initial

Jawbone/features.csv

@@ -0,0 +1,81 @@
+Field,Definition
+DATE,Date of column data (in year/month/day format). 
+age,Your age on the indicated date. 
+avg_bg,Average passive heart rate (in beats per minute).
+bmr,Basal Metabolic Rate (in calories). 
+body_fat,Body fat percentage. 
+e_caffeine,Caffeine content consumed (in milligrams).
+e_calcium,Calcium content consumed (in milligrams). 
+e_calories,Calories consumed. 
+e_carbs,Carbohydrates consumed (in grams). 
+e_cholesterol,Cholesterol consumed (in milligrams). 
+e_count,Number of meals logged. 
+e_fat,Fat consumed (in grams). 
+e_fiber,Fiber consumed (in grams). 
+e_iron,Percentage of recommended iron consumed (based on a 2000 calorie per day diet). 
+e_monounsaturated_fat,Monounsaturated fat consumed (in grams). 
+e_num_drinks,Number of drinks logged. 
+e_num_foods,Number of meal items logged. 
+e_num_mealitems_green,Number of meal items logged with a green UP Food Score.
+e_num_mealitems_red,Number of meal items logged with a red UP Food Score. 
+e_num_mealitems_with_score,Number of meal items logged with an UP Food Score. 
+e_num_mealitems_yellow,Number of meal items logged with a yellow UP Food Score. 
+e_num_water,Glasses of water logged. 
+e_polyunsaturated_fat,Polyunsaturated fat consumed (in grams). 
+e_potassium,Potassium consumed (in milligrams). 
+e_protein,Protein consumed (in grams). 
+e_sat_fat,Saturated fat consumed (in grams). 
+e_sodium,Sodium consumed (in milligrams). 
+e_sugar,Sugar consumed (in grams). 
+e_trans_fat,Trans fat consumed (in grams). 
+e_unsat_fat,Unsaturated fat consumed (in grams). 
+e_vitamin_a,Percentage of recommended Vitamin A consumed (based on a 2000 calorie per day diet). 
+e_vitamin_c,Percentage of recommended Vitamin C consumed (based on a 2000 calorie per day diet). 
+gender,Your specified gender (0=male 1=female). 
+goal_body_weight,Weight goal (in kilograms). 
+goal_body_weight_intent,Weight goal preference (0=lose 1=maintain 2=gain). 
+height,Your specified height (in meters). 
+m_active_time,Amount of total active time (in seconds). 
+m_calories,Total number of calories burned during active time (in seconds). 
+m_distance,Total distance traveled (in meters). 
+m_inactive_time,Total inactive time (in seconds). 
+m_lcat,Longest consecutive active time (in seconds). 
+m_lcit,Longest consecutive inactive time (in seconds). 
+m_steps,Total number of steps taken. 
+m_steps_3am,Total number of steps taken (before 3am). 
+m_total_calories,Total number of calories burned in the day
+m_workout_count,Number of workouts logged. 
+m_workout_time,Length of logged workouts (in seconds). 
+max_bg,Highest passive heart rate (in beats per minute). 
+min_bg,Lowest passive heart rate (in beats per minute). 
+n_asleep_time,Duration of naps/secondary sleep (in seconds). 
+n_awake,Total time awake during naps/secondary sleep (in seconds). 
+n_awake_time,Total time awake after waking from naps/secondary sleep (in seconds). 
+n_awakenings,Number of times awoken during naps/secondary sleep. 
+n_bedtime,Length of time band in sleep mode during naps/secondary sleep (in seconds). 
+n_clinical_deep,Length of Deep sleep during naps/secondary sleep (in seconds UP3 and UP4 only). 
+n_count,Number of naps/secondary sleep entries logged. 
+n_deep,Length of Sound sleep during naps/secondary sleep (in seconds). 
+n_duration,Duration of naps/secondary sleep (in seconds). 
+n_light,Length of Light sleep during naps/secondary sleep (in seconds). 
+n_quality,Not applicable. 
+n_rem,Length of REM sleep during naps/secondary sleep (in seconds). 
+n_to_bed_phr,Average passive heart rate 1 hour before naps/secondary sleep (in beats per minute). 
+num_readings,Number of background heart rate readings. 
+o_count,Number of moods logged. 
+o_mood,Average score for moods logged (10 = lowest 80 = highest). 
+rhr,Resting heart rate (in beats per minute). 
+s_asleep_time,Duration of primary sleep (in seconds). 
+s_awake,Total time awake during primary sleep (in seconds). 
+s_awake_time,Total time awake after waking from primary sleep (in seconds). 
+s_awakenings,Number of times awoken during primary sleep. 
+s_bedtime,Length of time band was in primary sleep mode (in seconds). 
+s_clinical_deep,Length of primary Deep sleep (in seconds UP3 and UP4 only). 
+s_count,Number of primary sleep entries logged. 
+s_deep,Length of primary Sound sleep (in seconds). 
+s_duration,Duration of primary sleep (in seconds). 
+s_light,Length of primary Light sleep (in seconds). 
+s_quality,Not applicable. 
+s_rem,Length of primary REM sleep (in seconds). 
+s_to_bed_phr,Average passive heart rate 1 hour before primary sleep (in beats per minute). 
+weight,Weight (in kilograms). 

