/ TIMESERIES

Time Series Analysis - Multivariage Time Series with LSTM & RNNs

Multivariate Time Series

import pandas as pd
import numpy as np
%matplotlib inline
import matplotlib.pyplot as plt

df = pd.read_csv('MultiVariate-RNN-with-TensorFlow-and-Keras-master/DATA/energydata_complete.csv', index_col='date',infer_datetime_format=True)

df.head()
Appliances lights T1 RH_1 T2 RH_2 T3 RH_3 T4 RH_4 ... T9 RH_9 T_out Press_mm_hg RH_out Windspeed Visibility Tdewpoint rv1 rv2
date
2016-01-11 17:00:00 60 30 19.89 47.596667 19.2 44.790000 19.79 44.730000 19.000000 45.566667 ... 17.033333 45.53 6.600000 733.5 92.0 7.000000 63.000000 5.3 13.275433 13.275433
2016-01-11 17:10:00 60 30 19.89 46.693333 19.2 44.722500 19.79 44.790000 19.000000 45.992500 ... 17.066667 45.56 6.483333 733.6 92.0 6.666667 59.166667 5.2 18.606195 18.606195
2016-01-11 17:20:00 50 30 19.89 46.300000 19.2 44.626667 19.79 44.933333 18.926667 45.890000 ... 17.000000 45.50 6.366667 733.7 92.0 6.333333 55.333333 5.1 28.642668 28.642668
2016-01-11 17:30:00 50 40 19.89 46.066667 19.2 44.590000 19.79 45.000000 18.890000 45.723333 ... 17.000000 45.40 6.250000 733.8 92.0 6.000000 51.500000 5.0 45.410389 45.410389
2016-01-11 17:40:00 60 40 19.89 46.333333 19.2 44.530000 19.79 45.000000 18.890000 45.530000 ... 17.000000 45.40 6.133333 733.9 92.0 5.666667 47.666667 4.9 10.084097 10.084097

5 rows × 28 columns

df.info()

<class 'pandas.core.frame.DataFrame'>
Index: 19735 entries, 2016-01-11 17:00:00 to 2016-05-27 18:00:00
Data columns (total 28 columns):
 #   Column       Non-Null Count  Dtype  
---  ------       --------------  -----  
 0   Appliances   19735 non-null  int64  
 1   lights       19735 non-null  int64  
 2   T1           19735 non-null  float64
 3   RH_1         19735 non-null  float64
 4   T2           19735 non-null  float64
 5   RH_2         19735 non-null  float64
 6   T3           19735 non-null  float64
 7   RH_3         19735 non-null  float64
 8   T4           19735 non-null  float64
 9   RH_4         19735 non-null  float64
 10  T5           19735 non-null  float64
 11  RH_5         19735 non-null  float64
 12  T6           19735 non-null  float64
 13  RH_6         19735 non-null  float64
 14  T7           19735 non-null  float64
 15  RH_7         19735 non-null  float64
 16  T8           19735 non-null  float64
 17  RH_8         19735 non-null  float64
 18  T9           19735 non-null  float64
 19  RH_9         19735 non-null  float64
 20  T_out        19735 non-null  float64
 21  Press_mm_hg  19735 non-null  float64
 22  RH_out       19735 non-null  float64
 23  Windspeed    19735 non-null  float64
 24  Visibility   19735 non-null  float64
 25  Tdewpoint    19735 non-null  float64
 26  rv1          19735 non-null  float64
 27  rv2          19735 non-null  float64
dtypes: float64(26), int64(2)
memory usage: 4.4+ MB

df['Windspeed'].plot(figsize=(12,8))

<AxesSubplot:xlabel='date'>

png

df['Appliances'].plot(figsize=(12,8))

<AxesSubplot:xlabel='date'>

png

Train Test Split

len(df)

