AAUBS Data Science 2023
Info, Schedule & Co
Modules
Literature & Resources
Semester Schedule
Semester Project Requirements
1.
Applied Data Science and Machine Learning
A) Introduction to Data Science (W35-36)
- Welcome Students!
- Data Handling and Manipulation
- Exploratory Data Analysis and Essential Statistics
- Data Visualization in Data Science
Mathematics (Brushup)
B) Rapid Prototyping (W37)
Introduction to Streamlit: Building an Employee Attrition Dashboard
Streamlit Development & Running Offline
- Real World Data to Online Dashboard
C) Intro to Unsupervised Machine Learning (W38)
- Introduction to Unsupervised ML
- Recommendation and Similarity Search
- Introduction to Clustering: K-means and Hierarchical Approaches
D) Intro to Supervised Machine Learning (W39)
- Introduction to Supervised ML
- SML - Further topics
- Time Series Forecasts
2.
Network Analysis & NLP
Natural Language Processing
Basics of NLP
NLP Applications Chatbot
Network Analysis
Basics Network Analysis
2 Mode Networks
NW Exercises
NW Cases
3.
Applied Deep Learning and Artificial Intelligence
Intro to Traditional Deep Learning
Group Assignment 1
Intro to Transformer Models
Group assignment 2
Intro to GPT Models
Group Assignment 3
Intro to Graph Neural Networks
Group Assignment 4
4.
Data Engineering and Machine Learning Operations in Business
Lecture 1 Introduction to Serverless ML and Databases
Lecture 2 Refactoring & First Serverless App
Lecture 3: Credit Card Prediction Service Project
Lecture 5 Feature Selection, Batch Inference Pipelines, Model Registry
Lecture 6 Feature Selection, Batch Inference Pipelines, Model Registry
Bonus Workshop: Using LLMs in your applications
Clear History
Built with
from
Grav
and
Hugo
Social / Business Data Science 2023
>
Data Engineering and Machine Learning Operations in Business
> Lecture 3: Credit Card Prediction Service Project
Lecture 3: Credit Card Prediction Service Project
Slides
Original Slides from the Serverless ML course
Exercise
Clone repository and execute the feature-pipelines
Modify parameters in the feature-pipeline to produce
