In [ ]:
#lineplot for themperature changes for time
plt.figure(figsize=(20,10))
sns.lineplot(x='date',y='Temperature',data=df)
plt.show()
The aim of this project is to predict whether a room is occupied or not based on the data collected from the sensors. The data set is collected from the UCI Machine Learning Repository. The data set contains 7 attributes. The attributes are date, temperature, humidity, light, CO2, humidity ratio and occupancy. The data set is divided into 3 data sets for training and testing. The data set provides experimental data used for binary classification (room occupancy of an office room) from Temperature, Humidity, Light and CO2. Ground-truth occupancy was obtained from time stamped pictures that were taken every minute.
| Column Position | Atrribute Name | Definition | Data Type | Example | % Null Ratios |
|---|---|---|---|---|---|
| 1 | Date | Date & time in year-month-day hour:minute:second format | Qualitative | 2/4/2015 17:57, 2/4/2015 17:55, 2/4/2015 18:06 | 0 |
| 2 | Temperature | Temperature in degree Celcius | Quantitative | 23.150, 23.075, 22.890 | 0 |
| 3 | Humidity | Relative humidity in percentage | Quantitative | 27.272000, 27.200000, 27.390000 | 0 |
| 4 | Light | Illuminance measurement in unit Lux | Quantitative | 426.0, 419.0, 0.0 | 0 |
| 5 | CO2 | CO2 in parts per million (ppm) | Quantitative | 489.666667, 495.500000, 534.500000 | 0 |
| 6 | HumidityRatio | Humadity ratio: Derived quantity from temperature and relative humidity, in kgwater-vapor/kg-air | Quantitative | 0.004986, 0.005088, 0.005203 | 0 |
| 7 | Occupancy | Occupied or not: 1 for occupied and 0 for not occupied | Quantitative | 1, 0 | 0 |
#importing the libraries
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
Loading two datasets and combining them into one dataset
#loading the datasets
df1 = pd.read_csv('datatest.csv')
df2 = pd.read_csv('datatraining.csv')
#combining the datasets
df = pd.concat([df1,df2])
df.head()
| date | Temperature | Humidity | Light | CO2 | HumidityRatio | Occupancy | |
|---|---|---|---|---|---|---|---|
| 0 | 2/2/2015 14:19 | 23.7000 | 26.272 | 585.200000 | 749.200000 | 0.004764 | 1 |
| 1 | 2/2/2015 14:19 | 23.7180 | 26.290 | 578.400000 | 760.400000 | 0.004773 | 1 |
| 2 | 2/2/2015 14:21 | 23.7300 | 26.230 | 572.666667 | 769.666667 | 0.004765 | 1 |
| 3 | 2/2/2015 14:22 | 23.7225 | 26.125 | 493.750000 | 774.750000 | 0.004744 | 1 |
| 4 | 2/2/2015 14:23 | 23.7540 | 26.200 | 488.600000 | 779.000000 | 0.004767 | 1 |
#number of rows and columns
df.shape
(10808, 7)
#checking for null values
df.isnull().sum()
date 0 Temperature 0 Humidity 0 Light 0 CO2 0 HumidityRatio 0 Occupancy 0 dtype: int64
#checking for duplicate values
df.duplicated().sum()
27
#removing the duplicate values
df.drop_duplicates(inplace=True)
#checking data types
df.dtypes
date object Temperature float64 Humidity float64 Light float64 CO2 float64 HumidityRatio float64 Occupancy int64 dtype: object
#converting the date and time to datetime format
df['date'] = pd.to_datetime(df['date'])
df.dtypes
date datetime64[ns] Temperature float64 Humidity float64 Light float64 CO2 float64 HumidityRatio float64 Occupancy int64 dtype: object
#checking the descriptive statistics
df.describe()
| date | Temperature | Humidity | Light | CO2 | HumidityRatio | Occupancy | |
|---|---|---|---|---|---|---|---|
| count | 10781 | 10781.000000 | 10781.000000 | 10781.000000 | 10781.000000 | 10781.000000 | 10781.000000 |
| mean | 2015-02-06 13:41:14.581207808 | 20.821800 | 25.638618 | 138.036704 | 634.460328 | 0.003904 | 0.250533 |
| min | 2015-02-02 14:19:00 | 19.000000 | 16.745000 | 0.000000 | 412.750000 | 0.002674 | 0.000000 |
| 25% | 2015-02-04 18:23:00 | 20.000000 | 21.390000 | 0.000000 | 441.000000 | 0.003323 | 0.000000 |
| 50% | 2015-02-06 15:24:00 | 20.700000 | 25.680000 | 0.000000 | 464.000000 | 0.003805 | 0.000000 |
| 75% | 2015-02-08 12:29:00 | 21.500000 | 28.323333 | 415.000000 | 763.000000 | 0.004373 | 1.000000 |
| max | 2015-02-10 09:33:00 | 24.408333 | 39.117500 | 1697.250000 | 2028.500000 | 0.006476 | 1.000000 |
| std | NaN | 1.078589 | 4.954838 | 212.330275 | 313.074686 | 0.000803 | 0.433340 |
#value counts for the target variable i.e. occupancy
df['Occupancy'].value_counts()
Occupancy 0 8080 1 2701 Name: count, dtype: int64
In the exploratory data analysis, we will be looking at the distribution of the data, along with the time series of the data. We will also be looking at the correlation between the variables.
