37.0136986301 ,                      "e_carbs" : null ,                      "e_fiber" : null ,                      "n_quality" : null ,                      "n_awake_time" : null ,                      "n_awake" : null ,                      "n_asleep_time" : null ,                      "m_active_time" : 201.0                  }              ]          ] ,          "links" : {              "next" : "/nudge/api/v.1.34/users/RGaCBFg9CsDYVvm2kchbcw/trends/?end_date=20110724&range=d&range_duration=730&bucket_size=d"          }      } }

---

数据预处理

下面的代码，采用 Python 实现了将 Trends 数据转换成平面CSV文件。

def list2string(list ): ret = "" for l in list: ret = "%s,%s" % (ret , l) return "%s\n" % ret[1:] def parseJawboneTrends(self , path , filename ): fn = "%s%s.json" % ( path , filename ) ofn = "%s.csv" % ( filename ) input = open( fn , 'r' ) str = input.read() v = json.loads(str) output = open( ofn , 'w' ) datas = v["data"]["data"] tmps = datas[0][1] headerKeys = sorted(tmps , key=lambda tmps:tmps[0] ) header = "datestr,%s" % self.list2string(headerKeys) print header output.write( header ) for d in datas: tmps = d[1] stmp = "" for key in headerKeys: stmp = "%s,%s" % ( stmp , tmps[ key ]) str = "%s,%s\n" % (d[0] , stmp[1:]) print str output.write( str ) input.close() output.close() if __name__ == "__main__": datapath = "../data/jawboneup/" parseJawboneTrends( path = datapath , filename = "users_trends-day" )

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40

def list2string ( list ) :      ret = ""      for l in list :          ret = "%s,%s" % ( ret , l )      return "%s\n" % ret [ 1 : ]   def parseJawboneTrends ( self , path , filename ) :      fn = "%s%s.json" % ( path , filename )      ofn = "%s.csv" % ( filename )        input = open ( fn , 'r' )      str = input . read ( )      v = json . loads ( str )        output = open ( ofn , 'w' )      datas = v [ "data" ] [ "data" ]        tmps = datas [ 0 ] [ 1 ]      headerKeys = sorted ( tmps , key = lambda tmps : tmps [ 0 ] )      header = "datestr,%s" % self . list2string ( headerKeys )      print header      output . write ( header )        for d in datas :          tmps = d [ 1 ]          stmp = ""          for key in headerKeys :              stmp = "%s,%s" % ( stmp , tmps [ key ] )            str = "%s,%s\n" % ( d [ 0 ] ,    stmp [ 1 : ] )          print str          output . write ( str )        input . close ( )      output . close ( )   if __name__ == "__main__" :      datapath = "../data/jawboneup/"      parseJawboneTrends ( path = datapath , filename = "users_trends-day" )

