Skip to content

GitLab

Commit 79e18df0 authored by Lukas-Stingl's avatar Lukas-Stingl
Browse files

Merge branch 'master' of https://git.scc.kit.edu/uflgi/bda-analytics-challenge-template

parents 4da354e6 95194b68
Expand all Hide whitespace changes
Inline Side-by-side
notebooks/Cluster0_LSTM.ipynb
View file @ 79e18df0
This diff is collapsed.
notebooks/Cluster2_LSTM.ipynb deleted 100644 → 0
View file @ 4da354e6
This diff is collapsed.
notebooks/DataTransformation.ipynb deleted 100644 → 0
View file @ 4da354e6
This diff is collapsed.
notebooks/DataVisualization.ipynb
View file @ 79e18df0
This diff is collapsed.
notebooks/IncludeWeatherAndHoliday.ipynb
View file @ 79e18df0
%% Cell type:code id: tags:
``` python
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import os
import seaborn as sn
import datetime
```
%% Cell type:code id: tags:
``` python
path_0 = '../data/preprocessed/clusters/0/'
path_1 = '../data/preprocessed/clusters/1/'
path_2 = '../data/preprocessed/clusters/2/'
dfs_c_0 = []
dfs_c_1 = []
dfs_c_2 = []
dfs_c_0_grouped = []
dfs_c_1_grouped = []
dfs_c_2_grouped = []
csv_files_0 = [csv for csv in os.listdir(path_0) if csv.endswith('.csv')]
csv_files_1 = [csv for csv in os.listdir(path_1) if csv.endswith('.csv')]
csv_files_2 = [csv for csv in os.listdir(path_2) if csv.endswith('.csv')]
for file in csv_files_0:
# import DataFrame
df = pd.read_csv(path_0 + file)
if file.startswith('g_'):
dfs_c_0_grouped.append(df)
else:
dfs_c_0.append(df)
for file in csv_files_1:
# import DataFrame
df = pd.read_csv(path_1 + file)
if file.startswith('g_'):
dfs_c_1_grouped.append(df)
else:
dfs_c_1.append(df)
for file in csv_files_2:
# import DataFrame
df = pd.read_csv(path_2 + file)
if file.startswith('g_'):
dfs_c_2_grouped.append(df)
else:
dfs_c_2.append(df)
import sys
from pathlib import Path
# in jupyter (lab / notebook), based on notebook path
module_path = str(Path.cwd().parents[0] / "src")
if module_path not in sys.path:
sys.path.append(module_path)
from preprocessing import import_preprocessed_data, assign_holidays, assign_weather
```
%% Cell type:code id: tags:
``` python
%load_ext autoreload
%autoreload 2
%aimport preprocessing
```
%% Cell type:markdown id: tags:
Include all data into DataFrames
%% Cell type:code id: tags:
``` python
dfs_c_0, dfs_c_1, dfs_c_2, dfs_c_0_grouped, dfs_c_1_grouped, dfs_c_2_grouped = import_preprocessed_data() # import all data
```
%% Cell type:markdown id: tags:
Include holiday data
Include holiday data, holiday = 1, no holiday = 0
%% Cell type:code id: tags:
``` python
holiday = '../data/raw/holidays.csv'
df_holiday = pd.read_csv(holiday)
df_holiday
```
%%%% Output: execute_result
Date Holiday
0 2020-01-01 1
1 2020-01-02 0
2 2020-01-03 0
3 2020-01-04 0
4 2020-01-05 1
.. ... ...