19735

df.loc['2016-05-01':]
Appliances lights T1 RH_1 T2 RH_2 T3 RH_3 T4 RH_4 ... T9 RH_9 T_out Press_mm_hg RH_out Windspeed Visibility Tdewpoint rv1 rv2
date
2016-05-01 00:00:00 50 0 21.200000 38.200000 18.390000 41.400000 23.200000 36.400000 19.73 37.730000 ... 19.166667 38.200000 4.300000 763.700000 86.000000 2.000000 40.000000 2.200000 15.310003 15.310003
2016-05-01 00:10:00 60 0 21.150000 38.172500 18.390000 41.500000 23.133333 36.466667 19.70 37.663333 ... 19.133333 38.290000 4.283333 763.716667 86.333333 2.166667 38.166667 2.216667 11.240067 11.240067
2016-05-01 00:20:00 50 0 21.133333 38.090000 18.323333 41.500000 23.200000 36.500000 19.70 37.590000 ... 19.133333 38.363333 4.266667 763.733333 86.666667 2.333333 36.333333 2.233333 21.808814 21.808814
2016-05-01 00:30:00 50 0 21.100000 38.090000 18.290000 41.500000 23.200000 36.500000 19.70 37.500000 ... 19.166667 38.500000 4.250000 763.750000 87.000000 2.500000 34.500000 2.250000 47.000534 47.000534
2016-05-01 00:40:00 60 0 21.100000 38.030000 18.290000 41.560000 23.200000 36.500000 19.70 37.500000 ... 19.166667 38.633333 4.233333 763.766667 87.333333 2.666667 32.666667 2.266667 8.059441 8.059441
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ...
2016-05-27 17:20:00 100 0 25.566667 46.560000 25.890000 42.025714 27.200000 41.163333 24.70 45.590000 ... 23.200000 46.790000 22.733333 755.200000 55.666667 3.333333 23.666667 13.333333 43.096812 43.096812
2016-05-27 17:30:00 90 0 25.500000 46.500000 25.754000 42.080000 27.133333 41.223333 24.70 45.590000 ... 23.200000 46.790000 22.600000 755.200000 56.000000 3.500000 24.500000 13.300000 49.282940 49.282940
2016-05-27 17:40:00 270 10 25.500000 46.596667 25.628571 42.768571 27.050000 41.690000 24.70 45.730000 ... 23.200000 46.790000 22.466667 755.200000 56.333333 3.666667 25.333333 13.266667 29.199117 29.199117
2016-05-27 17:50:00 420 10 25.500000 46.990000 25.414000 43.036000 26.890000 41.290000 24.70 45.790000 ... 23.200000 46.817500 22.333333 755.200000 56.666667 3.833333 26.166667 13.233333 6.322784 6.322784
2016-05-27 18:00:00 430 10 25.500000 46.600000 25.264286 42.971429 26.823333 41.156667 24.70 45.963333 ... 23.200000 46.845000 22.200000 755.200000 57.000000 4.000000 27.000000 13.200000 34.118851 34.118851

3853 rows × 28 columns

df = df.loc['2016-05-01':]

df = df.round(2)

df
Appliances lights T1 RH_1 T2 RH_2 T3 RH_3 T4 RH_4 ... T9 RH_9 T_out Press_mm_hg RH_out Windspeed Visibility Tdewpoint rv1 rv2
date
2016-05-01 00:00:00 50 0 21.20 38.20 18.39 41.40 23.20 36.40 19.73 37.73 ... 19.17 38.20 4.30 763.70 86.00 2.00 40.00 2.20 15.31 15.31
2016-05-01 00:10:00 60 0 21.15 38.17 18.39 41.50 23.13 36.47 19.70 37.66 ... 19.13 38.29 4.28 763.72 86.33 2.17 38.17 2.22 11.24 11.24
2016-05-01 00:20:00 50 0 21.13 38.09 18.32 41.50 23.20 36.50 19.70 37.59 ... 19.13 38.36 4.27 763.73 86.67 2.33 36.33 2.23 21.81 21.81
2016-05-01 00:30:00 50 0 21.10 38.09 18.29 41.50 23.20 36.50 19.70 37.50 ... 19.17 38.50 4.25 763.75 87.00 2.50 34.50 2.25 47.00 47.00
2016-05-01 00:40:00 60 0 21.10 38.03 18.29 41.56 23.20 36.50 19.70 37.50 ... 19.17 38.63 4.23 763.77 87.33 2.67 32.67 2.27 8.06 8.06
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ...
2016-05-27 17:20:00 100 0 25.57 46.56 25.89 42.03 27.20 41.16 24.70 45.59 ... 23.20 46.79 22.73 755.20 55.67 3.33 23.67 13.33 43.10 43.10
2016-05-27 17:30:00 90 0 25.50 46.50 25.75 42.08 27.13 41.22 24.70 45.59 ... 23.20 46.79 22.60 755.20 56.00 3.50 24.50 13.30 49.28 49.28
2016-05-27 17:40:00 270 10 25.50 46.60 25.63 42.77 27.05 41.69 24.70 45.73 ... 23.20 46.79 22.47 755.20 56.33 3.67 25.33 13.27 29.20 29.20
2016-05-27 17:50:00 420 10 25.50 46.99 25.41 43.04 26.89 41.29 24.70 45.79 ... 23.20 46.82 22.33 755.20 56.67 3.83 26.17 13.23 6.32 6.32
2016-05-27 18:00:00 430 10 25.50 46.60 25.26 42.97 26.82 41.16 24.70 45.96 ... 23.20 46.84 22.20 755.20 57.00 4.00 27.00 13.20 34.12 34.12