处理后的数据文件样本为：

datestr,age,bmr,e_protein,e_calories,e_cholesterol,e_sat_fat,e_calcium,e_unsat_fat,e_sugar,e_sodium,e_carbs,e_fiber,gender,height,m_distance,m_lcat,m_workout_time,m_lcit,m_active_time,m_calories,m_total_calories,m_steps,n_awakenings,n_deep,n_bedtime,n_duration,n_light,n_quality,n_awake_time,n_awake,n_asleep_time,pal,s_awakenings,s_quality,s_light,s_deep,s_bedtime,s_awake_time,s_awake,s_asleep_time,s_duration,weight 20130605,37.0136986301,1319.38485068,None,None,None,None,None,None,None,None,None,None,1.0,1.58,270.0,88.0,0.0,0.0,201.0,12.3110000491,1331.69585073,367.0,None,None,None,None,None,None,None,None,None,None,None,None,None,None,None,None,None,None,None,57.0 20130606,37.0164383562,1319.37204082,None,None,None,None,None,None,None,None,None,None,1.0,1.58,8561.0,1129.0,0.0,3480.0,6412.0,336.044000283,1655.4160411,12339.0,None,None,None,None,None,None,None,None,None,None,0.0,71.0,8742.0,11409.0,14.0,374.0,2082.0,41.0,20151.0,57.0 20130607,37.0191780822,1319.35923096,None,None,None,None,None,None,None,None,None,None,1.0,1.58,8818.0,1450.0,0.0,3660.0,7040.0,346.185998712,1665.54522967,13030.0,0.0,2580.0,1317.0,5342.0,2762.0,18.0,1422.0,1052.0,1333.0,None,2.0,64.0,14160.0,7225.0,1418.0,378.0,3297.0,1429.0,21385.0,57.0 20130608,37.0219178082,1319.3464211,None,None,None,None,None,None,None,None,None,None,1.0,1.58,4644.0,645.0,0.0,5640.0,3472.0,183.414000019,1502.76042111,6693.0,None,None,None,None,None,None,None,None,None,None,0.0,0.0,0.0,0.0,1410.0,390.0,0.0,1410.0,25200.0,57.0 20130609,37.0246575342,1319.33361123,15.3100000619,232.0,460.0,4.5640001595,70.0,9.69599956274,0.939999982715,663.0,4.76000010967,0.300000011921,1.0,1.58,7965.0,1662.0,1800.0,7260.0,7917.0,375.598999478,1694.93261071,11611.0,None,None,None,None,None,None,None,None,None,None,2.0,99.0,12406.0,16136.0,1317.0,372.0,2189.0,1321.0,28542.0,57.0

1 2 3 4 5 6 7

datestr , age , bmr , e_protein , e_calories , e_cholesterol , e_sat_fat , e_calcium , e_unsat_fat , e_sugar , e_sodium , e_carbs , e_fiber , gender , height , m_distance , m_lcat , m_workout_time , m_lcit , m_active_time , m_calories , m_total_calories , m_steps , n_awakenings , n_deep , n_bedtime , n_duration , n_light , n_quality , n_awake_time , n_awake , n_asleep_time , pal , s_awakenings , s_quality , s_light , s_deep , s_bedtime , s_awake_time , s_awake , s_asleep_time , s_duration , weight 20130605 , 37.0136986301 , 1319.38485068 , None , None , None , None , None , None , None , None , None , None , 1.0 , 1.58 , 270.0 , 88.0 , 0.0 , 0.0 , 201.0 , 12.3110000491 , 1331.69585073 , 367.0 , None , None , None , None , None , None , None , None , None , None , None , None , None , None , None , None , None , None , None , 57.0 20130606 , 37.0164383562 , 1319.37204082 , None , None , None , None , None , None , None , None , None , None , 1.0 , 1.58 , 8561.0 , 1129.0 , 0.0 , 3480.0 , 6412.0 , 336.044000283 , 1655.4160411 , 12339.0 , None , None , None , None , None , None , None , None , None , None , 0.0 , 71.0 , 8742.0 , 11409.0 , 14.0 , 374.0 , 2082.0 , 41.0 , 20151.0 , 57.0 20130607 , 37.0191780822 , 1319.35923096 , None , None , None , None , None , None , None , None , None , None , 1.0 , 1.58 , 8818.0 , 1450.0 , 0.0 , 3660.0 , 7040.0 , 346.185998712 , 1665.54522967 , 13030.0 , 0.0 , 2580.0 , 1317.0 , 5342.0 , 2762.0 , 18.0 , 1422.0 , 1052.0 , 1333.0 , None , 2.0 , 64.0 , 14160.0 , 7225.0 , 1418.0 , 378.0 , 3297.0 , 1429.0 , 21385.0 , 57.0 20130608 , 37.0219178082 , 1319.3464211 , None , None , None , None , None , None , None , None , None , None , 1.0 , 1.58 , 4644.0 , 645.0 , 0.0 , 5640.0 , 3472.0 , 183.414000019 , 1502.76042111 , 6693.0 , None , None , None , None , None , None , None , None , None , None , 0.0 , 0.0 , 0.0 , 0.0 , 1410.0 , 390.0 , 0.0 , 1410.0 , 25200.0 , 57.0 20130609 , 37.0246575342 , 1319.33361123 , 15.3100000619 , 232.0 , 460.0 , 4.5640001595 , 70.0 , 9.69599956274 , 0.939999982715 , 663.0 , 4.76000010967 , 0.300000011921 , 1.0 , 1.58 , 7965.0 , 1662.0 , 1800.0 , 7260.0 , 7917.0 , 375.598999478 , 1694.93261071 , 11611.0 , None , None , None , None , None , None , None , None , None , None , 2.0 , 99.0 , 12406.0 , 16136.0 , 1317.0 , 372.0 , 2189.0 , 1321.0 , 28542.0 , 57.0