726 2021-12-27 0
727 2021-12-28 0
728 2021-12-29 0
729 2021-12-30 0
730 2021-12-31 0
[731 rows x 2 columns]
%% Cell type:code id: tags:
``` python
import datetime
```
%% Cell type:code id: tags:
``` python
def assign_holidays(df, df_holiday):
df['holiday'] = 0
for i in range(0,len(df), 1):
form = "%Y-%m-%d"
d1 = datetime.datetime.strptime(df['time_stamp'][i], form)
#get holiday value
for j in range(0, len(df_holiday), 1):
d2 = datetime.datetime.strptime(df_holiday['Date'][j], form)
if d1 == d2:
df['holiday'][i] = df_holiday['Holiday'][j]
```
%% Cell type:code id: tags:
``` python
for df in dfs_c_0_grouped:
assign_holidays(df, df_holiday)
for df in dfs_c_1_grouped:
assign_holidays(df, df_holiday)
for df in dfs_c_2_grouped:
assign_holidays(df, df_holiday)
```
%%%% Output: stream
C:\Users\hendr\AppData\Local\Temp/ipykernel_8496/2432097566.py:12: SettingWithCopyWarning:
A value is trying to be set on a copy of a slice from a DataFrame
See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
df['holiday'][i] = df_holiday['Holiday'][j]
%% Cell type:markdown id: tags:
Include weather data
Include weather data of Frankfurt airport
%% Cell type:code id: tags:
``` python
weather_path = '../data/raw/weather_FrankfurtAirport.csv'
df_weather = pd.read_csv(weather_path)
df_weather
```
%%%% Output: execute_result
date tavg tmin tmax prcp snow wdir wspd wpgt pres tsun
0 2020-05-08 17.5 8.2 24.8 0.0 0 112.0 7.2 27.7 1018.0 616
1 2020-05-09 15.7 12.5 17.4 0.6 0 60.0 11.5 25.2 1013.5 10
2 2020-05-10 16.8 11.7 24.9 2.9 0 129.0 8.6 63.0 1006.5 308
3 2020-05-11 7.2 1.9 15.5 9.0 0 56.0 23.8 61.2 1009.2 38
4 2020-05-12 7.9 1.0 14.2 0.0 0 130.0 8.3 24.1 1018.7 736
.. ... ... ... ... ... ... ... ... ... ... ...
361 2021-05-04 10.5 8.0 13.7 0.0 0 215.0 32.4 77.8 1006.5 67
362 2021-05-05 7.9 4.1 11.4 0.7 0 248.0 24.1 65.5 1008.7 422
363 2021-05-06 6.9 3.2 12.1 5.9 0 174.0 11.5 34.9 1009.1 46
364 2021-05-07 7.5 1.5 12.1 0.7 0 273.0 14.4 59.4 1015.9 654
365 2021-05-08 10.0 -0.7 17.5 0.0 0 169.0 9.4 38.2 1017.6 620
[366 rows x 11 columns]
%% Cell type:code id: tags:
``` python
def assign_weather(df, df_weather):
df['temp avg'] = 0.0
df['temp min'] = 0.0
df['temp max'] = 0.0
df['rainfall sum'] = 0.0
df['snowfall sum'] = 0.0
df['sunshine minutes'] = 0
for i in range(0,len(df), 1):
form = "%Y-%m-%d"
d1 = datetime.datetime.strptime(df['time_stamp'][i], form)
#get weather values
for j in range(0, len(df_weather), 1):
d2 = datetime.datetime.strptime(df_weather['date'][j], form)
if d1 == d2:
df['temp avg'][i] = df_weather['tavg'][j]
df['temp min'][i] = df_weather['tmin'][j]
df['temp max'][i] = df_weather['tmax'][j]
df['rainfall sum'][i] = df_weather['prcp'][j]
df['snowfall sum'][i] = df_weather['snow'][j]
df['sunshine minutes'][i] = df_weather['tsun'][j]
```
%% Cell type:code id: tags:
``` python
# assign to all DataFrames
for df in dfs_c_0_grouped:
assign_weather(df, df_weather)
for df in dfs_c_1_grouped:
assign_weather(df, df_weather)
for df in dfs_c_2_grouped:
assign_weather(df, df_weather)
```
%%%% Output: stream
C:\Users\hendr\AppData\Local\Temp/ipykernel_8496/777759664.py:17: SettingWithCopyWarning:
A value is trying to be set on a copy of a slice from a DataFrame
See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
df['temp avg'][i] = df_weather['tavg'][j]
C:\Users\hendr\AppData\Local\Temp/ipykernel_8496/777759664.py:18: SettingWithCopyWarning:
A value is trying to be set on a copy of a slice from a DataFrame