3853 rows × 28 columns

len(df)

3853

# How many rows per day? We know its every 10 min
24*60/10

144.0

test_days = 2
test_ind = test_days * 144
test_ind

288

train = df.iloc[:-test_ind]
test = df.iloc[-test_ind:]

train
Appliances lights T1 RH_1 T2 RH_2 T3 RH_3 T4 RH_4 ... T9 RH_9 T_out Press_mm_hg RH_out Windspeed Visibility Tdewpoint rv1 rv2
date
2016-05-01 00:00:00 50 0 21.20 38.20 18.39 41.40 23.20 36.40 19.73 37.73 ... 19.17 38.20 4.30 763.70 86.00 2.00 40.00 2.20 15.31 15.31
2016-05-01 00:10:00 60 0 21.15 38.17 18.39 41.50 23.13 36.47 19.70 37.66 ... 19.13 38.29 4.28 763.72 86.33 2.17 38.17 2.22 11.24 11.24
2016-05-01 00:20:00 50 0 21.13 38.09 18.32 41.50 23.20 36.50 19.70 37.59 ... 19.13 38.36 4.27 763.73 86.67 2.33 36.33 2.23 21.81 21.81
2016-05-01 00:30:00 50 0 21.10 38.09 18.29 41.50 23.20 36.50 19.70 37.50 ... 19.17 38.50 4.25 763.75 87.00 2.50 34.50 2.25 47.00 47.00
2016-05-01 00:40:00 60 0 21.10 38.03 18.29 41.56 23.20 36.50 19.70 37.50 ... 19.17 38.63 4.23 763.77 87.33 2.67 32.67 2.27 8.06 8.06
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ...
2016-05-25 17:20:00 120 0 24.50 37.22 24.13 34.30 25.20 37.64 24.36 38.29 ... 21.89 37.03 16.17 756.17 52.67 1.33 31.33 6.43 33.46 33.46
2016-05-25 17:30:00 190 0 24.50 37.16 24.10 34.30 25.20 37.55 24.29 38.16 ... 21.89 37.20 16.25 756.15 53.50 1.50 33.50 6.75 0.43 0.43
2016-05-25 17:40:00 160 0 24.50 37.43 24.10 34.43 25.14 37.28 24.29 38.00 ... 21.89 37.33 16.33 756.13 54.33 1.67 35.67 7.07 16.67 16.67
2016-05-25 17:50:00 90 0 24.50 37.63 24.03 34.43 25.10 36.99 24.29 37.93 ... 22.00 37.36 16.42 756.12 55.17 1.83 37.83 7.38 39.36 39.36
2016-05-25 18:00:00 100 0 24.50 38.00 24.00 34.40 25.10 36.73 24.29 37.86 ... 22.00 37.36 16.50 756.10 56.00 2.00 40.00 7.70 38.63 38.63

3565 rows × 28 columns

Scale Data

from sklearn.preprocessing import MinMaxScaler

scaler = MinMaxScaler()

scaler.fit(train)

MinMaxScaler()

scaled_train = scaler.transform(train)
scaled_test = scaler.transform(test)

Time Series Generator

from tensorflow.keras.preprocessing.sequence import TimeseriesGenerator

# define generator
length = 144 # Length of the output sequences (in number of timesteps)
batch_size = 1 # Number of timeseries samples in each batch
generator = TimeseriesGenerator(scaled_train, scaled_train, length=length, batch_size=batch_size)

len(scaled_train)

3565

len(generator)

3421

# What does the first batch look like?
X, y = generator[0]

print(f'Given the Array: \n{X.flatten()}')
print(f'Predict this y: \n {y}')

