Cloud - Specialization in Analytics and Data Science.

Presented by:

• Danilo Diaz Valencia: danilo.diaz@udea.edu.co

• Santiago Jaramillo: santiago.jaramillog1@udea.edu.co

Titanic survival_probability PREDICTION

The current exercise involves predicting passenger survival in the Titanic catastrophe. The objective of this project is to deploy a web service on the public Azure cloud using AzureML. This service is capable of predicting the probability of passenger survival based on the gender and class they were traveling in. link

Data Description

Titanic Dataset

• passenger_id: Passenger ID
• pclass: Ticket class: 1 = 1st, 2 = 2nd, 3 = 3rd
• name: Passenger name
• sex: Passenger gender
• age: Passenger age
• sibsp: Siblings aboard
• parch: Parents/Children aboard
• ticket: Ticket number
• fare: Passenger fare
• cabin: Cabin number
• embarked: Port of embarkation: C = Cherbourg, Q = Queenstown, S = Southampton
• survival: Survived 0 = No, 1 = Yes

Requirements

• Azure subscription.
• Docker, only for local environment, recommended for debugging processes.
• Create environment
• Python 3.8 -
• IDE vs code or equivalent.

How to run it?

• Create a new environment
• Install the following dependencies

pip install -r requirements.txt

pip install azureml-core
pip install azureml
pip install azureml-contrib-services
pip install numpy
pip install pandas
pip install scikit-learn
pip install scikit-multilearn

• Execute the first script to create the environment

python ./01-create-environment.py

• Follow the steps of interactive authentication and wait for the script to provision us with a machine learning environment. At the end, there should be a configuration file within a folder called .azureml with the team's parameters.

• The configuration after the first script should generate the following resources within the group: Azure Machine Learning, Container registry, Application Insights, Key Vault, Storage Account.

NOTE: If desired, Azure CLI can be used!

• Execute the second script to provision the Machine Learning environment with a cluster to manage computation. Once the script finishes, a new cluster named "cpu-cluster" should appear in the compute tab of the machine learning general panel.

python ./02-create-compute.py

• Execute the third script to verify the workspace environment

python ./03-test-workspace-remote.py

• Run the fourth script to register the model in Azure

python ./04-azure-model-registration.py

• Run the fifth script to deploy the model

python ./05-deploy-azure-model-aci.py

Azure Service Test

The API is for educational purposes only and is therefore only available for a limited time!!!

• API Link Api-Titanic

{
    "data": ["mele", 4]
}

Postman Test

http://958b1618-b579-41c9-aee1-9b4582414e15.eastus2.azurecontainer.io/score