See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
df['temp min'][i] = df_weather['tmin'][j]
C:\Users\hendr\AppData\Local\Temp/ipykernel_8496/777759664.py:19: SettingWithCopyWarning:
A value is trying to be set on a copy of a slice from a DataFrame
See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
df['temp max'][i] = df_weather['tmax'][j]
C:\Users\hendr\AppData\Local\Temp/ipykernel_8496/777759664.py:20: SettingWithCopyWarning:
A value is trying to be set on a copy of a slice from a DataFrame
See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
df['rainfall sum'][i] = df_weather['prcp'][j]
C:\Users\hendr\AppData\Local\Temp/ipykernel_8496/777759664.py:21: SettingWithCopyWarning:
A value is trying to be set on a copy of a slice from a DataFrame
See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
df['snowfall sum'][i] = df_weather['snow'][j]
C:\Users\hendr\AppData\Local\Temp/ipykernel_8496/777759664.py:22: SettingWithCopyWarning:
A value is trying to be set on a copy of a slice from a DataFrame
See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
df['sunshine minutes'][i] = df_weather['tsun'][j]
%% Cell type:code id: tags:
``` python
# Loading the dataset
# Loading one container dataset
data = dfs_c_0_grouped[0]
# Numeric columns of the dataset
numeric_col = ['inter_pol','Temperature','Tilt','holiday','temp avg','temp min','temp max','rainfall sum','snowfall sum','sunshine minutes']
# Correlation Matrix formation
corr_matrix = data.loc[:,numeric_col].corr()
print(corr_matrix)
#Using heatmap to visualize the correlation matrix
sn.heatmap(corr_matrix, annot=True)
```
%%%% Output: stream
inter_pol Temperature Tilt holiday temp avg \
inter_pol 1.000000 0.272704 0.066792 0.036456 0.330773
Temperature 0.272704 1.000000 -0.253476 -0.101045 0.954478
Tilt 0.066792 -0.253476 1.000000 0.060309 -0.241740
holiday 0.036456 -0.101045 0.060309 1.000000 -0.109564
temp avg 0.330773 0.954478 -0.241740 -0.109564 1.000000
temp min 0.352179 0.858892 -0.164456 -0.095769 0.947680
temp max 0.289772 0.963847 -0.285066 -0.107842 0.976731
rainfall sum 0.059159 -0.054447 0.060050 -0.043965 0.011288
snowfall sum -0.051326 -0.223033 0.174676 -0.032307 -0.270178
sunshine minutes 0.110113 0.648934 -0.234403 -0.005993 0.495271
temp min temp max rainfall sum snowfall sum \
inter_pol 0.352179 0.289772 0.059159 -0.051326
Temperature 0.858892 0.963847 -0.054447 -0.223033
Tilt -0.164456 -0.285066 0.060050 0.174676
holiday -0.095769 -0.107842 -0.043965 -0.032307
temp avg 0.947680 0.976731 0.011288 -0.270178
temp min 1.000000 0.871725 0.097316 -0.246456
temp max 0.871725 1.000000 -0.025114 -0.270493
rainfall sum 0.097316 -0.025114 1.000000 -0.016413
snowfall sum -0.246456 -0.270493 -0.016413 1.000000
sunshine minutes 0.272622 0.610498 -0.252725 -0.126974
sunshine minutes
inter_pol 0.110113
Temperature 0.648934
Tilt -0.234403
holiday -0.005993
temp avg 0.495271
temp min 0.272622
temp max 0.610498
rainfall sum -0.252725
snowfall sum -0.126974
sunshine minutes 1.000000
%%%% Output: execute_result
<AxesSubplot:>
%%%% Output: display_data
![](data:image/png;base64,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)
%% Cell type:code id: tags:
``` python
```
%% Cell type:code id: tags:
%% Cell type:markdown id: tags:
``` python
```
The matrix shows, that there is poor correlation between the features and the target (inter_pol). Therefore, the features will not be used further.
