[go: up one dir, main page]
Scribd
Upload
3 views3 pages

Week_9_Deployment

The document outlines a series of videos focused on deploying results in data science projects, including anonymizing data, building applications, and hosting them. It covers tools like ARX for data anonymization, Streamlit for creating user interfaces, and Heroku for application deployment. Each video provides step-by-step guidance on the respective processes, ensuring users can effectively share and manage their data science projects.

Uploaded by

revanthkalla1
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
3 views3 pages

Week_9_Deployment

The document outlines a series of videos focused on deploying results in data science projects, including anonymizing data, building applications, and hosting them. It covers tools like ARX for data anonymization, Streamlit for creating user interfaces, and Heroku for application deployment. Each video provides step-by-step guidance on the respective processes, ensuring users can effectively share and manage their data science projects.

Uploaded by

revanthkalla1
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 3

Introduction

Deploy the results (Video 1)


Video summary

This video explains the final step in data science projects: deploying results. It covers
anonymizing data, building applications, and hosting them for public access.

Highlights:

• 00:133 Deploying results


o Share work with others
o Ensure data is anonymized
o Build and host applications
• 01:024 Anonymizing data
o Use tools like ARX
o Create fake data with libraries like Faker
o Ensure sensitive information is masked
• 04:465 Building applications
o Use notebooks, Streamlit, or full-fledged web apps
o Streamlit and Shiny for user-friendly apps
o Flask and Tornado for web apps
• 08:426 Hosting applications
o Host on platforms like Google Drive, GitHub, or Heroku
o Use AWS or Azure for more control
o Example of a Heroku app

Tools to anonymize data (Video 2)


Video summary

This video provides a comprehensive introduction to the ARX Data Anonymization Tool,
demonstrating its basic functionalities and how to use it for anonymizing data sets. The video
covers importing data, configuring anonymization processes, defining metadata, and
analyzing results.

Highlights:

• 00:003 Introduction to ARX Tool


o Overview of the graphical user interface
o Project-based structure
o Importing data from various sources
• 01:004 Configuring Anonymization
o Specifying attribute types and data types
o Creating generalization hierarchies
o Defining privacy criteria like K-anonymity
• 06:015 Anonymizing Data
o Setting up data utility measures
o Balancing generalization and suppression
o Running the anonymization process
• 07:006 Analyzing Results
o Comparing transformed data to original data
o Assessing data utility and privacy
o Viewing statistical information and metadata
• 09:047 Exploring Solution Space
o Browsing and organizing transformations
o Filtering and selecting optimal solutions
o Applying and exporting anonymized data

This summary captures the key steps and features of the ARX tool as demonstrated in the
video.

• List of Tools

Libraries to build web applications (Video 3)


Video summary

This video demonstrates how to deploy a credit card approval prediction task using Streamlit
and Google Colab. It covers the installation of necessary libraries, setting up the environment,
and creating a user interface for input and output.

Highlights:

• 00:003 Introduction to Streamlit


o Easy Python library for deploying ML models
o Allows creating interfaces for input and output
o Example: credit card approval prediction
• 00:484 Setting up the environment
o Install Streamlit and ngrok in Google Colab
o Authenticate ngrok with an ID from its website
o Use ngrok link to view the app locally
• 01:455 Creating the app interface
o Import necessary libraries like pandas and Streamlit
o Use st.write and st.header for text and headers
o Input methods: st.selectbox for categorical, st.number_input for continuous
variables
• 04:426 Handling user inputs
o Collect inputs as a dictionary
o Convert dictionary to pandas DataFrame
o Visualize inputs for user confirmation
• 05:137 Pre-processing and prediction
o Apply encoding and convert features to numerical values
o Load the stored model and make predictions
o Display prediction results and probabilities
This video provides a comprehensive guide to deploying a machine learning model using
Streamlit, making it accessible for users to interact with the model through a web interface.

• List of frameworks
• Code - Streamlit
• Notebook - Streamlit

Services to host web applications (Video 4)


Video summary

This video demonstrates how to deploy a Streamlit application using Heroku. It covers the
steps from logging into Heroku, connecting to a GitHub repository, and setting up the
necessary files and environment for deployment.

Highlights:

• 00:003 Introduction to Heroku and Streamlit deployment


o Heroku allows publishing web applications
o Login or create a free Heroku account
o Overview of the Heroku dashboard
• 01:094 Connecting to GitHub for deployment
o Choose GitHub as the deployment method
o Connect to the GitHub repository
o Ensure the repository is public
• 02:195 Setting up the project files
o Key files: Procfile, README.md, app.py, setup.sh, requirements.txt
o Importance of specifying packages in requirements.txt
o Overview of app.py and setup.sh
• 03:506 Configuring the Heroku environment
o Create a folder in the root directory
o Serve using the default port 8501
o Command to start the application
• 04:567 Deploying and accessing the application
o Deploy the branch from GitHub
o View build logs in Heroku
o Access the deployed application via a Heroku link

• List and comparison


• Heroku deployment

You might also like