---

利用 R 进行数据统计和可视化

下面的代码实现了数据的读取、预处理、并构建了基于日历的热度图（calendarHeatmap）

# 设置工作路径 setwd("D:\\PythonWorkspaces\\VisHealth\\R") # 读取 CSV 文件 trends.daily <- read.csv("users_trends-day.csv" , as.is=TRUE) # 仅选取所需的数据列 names( trends.daily ) sleeps <- trends.daily[ , c( "datestr" , "s_light" , "s_deep" , "s_duration" ) ] # 数据清洗 sleeps.clean <- subset( sleeps , ! (is.na( s_duration ) | is.na( s_deep ) | ( s_duration == 0) | ( s_deep == 0) | ( s_deep == "None" ) | ( s_duration == "None" ) ) ) # 格式转换 sleeps.clean$s_light <- as.numeric( sleeps.clean$s_light ) sleeps.clean$s_deep <- as.numeric( sleeps.clean$s_deep ) sleeps.clean$s_duration <- as.numeric( sleeps.clean$s_duration ) # 去掉忘记切换睡眠模式的数据 sleeps.clean <- subset( sleeps.clean , s_light > 0.1 ) # 睡眠质量取值 = 深度睡眠时长 / 完整睡眠时长 sleeps.clean$q <- sleeps.clean$s_deep / sleeps.clean$s_duration # 设置做图所需的日期变量 sleeps.clean$date <- strptime ( sleeps.clean$datestr , format="%Y%m%d" )

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

# 设置工作路径 setwd ( "D:\\PythonWorkspaces\\VisHealth\\R" )   # 读取 CSV 文件 trends . daily & lt ; - read . csv ( "users_trends-day.csv" , as . is = TRUE )   # 仅选取所需的数据列 names ( trends . daily ) sleeps & lt ; - trends . daily [ , c ( "datestr" , "s_light" , "s_deep" , "s_duration" ) ]   # 数据清洗 sleeps . clean & lt ; - subset ( sleeps , ! ( is . na ( s_duration ) | is . na ( s_deep ) | ( s_duration == 0 )      | ( s_deep == 0 ) | ( s_deep == "None" ) | ( s_duration == "None" ) ) )   # 格式转换 sleeps . clean $ s_light & lt ; - as . numeric ( sleeps . clean $ s_light ) sleeps . clean $ s_deep & lt ; - as . numeric ( sleeps . clean $ s_deep ) sleeps . clean $ s_duration & lt ; - as . numeric ( sleeps . clean $ s_duration )   # 去掉忘记切换睡眠模式的数据 sleeps . clean & lt ; - subset ( sleeps . clean , s_light & gt ; 0.1    )   # 睡眠质量取值 = 深度睡眠时长 / 完整睡眠时长 sleeps . clean $ q & lt ; - sleeps . clean $ s_deep / sleeps . clean $ s_duration   # 设置做图所需的日期变量 sleeps . clean $ date & lt ; - strptime ( sleeps . clean $ datestr , format = "%Y%m%d" )