......
notebooks/Overall_LSTM.ipynb
View file @ 79e18df0
This diff is collapsed.
notebooks/Overall_grouped_LSTM.ipynb
View file @ 79e18df0
This diff is collapsed.
notebooks/data_collection.ipynb deleted 100644 → 0
View file @ 4da354e6
This diff is collapsed.
notebooks/multiple_regression.ipynb deleted 100644 → 0
View file @ 4da354e6
%% Cell type:code id: tags:
```
import sys
sys.path.append('../src')
import preprocessing
```
%%%% Output: stream
/Users/davidblumenthal/opt/miniconda3/envs/bda/lib/python3.9/site-packages/tslearn/bases/bases.py:15: UserWarning: h5py not installed, hdf5 features will not be supported.
Install h5py to use hdf5 features: http://docs.h5py.org/
warn(h5py_msg)
%% Cell type:code id: tags:
```
```
src/predict.py
View file @ 79e18df0
import pandas as pd
import numpy as np
# function uses the model and the initial input set to predict a given number of following values
def predict_values(model, initial_set, values, window_size):
result = []
for i in range(0, values):
x_input = initial_set.reshape((1, window_size, 1))
yhat = model.predict(x_input, verbose=0)
result.append(yhat[0][0])
# update the model input for the next prediction
initial_set = np.append(initial_set, yhat)
initial_set = np.delete(initial_set, 0)
return result
# function uses the model and the initial input set to predict a given number of following values.
# additionally the gradient is
def predict_values_with_gradient(model, initial_set, initial, values, window_size):
result = []
for i in range(0, values):
x_input = initial_set.reshape((1, window_size, 1))
yhat = model.predict(x_input, verbose=0)
result.append(yhat[0][0])
# update the model input for the next prediction
initial_set = np.append(initial_set, yhat)
initial_set = np.delete(initial_set, 0)
# calculate the gradient of the last values. If it falls under a threshold,
# the input data will be reseted to 140.
gradients = np.gradient(initial_set)
if ((sum(gradients) / len(gradients))) > -1 and (initial_set[window_size-1] < 20):
initial_set = initial
return result
src/preprocessing.py
View file @ 79e18df0
......@@ -2,6 +2,10 @@ import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import os
import gc
from mat4py import loadmat
import json
import datetime
# function loads the preprocessed data into DataFrames.
def import_preprocessed_data():
......@@ -75,3 +79,86 @@ def add_empties_column(df, values):
df_final['empties'][i] = 1
return df_final
# function parses JSON data into csv file format. Only the important information is saved in csv
def create_csv_files():
path_to_json = '../data/raw/data/'
json_files = [pos_json for pos_json in os.listdir(path_to_json) if pos_json.endswith('.txt')]
csv_folder = '../data/preprocessed/CSV/'
for filename in json_files:
f = open(path_to_json + filename,)
data = json.load(f)
f.close()
json_objects = data[1]
# create important columns
columns = ['id','deveui', 'unix_time', 'client_id', 'created_at', 'Status', 'Sensor ID', 'Events', 'Height', 'Voltage', 'Temperature', 'Tilt', 'Tx Event', 'Messagetype']
df = pd.DataFrame(columns=columns)
# import all json data into dataframes
for i in range(0, len(json_objects)):
new_data = {'id': int(json_objects[i]['id']),'deveui':str(json_objects[i]['deveui']), 'unix_time':int(json_objects[i]['unix_time']), 'client_id':str(json_objects[i]['client_id']), 'created_at':str(json_objects[i]['created_at']), 'Status':json_objects[i]['decoded_data']['sensor_data']['Status'], 'Sensor ID':str(json_objects[i]['decoded_data']['sensor_data']['Sensor ID']), 'Events':str(json_objects[i]['decoded_data']['sensor_data']['Events']), 'Height':str(json_objects[i]['decoded_data']['sensor_data']['Height 1']), 'Voltage':str(json_objects[i]['decoded_data']['sensor_data']['Voltage']), 'Temperature':str(json_objects[i]['decoded_data']['sensor_data']['Temperature']), 'Tilt':str(json_objects[i]['decoded_data']['sensor_data']['Tilt']), 'Tx Event':str(json_objects[i]['decoded_data']['sensor_data']['Tx Event'])}
