Documenting my learnings for Data Science, and how anyone can approach learning ML/AI.
See my qualifications on Linkedin, Personal Website.
I was a part of the Applied AI team in Deloitte USI, with experiences on analysis, cleansing and understanding the story of a dataset to use Machine Learning to build and run custom models, evaluate performances, at times also requiring software development to build relevant dashboards, for various clients.
There are tons of articles, courses (paid and free), degrees available in this field. But, at the end of the day, everything you read should fall into one of these topics.
This article is aimed at teaching those few people who are starting their Data Science from scratch, or are at beginner or intermediate levels.
Beginners can follow through everything provided, and intermediate experienced people can try to identify where all their ML learning really fits in the large canvas of Machine Learning.
Set a weekly target for each of these to atleast have a basic idea of the ML Universe in 5 weeks maximum.
- Setup and get familiar with Jupyter Notebook Usage
- Refer Jupyter Link
This resource by w3 is really great, and will help you build basic python fundamentals and provide syntax-level comfort. You will need to solve the first 10 questions, and subsequently every 2nd/4th question to practice more efficient questions.
- Python Basic (Part -I)
- Python Basic (Part -II)
- Python built-in Modules
- Python Data Types - String
- Python JSON
- Python Data Types - List
- Python Data Types - Dictionary
- Python Conditional statements and loops
- Python functions
- Python Lambda
Follow everything on this Roadmap.
There are literally no shortcuts here, and if you've no idea what to read, just download this book, and start reading the index page to find out topics that you may have left before.
- https://www.analyticsvidhya.com/blog/2021/07/basic-statistics-concepts-for-machine-learning-newbies/
- https://towardsdatascience.com/machine-learning-probability-statistics-f830f8c09326
Although, a strong fundamentals over Maths is required, you may skip this section entirely for now, and revisit this once you will have more experience on the overall ML concepts.
You will need to cover the basic libraries like Pandas, MatPlotLib/Seaborn, Numpy to grasp a basic syntax level understanding of how you can work on dataframe & datasets in large and in Python.
- Cleaning, imputations, merge, joins, iterating over rows, lambda function usage
- Lambda Functions
- Merge, Joins, concatenation of multiple dataframes
- Interview Questions
You will find these few applications on a very large scale, either in projects, and in interviews both.
- NLP (wherever there is text data)
- Churn Modelling (everywhere)
- Customer Segmentation (everywhere)
- Retail
- Sales Forecasting
- UpSell & CrossSell
- Predictive Maintenance
Pick any of these approaches, or mix them up. The idea here is to give you enough understanding of the ML universe, so that you can start picking up things on your own.
- Dataset Specific
(Some famous beginner datasets are given below)
- AirBNB Price Prediction (Regression)
- IMDB Sentiment Analysis (Classification)
- Titanic (Classification)
- Algorithm/Concept Specific
- Classification
- Regression
- Clustering
- Deep Learning
- Decision Trees
- NLP
-
This book will take you on an end-to-end journey of ML, right from a project start to production deployment, and the kind of approaches that you're going need. Read it here - Machine Learning System Design