数据显示样本：

> head(sleeps.clean , 10) datestr s_light s_deep s_duration q date 2 20130606 8742 11409 20151 0.5661754 2013-06-06 3 20130607 14160 7225 21385 0.3378536 2013-06-07 5 20130609 12406 16136 28542 0.5653423 2013-06-09 6 20130610 13966 10080 24046 0.4191965 2013-06-10 7 20130611 11393 4408 15801 0.2789697 2013-06-11 8 20130612 20652 15093 35745 0.4222409 2013-06-12 9 20130613 9480 11118 20598 0.5397611 2013-06-13 10 20130614 8840 10967 19807 0.5536931 2013-06-14 > summary( sleeps.clean[,2:5] ) s_light s_deep s_duration q Min. : 6300 Min. : 3613 Min. :11235 Min. :0.2052 1st Qu.:11331 1st Qu.: 6728 1st Qu.:19720 1st Qu.:0.3420 Median :13070 Median : 8528 Median :21283 Median :0.3982 Mean :13153 Mean : 9268 Mean :22421 Mean :0.4065 3rd Qu.:14430 3rd Qu.:11191 3rd Qu.:24744 3rd Qu.:0.4814 Max. :21180 Max. :16740 Max. :35745 Max. :0.6042

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

& gt ; head ( sleeps . clean , 10 )      datestr s_light s_deep s_duration          q        date 2    20130606      8742    11409        20151 0.5661754 2013 - 06 - 06 3    20130607    14160    7225        21385 0.3378536 2013 - 06 - 07 5    20130609    12406    16136        28542 0.5653423 2013 - 06 - 09 6    20130610    13966    10080        24046 0.4191965 2013 - 06 - 10 7    20130611    11393    4408        15801 0.2789697 2013 - 06 - 11 8    20130612    20652    15093        35745 0.4222409 2013 - 06 - 12 9    20130613      9480    11118        20598 0.5397611 2013 - 06 - 13 10 20130614      8840    10967        19807 0.5536931 2013 - 06 - 14   & gt ; summary ( sleeps . clean [ , 2 : 5 ] )      s_light           s_deep         s_duration            q          Min .    : 6300    Min .    : 3613    Min .    : 11235    Min .    : 0.2052    1st Qu . : 11331    1st Qu . : 6728    1st Qu . : 19720    1st Qu . : 0.3420    Median : 13070    Median : 8528    Median : 21283    Median : 0.3982    Mean    : 13153    Mean    : 9268    Mean    : 22421    Mean    : 0.4065    3rd Qu . : 14430    3rd Qu . : 11191    3rd Qu . : 24744    3rd Qu . : 0.4814    Max .    : 21180    Max .    : 16740    Max .    : 35745    Max .    : 0.6042