#lineplot for themperature changes for time
plt.figure(figsize=(20,10))
sns.lineplot(x='date',y='Temperature',data=df)
plt.show()
The spikes in the graph clearly indicates that the room temperature incresases suddenly which might be due to the presence of people in the room. The temperature of the room may increase due to the heat emitted by the human body.
#lineplot for humidity changes for time
plt.figure(figsize=(20,10))
sns.lineplot(x='date',y='Humidity',data=df)
plt.show()
The line graph between 3rd of February to 6th of February shows some similarity with the temperature graph, which might be due to the presence of people in the room. However 7th of February onwards there has been a significant rise in the humidity levels which might be due to cleaning of the room, or change in the weather conditions. Out of which room cleaning such sweeping the floor might be the reason for the sudden rise in the humidity levels. But it couldn't explain the increase in the humidity levels near 10th of February.
#lineplot for light changes for time
plt.figure(figsize=(20,10))
sns.lineplot(x='date',y='Light',data=df)
plt.show()
If we look closely, we can see that the number of peaks in this graph and in the temperature graph are same. This indicates that lights were turned on when there was a person in the room. This is a good indicator of the occupancy of the room.
#lineplot for co2 changes for time
plt.figure(figsize=(20,10))
sns.lineplot(x='date',y='CO2',data=df)
plt.show()
The co2 graph also shows the spikes in the co2 levels which indicates the presence of person in the room, assuming that there is no other source of co2 in the room.In addition to that the spikes also shows correspondence with the temperature graph and light graph. However from 7th of February to 9th of February, the co2 levels where minimum, which indicstes that the room was not occupied during that time. This observation contradicts with the humidity graph and temperature graph.
#lineplot for humidity ratio changes for time
plt.figure(figsize=(20,10))
sns.lineplot(x='date',y='HumidityRatio',data=df)
plt.show()
The humidity ratio graph is quite similar to the humidity graph. The spikes in the graph indicates the presence of people in the room. Moreover the same assumption is made about the humidity ratio after 9th of February.
#correlation heatmap
plt.figure(figsize=(20,10))
sns.heatmap(df.corr(),annot=True)
plt.show()
There is a strong coorelation between light and occupancy as well as between humidity and humidity ratio. The co2 levels and temperature also shows a strong correlation with the occupancy. However the humidity and humidity ratio has very less correlation with the occupancy.
#violinplot for temperature
sns.violinplot(y = df['Temperature'],x = df['Occupancy'])
plt.xlabel('Occupancy')
plt.ylabel('Temperature')
plt.show()
The temperature and occupancy graph shows that the temperature of the room increases when there is a person in the room. This is because of the heat emitted by the human body. The temperature of the room decreases when there is no person in the room. This proves the hypothesis regarding the peaks in the temperature graph.