Given the Array: 
[0.03896104 0.         0.13798978 ... 0.14319527 0.75185111 0.75185111]
Predict this y: 
 [[0.03896104 0.         0.30834753 0.29439421 0.16038492 0.49182278
  0.0140056  0.36627907 0.24142857 0.24364791 0.12650602 0.36276002
  0.12       0.28205572 0.06169297 0.15759185 0.34582624 0.39585974
  0.09259259 0.39649608 0.18852459 0.96052632 0.59210526 0.1
  0.58333333 0.13609467 0.4576746  0.4576746 ]]

Create the Model

from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import Dense, LSTM

scaled_train.shape

(3565, 28)

# define model
model = Sequential()
model.add(LSTM(25, input_shape=(length, scaled_train.shape[1])))
model.add(Dense(scaled_train.shape[1]))
model.compile(optimizer='adam', loss='mse')

model.summary()

Model: "sequential"
_________________________________________________________________
Layer (type)                 Output Shape              Param #   
=================================================================
lstm (LSTM)                  (None, 25)                5400      
_________________________________________________________________
dense (Dense)                (None, 28)                728       
=================================================================
Total params: 6,128
Trainable params: 6,128
Non-trainable params: 0
_________________________________________________________________

EarlyStopping

from tensorflow.keras.callbacks import EarlyStopping
es = EarlyStopping(monitor='val_loss', patience=1)
validation_generator = TimeseriesGenerator(scaled_test, scaled_test,
                                          length=length, batch_size=batch_size)

model.fit_generator(generator, epochs=10,
                   validation_data = validation_generator,
                   callbacks=[es])

WARNING:tensorflow:From <ipython-input-32-d0b564e3a4b2>:3: Model.fit_generator (from tensorflow.python.keras.engine.training) is deprecated and will be removed in a future version.
Instructions for updating:
Please use Model.fit, which supports generators.
Epoch 1/10
3421/3421 [==============================] - 13s 4ms/step - loss: 0.0172 - val_loss: 0.0131
Epoch 2/10
3421/3421 [==============================] - 13s 4ms/step - loss: 0.0090 - val_loss: 0.0106
Epoch 3/10
3421/3421 [==============================] - 13s 4ms/step - loss: 0.0081 - val_loss: 0.0092
Epoch 4/10
3421/3421 [==============================] - 13s 4ms/step - loss: 0.0077 - val_loss: 0.0087
Epoch 5/10
3421/3421 [==============================] - 13s 4ms/step - loss: 0.0074 - val_loss: 0.0086
Epoch 6/10
3421/3421 [==============================] - 13s 4ms/step - loss: 0.0073 - val_loss: 0.0080
Epoch 7/10
3421/3421 [==============================] - 13s 4ms/step - loss: 0.0072 - val_loss: 0.0086





<tensorflow.python.keras.callbacks.History at 0x26e52aa8cc8>

model.history.history.keys()

dict_keys(['loss', 'val_loss'])

losses = pd.DataFrame(model.history.history)
losses.plot()

<AxesSubplot:>

png

Evaluate on Test Data

first_eval_batch = scaled_train[-length:]

first_eval_batch

array([[0.1038961 , 0.        , 0.72231687, ..., 0.53550296, 0.15909546,
        0.15909546],
       [0.11688312, 0.        , 0.73424191, ..., 0.52662722, 0.40344207,
        0.40344207],
       [0.11688312, 0.        , 0.73424191, ..., 0.51775148, 0.20452271,
        0.20452271],
       ...,
       [0.18181818, 0.        , 0.70017036, ..., 0.50118343, 0.33340004,
        0.33340004],
       [0.09090909, 0.        , 0.70017036, ..., 0.51952663, 0.78747248,
        0.78747248],
       [0.1038961 , 0.        , 0.70017036, ..., 0.53846154, 0.77286372,
        0.77286372]])

first_eval_batch = first_eval_batch.reshape((1, length, scaled_train.shape[1]))

model.predict(first_eval_batch)

array([[ 0.12962861,  0.11173069,  0.7100676 ,  0.39257962,  0.5337119 ,
         0.4407935 ,  0.43118235,  0.4344098 ,  0.7135023 ,  0.3761254 ,
         0.328368  ,  0.37035555,  0.6467758 , -0.00678065,  0.62732697,
         0.40806448,  0.50039965,  0.33460236,  0.5257516 ,  0.35319155,
         0.6283157 ,  0.5327904 ,  0.43472147,  0.19735846,  0.6059718 ,
         0.5548344 ,  0.5132371 ,  0.5090848 ]], dtype=float32)

