pip install seaborn==0.11.0

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns
import random
import os
from sklearn.linear_model import LinearRegression
import datetime as dt
sns.set_theme()

filename = os.path.join(os.getcwd(), "FitabaseData", "dailyActivity_merged.csv")
df = pd.read_csv(filename, header=0)

df.columns

df.head(10)

len(df["Id"].unique()) #33 unique users tracked 

df["Id"].value_counts() # days tracked ranges: 4, 18-31

df.shape #940 total data examples

df_summ = df.describe(include = 'all')
df_summ

df["SedentaryActiveDistance"].unique()
# SedentaryActiveDistance does not contribute much value to the total distance feature
# WILL DROP 

# Question: does LoggedActivitiesDistance get added into the TotalDistance. 
# If so, will drop
df["LoggedActivitiesDistance"].unique()

# testing above question

df["test_total_distance"] = df["TrackerDistance"] + df["LoggedActivitiesDistance"] + df["VeryActiveDistance"] + df["ModeratelyActiveDistance"] + df["LightActiveDistance"] + df["SedentaryActiveDistance"]
total_diff_vals = df["test_total_distance"] == df["TotalDistance"]
total_diff_vals.value_counts()

# conclustion: 85 out of 940 have a different values between the app provided TotalDistance
# and the summed distance features I created. 

# due to time constraints, it's a low priority to understand the unique differnces between
# the distance features

# will use just total distance as the main distance feature for the analysis 

# question: are TotalDistance and TrackerDistance the same values? 
# if so will drop TrackerDist

df["same_val_tot_dist"]= (df["TotalDistance"]==df["TrackerDistance"])

# answer: no there are some values that are not the same:
df[~df["same_val_tot_dist"]]

print(df[~df["same_val_tot_dist"]].shape)

#conclusion: do not use TrackerDistance as a feature. 
#            only 15 out of 940 are different in value
#            purpose/metric of TrackerDistance unknown at this time

sns.histplot(data=df, x="Calories")

sns.histplot(data=df, x="TotalSteps")

# combining all minute features to create "total_minutes" feature
df["total_minutes"] = df["VeryActiveMinutes"] + df["FairlyActiveMinutes"] + df["LightlyActiveMinutes"]
df["total_minutes"]

df.head()

# creating df of features I will use and dropping rest


features_to_drop = ['same_val_tot_dist','TrackerDistance', 'LoggedActivitiesDistance', 'VeryActiveDistance',
                   'ModeratelyActiveDistance', 'LightActiveDistance', 'SedentaryActiveDistance', 'VeryActiveMinutes',
                    'FairlyActiveMinutes', 'LightlyActiveMinutes', 'SedentaryMinutes', 'test_total_distance']


df_features = df.drop(columns = features_to_drop, axis = 1, inplace=False)

df_features.head(10)

# converting the date data so it can be read as a number

df_features['Date'] = pd.to_datetime(df_features['ActivityDate'])
df_features['Date'] = df_features['Date'].map(dt.datetime.toordinal)
df_features.head()

#drop the ActivityDate
df_features.drop(columns = 'ActivityDate', axis = 1, inplace=True)
df_features

#randomly select user:

rand_user = random.choice(df_features['Id'])
rand_user

# first I want to get a list of index values for NOT this user

index_list_for_user = df_features[df_features['Id'] != rand_user ].index
index_list_for_user # returns a list of index values in df_features for everyone BUT this user

# now I want to build a dataframe for just that user at those index values:
df_for_user = df_features.drop(index_list_for_user, inplace=False)
df_for_user

print(df_for_user.shape)

#define data
x = np.array(df_for_user['Date'])
y = np.array(df_for_user['Calories'])

#find line of best fit
a, b = np.polyfit(x, y, 1)

#add points to plot
plt.scatter(x, y, color='purple')

#add line of best fit to plot
plt.plot(x, a*x+b, color='steelblue', linestyle='--', linewidth=2)

#define data
x = np.array(df_for_user['Date'])
y = np.array(df_for_user['TotalDistance'])

#find line of best fit
a, b = np.polyfit(x, y, 1)

#add points to plot
plt.scatter(x, y, color='purple')

#add line of best fit to plot
plt.plot(x, a*x+b, color='steelblue', linestyle='--', linewidth=2)

#define data
x = np.array(df_for_user['Date'])
y = np.array(df_for_user['TotalSteps'])

#find line of best fit
a, b = np.polyfit(x, y, 1)

#add points to plot
plt.scatter(x, y, color='purple')

#add line of best fit to plot
plt.plot(x, a*x+b, color='steelblue', linestyle='--', linewidth=2)

#define data
x = np.array(df_for_user['Date'])
y = np.array(df_for_user['total_minutes'])

#find line of best fit
a, b = np.polyfit(x, y, 1)

#add points to plot
plt.scatter(x, y, color='purple')

#add line of best fit to plot
plt.plot(x, a*x+b, color='steelblue', linestyle='--', linewidth=2)