calendarHeatmap 的实现和使用

calendarHeatmap 将按日聚合数据以热度图的形式展现在日历上。用户可以指定聚合所使用的计算方法。

labelMon <- c("JAN","FEB","MAR","APR","MAY","JUN","JUL", "AUG","SEP","OCT","NOV","DEC") # labelMon <- c("一月","二月","三月","四月","五月","六月","七月", "八月","九月","十月","十一月","十二月") labelWeek <- c("Mon", "Tue", "Wed", "Thu", "Fri", "Sat" , "Sun") # labelWeek <- c("周一","周二","周三","周四","周五","周六","周日") # ================================ # calander heatmap # 参数说明 # data， 所需展现的数据框 # dateCol， 用于标识时间的列名，缺省为“date”，可以根据您的数据指定。此列的数据格式为 POSIXlt 格式日期。 # valueCol， 需要进行统计的列名，此列的数据格式需要是数字。 # statics， 统计分析函数方法，用于对 valueCol 进行计算，如：sum、mean、sd 等均可。 # islog2， 是否对数据进行 log2 操作，此参数主要用于更好的可视化符合幂律分布的数据 # picTitle， 图标标题，缺省会根据此图标的 valueCol 等信息自动生成 # isguide， 是否显示图例 calendarHeatmap <- function( data , dateCol = "date" , valueCol = "steps" , statics = "sum", islog2 = FALSE , picTitle = NA , isguide = FALSE ){ require("ggplot2") require("reshape2") data["daystr"] <- format ( data[, dateCol] , "%Y-%m-%d") data.mtx <- tapply(data[ , valueCol ] , list( data$daystr ) , statics) data.mtx [is.na(data.mtx )]= 0 if( islog2 ){ data.mtx[ data.mtx == 0 ] = 1 data.mtx <- log(data.mtx , 2) } m <- melt( data.mtx ) names(m) <- c( "daystr" , "values" ) m$date <- strptime ( m$daystr , format="%Y-%m-%d" ) m$mday = m$date$mday m$month <- m$date$mon + 1 m$monthf <- factor(m$month , levels=as.character(1:12), labels=labelMon, ordered=TRUE) m$weekday = m$date$wday m$weekday[ m$weekday == 0 ] = 7 m$weekdayf <- factor( m$weekday ,levels=c(1:7), labels=labelWeek,ordered=TRUE) m$week <- as.numeric(format( m$date ,"%W")) print(head(m)) pic <- ggplot(m, aes( weekdayf, week, fill = values)) + geom_tile(color="white", size= 1 ) + # facet_wrap(~ monthf ,nrow=3) + theme_bw() + xlab("") + ylab("Week of Year") + scale_y_reverse() + geom_text(aes(label = round( mday , 2 ) ) , vjust = 1) if( isguide ) pic <- pic + scale_fill_gradient(limits=c(0, max(m$values)), low='white', high='red') else pic <- pic + scale_fill_gradient(limits=c(0, max(m$values)), low='white', high='red' , guide="none") if( ! is.na(picTitle) ) pic <- pic + ggtitle( picTitle ) pic }

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69

labelMon & lt ; - c ( "JAN" , "FEB" , "MAR" , "APR" , "MAY" , "JUN" , "JUL" , "AUG" , "SEP" , "OCT" , "NOV" , "DEC" ) # labelMon <- c("一月","二月","三月","四月","五月","六月","七月", "八月","九月","十月","十一月","十二月") labelWeek & lt ; - c ( "Mon" , "Tue" , "Wed" , "Thu" , "Fri" , "Sat" , "Sun" ) # labelWeek <- c("周一","周二","周三","周四","周五","周六","周日")   # ================================ # calander heatmap # 参数说明 # data， 所需展现的数据框 # dateCol， 用于标识时间的列名，缺省为“date”，可以根据您的数据指定。此列的数据格式为 POSIXlt 格式日期。 # valueCol， 需要进行统计的列名，此列的数据格式需要是数字。 # statics， 统计分析函数方法，用于对 valueCol 进行计算，如：sum、mean、sd 等均可。     # islog2， 是否对数据进行 log2 操作，此参数主要用于更好的可视化符合幂律分布的数据 # picTitle， 图标标题，缺省会根据此图标的 valueCol 等信息自动生成 # isguide， 是否显示图例   calendarHeatmap & lt ; - function ( data , dateCol = "date" , valueCol = "steps" , statics = "sum" ,      islog2 = FALSE , picTitle = NA , isguide = FALSE ) {      require ( "ggplot2" )      require ( "reshape2" )        data [ "daystr" ] & lt ; - format ( data [ , dateCol ] , "%Y-%m-%d" )        data . mtx & lt ; - tapply ( data [ , valueCol ] , list ( data $ daystr ) , statics )      data . mtx [ is . na ( data . mtx ) ] = 0        if ( islog2 ) {          data . mtx [ data . mtx == 0 ] = 1          data . mtx & lt ; - log ( data . mtx , 2 )      }        m & lt ; - melt ( data . mtx )      names ( m ) & lt ; - c ( "daystr" , "values" )        m $ date & lt ; - strptime ( m $ daystr , format = "%Y-%m-%d" )        m $ mday = m $ date $ mday        m $ month & lt ; - m $ date $ mon + 1      m $ monthf & lt ; - factor ( m $ month , levels = as . character ( 1 : 12 ) , labels = labelMon , ordered = TRUE )        m $ weekday = m $ date $ wday      m $ weekday [ m $ weekday == 0 ] = 7      m $ weekdayf & lt ; - factor ( m $ weekday , levels = c ( 1 : 7 ) , labels = labelWeek , ordered = TRUE )        m $ week & lt ; -    as . numeric ( format ( m $ date , "%W" ) )        print ( head ( m ) )        pic & lt ; - ggplot ( m , aes ( weekdayf , week , fill = values ) ) +          geom_tile ( color = "white" , size = 1 ) +          # facet_wrap(~ monthf ,nrow=3) +          theme_bw ( ) +          xlab ( "" ) +          ylab ( "Week of Year" ) +          scale_y_reverse ( ) +          geom_text ( aes ( label = round ( mday , 2 ) ) , vjust = 1 )        if ( isguide )          pic & lt ; - pic + scale_fill_gradient ( limits = c ( 0 , max ( m $ values ) ) , low = 'white' , high = 'red' )      else          pic & lt ; - pic + scale_fill_gradient ( limits = c ( 0 , max ( m $ values ) ) , low = 'white' , high = 'red' , guide = "none" )        if ( ! is . na ( picTitle ) )          pic & lt ; - pic + ggtitle ( picTitle )        pic }

