黑料情报站

Track the Resource/Personal Real Time

The need for these many technical breakthroughs has increased due to the steadily expanding demands for security, safety, communication, and flawless company operations. Safety becomes a significant concern when you take the industrial and construction industries into account. The majority of the world's building industries are seeing unprecedented growth. And integrating such cutting-edge technology into the building site would not only ensure safety but also let the business monitor the activity of various personnel remotely.

Smart device for construction field. Effective Monitoring product for Construction Workers:

Getting the most out of their employees is one of the most important goals for construction site managers. They are aware that if the site's employees take excessive pauses between tasks or waste time talking about unimportant things, they will not finish the day's job by the deadline. This GPS-equipped smart gadget enables managers to keep an eye on every action taken by their employees. For instance, by knowing which helmets are in use, you can determine whether an employee is there. The signal will let you know if any employee stays longer than necessary at the same location. Additionally, you may determine each worker's task and location according to their function and determine whether or not they are actively engaged in their assigned work on a certain floor of the building. In the end, this results in increased productivity of the particular work.

DATA GATHERING : OUTPUT OBTAINED FROM THESE SENSORS:

GPS stands for Global Positioning System and used to detect the Latitude and Longitude of any location on the Earth, with exact UTC time (Universal Time Coordinated).

RFID:Some UHF RFID tags are delivered from the manufacturer

TEMPERATURE SENSOR

HEART BEAT SENSOR

MQ5 GAS SENSOR

AGE

WORKING DATE

WORKING HRS

EMP ID

EDA and data preprocessing :

PAIR PLOT

In a dataset, plot pairwise relationships. By default, this function will generate a grid of axes with each numerical variable in the data spread over a single row and column on the y-axis and a single column on the x-axis. Different rules are used to the diagonal axes when creating a plot to display the data's univariate distribution for the parameter in that column.

BAR GRAPH:

A bar graph is a diagram that uses rectangular bars with heights or lengths proportionate to the values they represent to display categorical data. Both a vertical and a horizontal bar plot are possible. A column chart is another name for a vertical bar graph. Bar graphs are used to track changes over time or to compare data between groups. However, bar graphs work better when the changes are more significant when trying to gauge change over time. Figure #bar

Data preprocessing

It is a data mining technique which is used to transform the raw data in a useful and efficient format. Our Data set consists the following Columns: EmployeeID, Work_date, Age, Working_hours, Temperature, GPS Location N, GPS Location S, MQ5 RFID with 255 records. Here some of the columns are removed which are not required for , Finally we are taking the following columns: Age, Temperature,Heartbeat sensor, MQ5.

iot_refined = iot.iloc[:,[2,4,7,8]] #removing columns which are not required

iot_refined.sample(5)

Criteria to define the output classes

iot_refined['performance'] = ["Idle" if i<88 or i>157 else "Working" for i in

iot_refined['Heart_beat']]

iot_refined.loc[iot_refined.Temperature>38,['performance']]="Idle"

iot_refined.loc[iot_refined.MQ5>665,['performance']]="Idle"

#Defining Independent and Dependent Variables :

X = iot_refined.iloc[:, 0:4]

Y = iot_refined['performance']

#Designing Models and Assigning them to a variable :

logit = LogisticRegression()

svm = SVC(kernel='linear', gamma=0.01, probability= True)

knn = KNeighborsClassifier(n_neighbors = 6)

mlp = MLPClassifier(hidden_layer_sizes=(100, ), learning_rate_init=0.02)

bayes = GaussianNB()

#Building an Ensemble Model - Stacking Classifier :

#define the base models

level0 = list()

level0.append(('logit', logit))

level0.append(('svm', svm))

level0.append(('knn', knn))

level0.append(('nb', bayes))

#define meta learner model

level1 = LogisticRegression()

#define the stacking ensemble

stack = StackingClassifier(estimators=level0, final_estimator=level1, cv=5)

A CustomFunction That Returns The Models on a Function Call :

def get_models():

models=dict() #Creatinganemptypython

dictionary models['logit'] = logit

models['svm'] = svm

models['knn'] = knn

models['mlp'] = mlp

models['nb'] = bayes

models['stacking'] = stack

return models

A Custom Function to Give Us the Cross Validation Scores in Helping with the Model Selection :

def evaluate_model(model):

skf = StratifiedKFold(n_splits=10, random_state = 1)

scores = cross_val_score(model, X, Y, scoring='accuracy', cv=skf, n_jobs=-1, error_score='raise')

return scores

Calculating theCross Validation Scores for Different Models:

# get the models to evaluate

models = get_models()

# evaluate the models and store results

#creating two empty lists to store the name and scores of each model

results, names = list(), list()

for name, model in models.items():

scores = evaluate_model(model)

results.append(scores)

names.append(name)

print(name,':', mean(scores))

Building Model with the Stacking Classifier :

stack.fit(X,Y) #fitting the model with the data

Conclusion

The following benefits of using Python Flask to deploy the model increased productivity, efficient staff management, worker safety, health monitoring, and worker trust. We can design stunning dashboards and monitor an employee's working environment.