TODO.org

@@ -0,0 +1,6 @@
+https://github.com/crowoy/Health-Analysis
+https://github.com/joytafty-work/SleepModel
+https://github.com/carleshf/jawboneup2tex
+https://github.com/search?l=Jupyter+Notebook&q=s_awakenings&type=Code&utf8=%E2%9C%93
+https://github.com/oshev/colifer/blob/592cc6b4d1ac9005c52fccdfb4e207513812baaa/colifer.py
+https://github.com/oshev/colifer/blob/592cc6b4d1ac9005c52fccdfb4e207513812baaa/reportextenders/jawbone/jawbone_sleep.py

main.py

@@ -0,0 +1,219 @@
+#!/usr/bin/env python3.6
+# TODO
+from csv import DictReader
+from itertools import islice
+
+from typing import Dict
+
+# sleep = []
+# with open('2017.csv', 'r') as fo:
+#     reader = DictReader(fo)
+#     for line in islice(reader, 0, 10):
+#         sleep
+#         print(line)
+
+import numpy as np
+import matplotlib.pyplot as plt
+from numpy import genfromtxt
+import matplotlib.pylab as pylab
+
+pylab.rcParams['figure.figsize'] = (32.0, 24.0)
+pylab.rcParams['font.size'] = 10
+
+
+
+dimensions = 3 # Number of dimensions to reduce to
+jawboneDataFile = "/L/Dropbox/backups/jawbone/2017.csv" # Data File Path
+
+jawboneDataFeatures = "Jawbone/features.csv" # Data File Path
+featureDesc: Dict[str, str] = {}
+for x in genfromtxt(jawboneDataFeatures, dtype='unicode', delimiter=','):
+    featureDesc[x[0]] = x[1]
+
+def filterData_Jawbone (data):
+    #Removes null data (and corresponding features)
+    data = data[0:,:]
+    # for i in range(16):
+    #     data = np.delete(data, 0, 1)
+    # print(data)
+    h, w = data.shape
+    data = np.where((data == ''), 0, data)
+    allZero = [np.all(np.delete([0 if col[i] == '' else col[i] for col in data], [0]).astype(float)
+              == 0) for i in range(w)]
+    allSame = [np.all(np.delete([0 if col[i] == '' else col[i] for col in data], [0]).astype(float)
+              == np.delete([0 if col[i] == '' else col[i] for col in data], [0]).astype(float)[0]) for i in range(w)]
+    empty = np.logical_or(allZero, allSame)
+    n = [i for i in range(np.array(empty).size) if empty[i] == True]
+    return np.delete(data, n, axis=1)
+
+dataAll = filterData_Jawbone(genfromtxt(jawboneDataFile, dtype='unicode', delimiter=','))
+features = dataAll[0]
+features = [
+    's_light', # 'light sleep' from app
+    's_awake', # 'woke up' from app (how many times you were awake)
+    's_deep' # 'sound sleep' from app
+]
+# TODO filter more carefully...
+
+
+def getIndex (data, features):
+    index = []
+    for f in features:
+        index.append(np.where((data[0] == f) == True)[0][0])
+    return index
+
+def getFeatures (data, features):
+    h, w = data.shape
+    index = getIndex(data, features)
+    extracted = np.zeros(h-1)
+    for i in index:
+        temp = np.delete([0 if col[i] == '' else col[i] for col in data], [0]).astype(float)
+        temp /= np.amax(temp)
+        extracted = np.vstack((extracted, temp))
+    extracted = np.delete(extracted, 0, 0)
+    return extracted
+
+
+# print(dataAll)
+data = getFeatures(dataAll, features)
+
+
+def remNull(x, y):
+    nx = np.where(x == 0)
+    ny = np.where(y == 0)
+    nulli = np.concatenate((nx[0], ny[0]))
+    x = np.delete(x, nulli, 0)
+    y = np.delete(y, nulli, 0)
+    return x, y
+