scaled_test[0]

array([0.19480519, 0.        , 0.70017036, 0.3920434 , 0.53007217,
       0.41064526, 0.40616246, 0.41913319, 0.72714286, 0.4115245 ,
       0.30722892, 0.36445121, 0.66777778, 0.        , 0.61119082,
       0.39840637, 0.51618399, 0.32953105, 0.53703704, 0.34024896,
       0.6057377 , 0.52631579, 0.41881579, 0.2       , 0.55283333,
       0.53372781, 0.76305783, 0.76305783])

n_features = scaled_train.shape[1]
test_predictions = []

first_eval_batch = scaled_train[-length:]
current_batch = first_eval_batch.reshape((1, length, n_features))

for i in range(len(test)):
    current_pred = model.predict(current_batch)[0]
    test_predictions.append(current_pred)
    current_batch = np.append(current_batch[:, 1:, :],[[current_pred]], axis=1)

scaled_test

array([[0.19480519, 0.        , 0.70017036, ..., 0.53372781, 0.76305783,
        0.76305783],
       [0.37662338, 0.        , 0.70017036, ..., 0.52840237, 0.62337402,
        0.62337402],
       [0.12987013, 0.        , 0.70017036, ..., 0.52366864, 0.08785271,
        0.08785271],
       ...,
       [0.32467532, 0.33333333, 0.87052811, ..., 0.86804734, 0.58415049,
        0.58415049],
       [0.51948052, 0.33333333, 0.87052811, ..., 0.86568047, 0.12627577,
        0.12627577],
       [0.53246753, 0.33333333, 0.87052811, ..., 0.86390533, 0.68260957,
        0.68260957]])

Inverse Transformations and Compare

true_predictions = scaler.inverse_transform(test_predictions)

true_predictions

array([[119.81403247,   3.35192084,  24.55809664, ...,   7.97670179,
         25.65645882,  25.44896849],
       [135.39294168,   5.15252039,  24.49285675, ...,   8.37637454,
         25.66081033,  25.26328633],
       [144.21259135,   6.69974104,  24.44082725, ...,   8.80649047,
         25.73894385,  25.18596294],
       ...,
       [222.21735567,  16.03211939,  24.13516293, ...,  21.59099305,
         23.42215515,  22.95250515],
       [222.21735567,  16.03211939,  24.13516363, ...,  21.59099506,
         23.42215218,  22.95250515],
       [222.21733272,  16.03211939,  24.13516398, ...,  21.59099103,
         23.42215218,  22.95250217]])

pred_df = pd.DataFrame(data=true_predictions, index=test.index, columns = test.columns)

test
Appliances lights T1 RH_1 T2 RH_2 T3 RH_3 T4 RH_4 ... T9 RH_9 T_out Press_mm_hg RH_out Windspeed Visibility Tdewpoint rv1 rv2
date
2016-05-25 18:10:00 170 0 24.50 37.86 24.00 34.27 25.00 36.70 24.29 37.79 ... 22.00 37.23 16.48 756.1 55.83 2.00 38.17 7.62 38.14 38.14
2016-05-25 18:20:00 310 0 24.50 37.30 23.86 34.33 24.94 36.67 24.29 37.79 ... 22.00 37.36 16.47 756.1 55.67 2.00 36.33 7.53 31.16 31.16
2016-05-25 18:30:00 120 0 24.50 36.96 23.73 34.33 24.85 36.50 24.29 37.79 ... 22.03 37.39 16.45 756.1 55.50 2.00 34.50 7.45 4.40 4.40
2016-05-25 18:40:00 120 0 24.50 37.00 23.70 34.40 24.84 36.45 24.29 37.90 ... 22.10 37.72 16.43 756.1 55.33 2.00 32.67 7.37 27.12 27.12
2016-05-25 18:50:00 120 0 24.49 37.07 23.68 34.52 24.84 36.49 24.28 37.93 ... 22.10 37.81 16.42 756.1 55.17 2.00 30.83 7.28 10.27 10.27
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ...
2016-05-27 17:20:00 100 0 25.57 46.56 25.89 42.03 27.20 41.16 24.70 45.59 ... 23.20 46.79 22.73 755.2 55.67 3.33 23.67 13.33 43.10 43.10
2016-05-27 17:30:00 90 0 25.50 46.50 25.75 42.08 27.13 41.22 24.70 45.59 ... 23.20 46.79 22.60 755.2 56.00 3.50 24.50 13.30 49.28 49.28
2016-05-27 17:40:00 270 10 25.50 46.60 25.63 42.77 27.05 41.69 24.70 45.73 ... 23.20 46.79 22.47 755.2 56.33 3.67 25.33 13.27 29.20 29.20
2016-05-27 17:50:00 420 10 25.50 46.99 25.41 43.04 26.89 41.29 24.70 45.79 ... 23.20 46.82 22.33 755.2 56.67 3.83 26.17 13.23 6.32 6.32
2016-05-27 18:00:00 430 10 25.50 46.60 25.26 42.97 26.82 41.16 24.70 45.96 ... 23.20 46.84 22.20 755.2 57.00 4.00 27.00 13.20 34.12 34.12