#boxplot for light
sns.boxplot(y = df['Light'],x = df['Occupancy'])
plt.xlabel('Occupancy')
plt.ylabel('Light')
plt.show()
The light intensity of the room increases when there is a person in the room. This is because the lights are turned on when there is a person in the room. The light intensity of the room decreases when there is no person in the room. This proves the hypothesis regarding the peaks in the light graph. The outliers in the boxplot and the curves in the ligh graph might be due to sunlight entering the room.
#violinlot for co2
sns.violinplot(y = df['CO2'],x = df['Occupancy'])
plt.xlabel('Occupancy')
plt.ylabel('CO2')
plt.show()
The CO2 levels of the room increases when there is a person in the room. This is because of the CO2 emitted by the human body. The CO2 levels of the room decreases when there is no person in the room. This proves the hypothesis regarding the peaks in the CO2 graph.
From the above EDA, it is quite clear that the temperature, light and CO2 levels of the room are a good indicator of the occupancy of the room. Therefore we will be using these three variables for our classification model.
#dropping columns humidity, date and humidity ratio
df.drop(['Humidity','date','HumidityRatio'],axis=1,inplace=True)
df.head(10)
| Temperature | Light | CO2 | Occupancy | |
|---|---|---|---|---|
| 0 | 23.7000 | 585.200000 | 749.200000 | 1 |
| 1 | 23.7180 | 578.400000 | 760.400000 | 1 |
| 2 | 23.7300 | 572.666667 | 769.666667 | 1 |
| 3 | 23.7225 | 493.750000 | 774.750000 | 1 |
| 4 | 23.7540 | 488.600000 | 779.000000 | 1 |
| 5 | 23.7600 | 568.666667 | 790.000000 | 1 |
| 6 | 23.7300 | 536.333333 | 798.000000 | 1 |
| 7 | 23.7540 | 509.000000 | 797.000000 | 1 |
| 8 | 23.7540 | 476.000000 | 803.200000 | 1 |
| 9 | 23.7360 | 510.000000 | 809.000000 | 1 |
from sklearn.model_selection import train_test_split
x_train,x_test,y_train,y_test = train_test_split(df.drop(['Occupancy'],axis=1),df['Occupancy'],test_size=0.2,random_state=42)
from sklearn.ensemble import RandomForestClassifier
rfc = RandomForestClassifier()
rfc
RandomForestClassifier()In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook.
RandomForestClassifier()
#training the model
rfc.fit(x_train,y_train)
#training accuracy
rfc.score(x_train,y_train)
1.0
rfc_pred = rfc.predict(x_test)
#confusion matrix heatmap
from sklearn.metrics import confusion_matrix
sns.heatmap(confusion_matrix(y_test,rfc_pred),annot=True)
plt.ylabel('Predicted')
plt.xlabel('Actual')
plt.show()
#distribution plot for the predicted and actual values
ax = sns.distplot(y_test,hist=False,label='Actual', color='r')
sns.distplot(rfc_pred,hist=False,label='Predicted',color='b',ax=ax)
plt.show()
C:\Users\DELL\AppData\Local\Temp\ipykernel_17804\748502899.py:2: UserWarning: `distplot` is a deprecated function and will be removed in seaborn v0.14.0. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `kdeplot` (an axes-level function for kernel density plots). For a guide to updating your code to use the new functions, please see https://gist.github.com/mwaskom/de44147ed2974457ad6372750bbe5751 ax = sns.distplot(y_test,hist=False,label='Actual', color='r') C:\Users\DELL\AppData\Local\Temp\ipykernel_17804\748502899.py:3: UserWarning: `distplot` is a deprecated function and will be removed in seaborn v0.14.0. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `kdeplot` (an axes-level function for kernel density plots). For a guide to updating your code to use the new functions, please see https://gist.github.com/mwaskom/de44147ed2974457ad6372750bbe5751 sns.distplot(rfc_pred,hist=False,label='Predicted',color='b',ax=ax)
from sklearn.metrics import classification_report
print(classification_report(y_test,rfc_pred))
precision recall f1-score support