+def calculateVar(x, y) -> float:
+    x, y = remNull(x,y)
+    if len(x) == 0:
+        # TODO needs date?
+        print("Warning")
+        return 0.0 # TODO ???
+    meanX = np.mean(x)
+    meanY = np.mean(y)
+    n = float(x.shape[0])
+    print(n)
+    return ((1/n)*(np.sum((x-meanX)*(y-meanY))))
+    # return ((1/(n + 1))*(np.sum((x-meanX)*(y-meanY)))) # TODO fixme..
+
+def calculateCov(data):
+    h, w = data.shape
+    cov = np.zeros([h, h])
+
+    for i in range(h):
+        for j in range(h):
+            cov[i][j] = calculateVar(data[i], data[j])
+    return cov
+
+
+# In[119]:
+# a = np.array([[1, 2, 3], [1, 2, 3]])
+# print(a)
+# print(calculateCov(a))
+# print(np.cov(a))
+# print("VAR")
+# print(np.var(a[0]))
+
+# print("DATA")
+# print(data)
+
+# print("NPCOV")
+# print(np.cov(data))
+
+# cov = calculateCov (data)
+# print("COV")
+# print(cov)
+cov = np.cov(data) # TODO ??? 
+
+
+# In[120]:
+
+
+def plotFeatures (title, label1, label2, feature1, feature2):
+    plt.scatter(feature1, feature2)
+    
+    plt.title(title)
+    plt.xlabel(label1)
+    plt.ylabel(label2)
+    
+    plt.xlim(0, 1)
+    plt.ylim(0, 1)
+    
+    plt.show()
+    
+def plotMatrix(data):
+    r, c = data.shape
+    c=2
+    fig = plt.figure()
+    plotID = 1
+    for i in range(c):
+        for j in range(c):
+            f1 = getFeature(data, data[0][i])
+            f2 = getFeature(data, data[0][j])
+            ax = fig.add_subplot( c, c, plotID )
+            ax.scatter(f1, f2)
+            ax.set_title(data[0][i] + ' vs ' + data[0][j])
+            ax.axis('off')
+            plotID += 1
+    plt.show()
+    
+def plotMatrix1(features, data):
+    for f in features:
+        print(f"{f}: {featureDesc[f]}")
+    r, c = data.shape
+    fig = plt.figure()
+    plotID = 1
+    for i in range(r):
+        for j in range(r):
+            ax = fig.add_subplot( r, r, plotID )
+            x,y = remNull(data[i], data[j])
+            ax.scatter(x, y, s=2)
+            ax.set_title(features[i] + ' vs ' + features[j], fontsize=15)
+            ax.tick_params(axis='x', which='major', labelsize=8)
+            ax.tick_params(axis='y', which='major', labelsize=8)
+#             ax.set_xlim(0,1)
+#             ax.set_ylim(0,1)
+            plotID += 1
+    plt.show()
+
+
+# In[121]:
+
+
+# plotMatrix1(features, data)
+
+
+# In[ ]:
+
+
+def rankF(features, cov):
+    n = len(features)
+    eigenV = np.linalg.eig(cov)
+    eigVal = np.matrix(eigenV[0])
+    eigVec = np.matrix(eigenV[1])
+    order  = (n-1) - np.argsort(eigVal)
+    
+    rankFeatures = np.empty(n, dtype='<U30') # TODO
+    # print(rankFeatures.shape)
+    for i in range(n):
+        rankFeatures[i] = features[(np.where(order == i)[1][0])]
+    
+    return rankFeatures, eigVal, eigVec
+
+# print(features)
+# rankFeatures, eigVal, eigVec = rankF(features, cov)
+rankFeatures = features
+# print(rankFeatures)
+# print(len(rankFeatures))
+
+
+r1, r2 = 0, dimensions
+selectedFeatures = features
+# selectedFeatures = np.take(rankFeatures, np.arange(r1, r2))
+selectedData = getFeatures(dataAll, selectedFeatures)
+# plotFeatures('111', 'f1', 'f2', selectedData[0], selectedData[1])
+
+plotMatrix1(rankFeatures, selectedData)