对每日睡眠质量的可视化展现。

pic8 <- calendarHeatmap( sleeps.clean , valueCol = "q", picTitle = "Heatmap" , islog2 = FALSE , isguide = TRUE) ggsave(plot = pic8, filename = "sleep_q_calendarheatmap.png")

1 2 3

pic8 & lt ; - calendarHeatmap (    sleeps . clean , valueCol = "q" , picTitle = "Heatmap" , islog2 = FALSE , isguide = TRUE ) ggsave ( plot = pic8 , filename = "sleep_q_calendarheatmap.png" )

多图合并

将睡眠质量的柱状图(参见Jawbone Up 数据的按小时统计及可视化)、数据Summary，和基于日历的热度图合并在一张图上，以更全面的描述睡眠质量：

library(gridExtra) g1 <- tableGrob(summary( sleeps.clean[,2:5], digits = 2 )) pic7 <- bucketBar( sleeps.clean , valueCol = "q" , bucketSize = "mday", statics = "mean" , picTitle = "Days in Month") g2 <- arrangeGrob(g1, pic7, ncol=2) png(file = "sleep_q_all.png", bg = "transparent" , width = 800, height = 800, units = "px") grid.arrange(pic8, g2, ncol=1, main="Sleeps") dev.off()

1 2 3 4 5 6 7 8 9 10

library ( gridExtra ) g1 & lt ; - tableGrob ( summary ( sleeps . clean [ , 2 : 5 ] , digits = 2 ) )   pic7 & lt ; - bucketBar ( sleeps . clean , valueCol = "q" , bucketSize = "mday" , statics = "mean" , picTitle = "Days in Month" ) g2 & lt ; - arrangeGrob ( g1 , pic7 , ncol = 2 )   png ( file = "sleep_q_all.png" , bg = "transparent" , width = 800 , height = 800 , units = "px" )      grid . arrange ( pic8 , g2 ,    ncol = 1 , main = "Sleeps" ) dev . off ( )

---

代码地址

https://github.com/iascchen/VisHealth/

数据预处理 R/json2csv.py; 可视化展现 R/vischarts.R 。

---

玩的开心！

---

Author : iascchen(at)gmail(dot)com

Date : 2013-08-05

新浪微博 : @问天鼓

原文链接：http://www.wearable.pw/index.php/archives/281