df = df.append(new_data, ignore_index=True)
df = df[::-1] #reverse values
# parse data into correct data types
df['Height'] = df['Height'].apply(lambda x: str(x).split(' ')[0])
df['Voltage'] = df['Voltage'].apply(lambda x: str(x).split(' ')[0])
df['Temperature'] = df['Temperature'].apply(lambda x: str(x).split(' ')[0])
df['Tilt'] = df['Tilt'].apply(lambda x: str(x).split(' ')[0])
df['Tx Event'] = df['Tx Event'].apply(lambda x: str(x).split(' ')[0])
df['Height'] = df['Height'].astype('int')
df['Voltage'] = df['Voltage'].astype('int')
df['Temperature'] = df['Temperature'].astype('int')
df['Tilt'] = df['Tilt'].astype('int')
df['Tx Event'] = df['Tx Event'].astype('int')
# save DataFrames in csv files
filename = df['deveui'][0]
filename = filename + ".csv"
df.to_csv(csv_folder + filename)
# function will assign holiday value the given container data frame
def assign_holidays(df, df_holiday):
df['holiday'] = 0
for i in range(0,len(df), 1):
form = "%Y-%m-%d"
d1 = datetime.datetime.strptime(df['time_stamp'][i], form)
#get holiday value
for j in range(0, len(df_holiday), 1):
d2 = datetime.datetime.strptime(df_holiday['Date'][j], form)
if d1 == d2:
df['holiday'][i] = df_holiday['Holiday'][j]
# function will assign weather data of Frankfurt airport to container data frame
def assign_weather(df, df_weather):
df['temp avg'] = 0.0
df['temp min'] = 0.0
df['temp max'] = 0.0
df['rainfall sum'] = 0.0
df['snowfall sum'] = 0.0
df['sunshine minutes'] = 0
for i in range(0,len(df), 1):
form = "%Y-%m-%d"
d1 = datetime.datetime.strptime(df['time_stamp'][i], form)
#get weather values
for j in range(0, len(df_weather), 1):
d2 = datetime.datetime.strptime(df_weather['date'][j], form)
if d1 == d2:
df['temp avg'][i] = df_weather['tavg'][j]
df['temp min'][i] = df_weather['tmin'][j]
df['temp max'][i] = df_weather['tmax'][j]
df['rainfall sum'][i] = df_weather['prcp'][j]
df['snowfall sum'][i] = df_weather['snow'][j]
df['sunshine minutes'][i] = df_weather['tsun'][j]
\ No newline at end of file
src/train.py
View file @ 79e18df0
from keras.preprocessing.sequence import TimeseriesGenerator
from sklearn.model_selection import train_test_split
import tensorflow as tf
import numpy
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import os
# function will take container data as input and return LSTM ready input format data.
# The window length and overlap can be set and a timeseries generator is used.
def create_windows_smoothed(dfs, length, batch_size, stride):
features = []
targets = []
X = []
y = []
# Height as feature and target
for df in dfs:
height = df['inter_pol'].to_numpy().tolist()
# apply TimeSeriesGenerator
ts_generator = TimeseriesGenerator(height,height,length=length, batch_size=batch_size, stride=stride)
for j in range(len(ts_generator)):
features.append(ts_generator[j][0])
targets.append(ts_generator[j][1])
#reshape data for neural network
for i in range(len(features)):
x = np.reshape(features[i], (length,1))
X.append(x)
X = np.array(X)
y = np.array(targets)
return X, y
# function creates train/validation/test split on data
def create_train_val_test_split(X,y):
#Split data into train & test set & validation set
X_tr, X_val, y_tr, y_val = train_test_split(X, y, test_size=0.2)
X_train, X_test, y_train, y_test = train_test_split(X_tr, y_tr, test_size=0.2)
return X_train, y_train, X_test, y_test, X_val, y_val
\ No newline at end of file