0 1.00 0.99 1.00 1623
1 0.97 1.00 0.99 534
accuracy 0.99 2157
macro avg 0.99 0.99 0.99 2157
weighted avg 0.99 0.99 0.99 2157
from sklearn.metrics import accuracy_score
from sklearn.metrics import precision_score
from sklearn.metrics import recall_score
from sklearn.metrics import f1_score
print('Accuracy Score : ' + str(accuracy_score(y_test,rfc_pred)))
print('Precision Score : ' + str(precision_score(y_test,rfc_pred)))
print('Recall Score : ' + str(recall_score(y_test,rfc_pred)))
print('F1 Score : ' + str(f1_score(y_test,rfc_pred)))
Accuracy Score : 0.9925822902178952 Precision Score : 0.9743589743589743 Recall Score : 0.9962546816479401 F1 Score : 0.9851851851851853
df_new = pd.read_csv('datatest2.csv')
df_new.head()
| date | Temperature | Humidity | Light | CO2 | HumidityRatio | Occupancy | |
|---|---|---|---|---|---|---|---|
| 0 | 2/11/2015 14:48 | 21.7600 | 31.133333 | 437.333333 | 1029.666667 | 0.005021 | 1 |
| 1 | 2/11/2015 14:49 | 21.7900 | 31.000000 | 437.333333 | 1000.000000 | 0.005009 | 1 |
| 2 | 2/11/2015 14:50 | 21.7675 | 31.122500 | 434.000000 | 1003.750000 | 0.005022 | 1 |
| 3 | 2/11/2015 14:51 | 21.7675 | 31.122500 | 439.000000 | 1009.500000 | 0.005022 | 1 |
| 4 | 2/11/2015 14:51 | 21.7900 | 31.133333 | 437.333333 | 1005.666667 | 0.005030 | 1 |
#dropping columns humidity, date and humidity ratio
df_new.drop(['Humidity','date','HumidityRatio'],axis=1,inplace=True)
#splitting the target variable
x = df_new.drop(['Occupancy'],axis=1)
y = df_new['Occupancy']
#predicting the values
pred = rfc.predict(x)
#confusion matrix heatmap
sns.heatmap(confusion_matrix(y,pred),annot=True)
plt.ylabel('Predicted')
plt.xlabel('Actual')
plt.show()
#distribution plot for the predicted and actual values
ax = sns.distplot(y,hist=False,label='Actual', color='r')
sns.distplot(pred,hist=False,label='Predicted',color='b',ax=ax)
plt.show()
C:\Users\DELL\AppData\Local\Temp\ipykernel_17804\4080147354.py:2: UserWarning: `distplot` is a deprecated function and will be removed in seaborn v0.14.0. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `kdeplot` (an axes-level function for kernel density plots). For a guide to updating your code to use the new functions, please see https://gist.github.com/mwaskom/de44147ed2974457ad6372750bbe5751 ax = sns.distplot(y,hist=False,label='Actual', color='r') C:\Users\DELL\AppData\Local\Temp\ipykernel_17804\4080147354.py:3: UserWarning: `distplot` is a deprecated function and will be removed in seaborn v0.14.0. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `kdeplot` (an axes-level function for kernel density plots). For a guide to updating your code to use the new functions, please see https://gist.github.com/mwaskom/de44147ed2974457ad6372750bbe5751 sns.distplot(pred,hist=False,label='Predicted',color='b',ax=ax)
print(classification_report(y,pred))
precision recall f1-score support
0 1.00 0.99 0.99 7703
1 0.96 0.99 0.97 2049
accuracy 0.99 9752
macro avg 0.98 0.99 0.98 9752
weighted avg 0.99 0.99 0.99 9752
print('Accuracy Score : ' + str(accuracy_score(y,pred)))
print('Precision Score : ' + str(precision_score(y,pred)))
print('Recall Score : ' + str(recall_score(y,pred)))
print('F1 Score : ' + str(f1_score(y,pred)))
Accuracy Score : 0.9883100902379 Precision Score : 0.9565832940066069 Recall Score : 0.9892630551488532 F1 Score : 0.9726487523992322
From the above models we can see that the Random Forest Classifier has the highest accuracy score of 98%. Therefore we will be using the Random Forest Classifier for our final model. I also conclude that from the exploratory data analysis, it was found that the change in room temperature, CO levels and light intensity can be used to predict the occupancy of the room, inplace of humidity and humidity ratio.