288 rows × 28 columns

pred_df
Appliances lights T1 RH_1 T2 RH_2 T3 RH_3 T4 RH_4 ... T9 RH_9 T_out Press_mm_hg RH_out Windspeed Visibility Tdewpoint rv1 rv2
date
2016-05-25 18:10:00 119.814032 3.351921 24.558097 37.874826 24.045388 35.283886 25.178642 36.989033 24.194516 37.009804 ... 21.939059 37.510725 17.030903 756.247622 57.038832 1.973585 41.358309 7.976702 25.656459 25.448968
2016-05-25 18:20:00 135.392942 5.152520 24.492857 37.657199 23.754765 35.277770 25.187547 37.040140 24.122789 36.663892 ... 21.886550 37.796598 17.147620 756.141145 57.685050 1.954602 41.986854 8.376375 25.660810 25.263286
2016-05-25 18:30:00 144.212591 6.699741 24.440827 37.524560 23.471283 35.436711 25.215640 37.063246 24.070092 36.440047 ... 21.844848 38.057945 17.257024 755.974725 58.609539 1.932213 42.552782 8.806490 25.738944 25.185963
2016-05-25 18:40:00 149.768802 8.069998 24.386769 37.446713 23.180236 35.667879 25.250065 37.088631 24.017115 36.277079 ... 21.810150 38.261318 17.332690 755.705818 59.766602 1.899848 42.988523 9.248289 25.786471 25.105795
2016-05-25 18:50:00 153.218038 9.331562 24.329630 37.435610 22.893000 35.979222 25.289133 37.152087 23.964464 36.188050 ... 21.782987 38.419516 17.376678 755.337311 61.196783 1.864786 43.348687 9.716914 25.818307 25.025800
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ...
2016-05-27 17:20:00 222.217310 16.032119 24.135162 60.898420 22.814009 60.913240 27.341344 52.462853 25.639273 52.654667 ... 27.455314 61.039082 22.957183 722.226312 84.129677 3.372346 56.142634 21.590995 23.422155 22.952505
2016-05-27 17:30:00 222.217333 16.032119 24.135162 60.898420 22.814009 60.913240 27.341345 52.462853 25.639273 52.654667 ... 27.455314 61.039084 22.957180 722.226312 84.129681 3.372346 56.142634 21.590993 23.422155 22.952502
2016-05-27 17:40:00 222.217356 16.032119 24.135163 60.898420 22.814007 60.913240 27.341345 52.462853 25.639273 52.654667 ... 27.455313 61.039084 22.957180 722.226312 84.129677 3.372346 56.142631 21.590993 23.422155 22.952505
2016-05-27 17:50:00 222.217356 16.032119 24.135164 60.898420 22.814009 60.913240 27.341347 52.462853 25.639273 52.654667 ... 27.455313 61.039087 22.957183 722.226312 84.129672 3.372346 56.142631 21.590995 23.422152 22.952505
2016-05-27 18:00:00 222.217333 16.032119 24.135164 60.898420 22.814011 60.913240 27.341347 52.462853 25.639273 52.654667 ... 27.455313 61.039092 22.957183 722.226309 84.129672 3.372347 56.142631 21.590991 23.422152 22.952502

288 rows × 28 columns

columns = pred_df.columns

for col in columns:
    plt.plot(pred_df[col], label='pred')
    plt.plot(test[col], label='real')
    plt.legend()
    plt.title(col)
    plt.show()

png

png

png

png

png

png

png

png

png

png

png

png

png

png

png

png

png

png

png

png

png

png

png

png

png

png

png

png