​ ull Data Science Roadmap (Beginner to​

F
​Advanced)​
​ his roadmap is designed to guide you step-by-step through learning data science, from​
T
​fundamentals to advanced projects. Start by following each phase in sequence, mastering​
​the core concepts before moving ahead. For every topic, you'll find carefully selected free​
​resources—mostly hands-on tutorials and practical projects—to help you learn by doing.​
​Take your time to practice, explore different datasets, and build your portfolio as you​
​progress. Later, we will break down each topic further with detailed, free resources and brief​
​explanations to help you choose the best learning path for your needs.​

​Phase 1: Prerequisites & Foundation​

​1. Mathematics for Data Science​

​●​ ​Linear Algebra​

​●​ ​Probability​

​●​ ​Statistics (Descriptive & Inferential)​

​●​ ​Calculus (basic level)​

​2. Programming with Python​

​●​ ​Syntax & Logic​

​●​ ​Functions & Loops​

​●​ ​File Handling​

​●​ ​OOPs Concepts​

NumPy​
​●​ ​Libraries:​​ Pandas​
​,​​

​Phase 2: Data Handling & Analysis​

​3. Data Analysis & Manipulation​

​●​ ​Pandas for DataFrames​

 ​●​ ​NumPy for numerical ops​

​●​ ​Data Cleaning & Wrangling​

​●​ ​Working with Real Datasets (CSV, Excel, JSON)​

​4. Data Visualization​

​●​ ​Matplotlib & Seaborn​

​●​ ​Plotly & Dash (Optional)​

​●​ ​Storytelling with Charts​

​●​ ​Visualizing Correlations and Trends​

​Phase 3: Machine Learning Core​

​5. Supervised Learning​

​●​ ​Linear Regression​

​●​ ​Logistic Regression​

​●​ ​Decision Trees & Random Forest​

​●​ ​SVM, KNN​

​6. Unsupervised Learning​

​●​ ​K-Means Clustering​

​●​ ​Hierarchical Clustering​

​●​ ​PCA (Dimensionality Reduction)​

​7. Model Evaluation & Tuning​

​●​ ​Train/Test Split​

​●​ ​Cross-Validation​
 ​●​ ​Confusion Matrix, Accuracy, Precision/Recall​

​●​ ​Hyperparameter Tuning (GridSearch, RandomSearch)​

​Phase 4: Advanced Concepts​

​8. Feature Engineering & Selection​

​●​ ​Handling Missing Data​

​●​ ​Encoding & Scaling​

​●​ ​Feature Importance​

​●​ ​Domain-Based Features​

​9. Time Series Analysis​

​●​ ​Stationarity​

​●​ ​ARIMA/SARIMA​

​●​ ​Forecasting Models​

​●​ ​Real-world datasets (stock, weather)​

​10. Introduction to Deep Learning​

​●​ ​Basics of Neural Networks​

​●​ ​Perceptrons​

​●​ ​Forward/Backward Propagation​

​●​ ​Keras/TensorFlow Basics​

​Phase 5: Real-World Applications & Deployment​

​11. Project Building & Capstone Projects​

​●​ ​Regression: House Price Prediction​

 ​●​ ​Classification: Titanic Survival, Loan Default​

​●​ ​NLP: Sentiment Analysis​

​●​ ​Clustering: Customer Segmentation​

​12. Model Deployment (MLOps Basics)​

​●​ ​Streamlit/Flask APIs​

​●​ ​Docker Basics​

​●​ ​Deploy to Hugging Face/Render/Heroku​

​●​ ​Git & GitHub for Project Management​

​Phase 6: Domain Integration & Extras​

​13. SQL for Data Science​

​●​ ​SELECT, JOIN, GROUP BY​

​●​ ​Subqueries​

​●​ ​Window Functions​

​●​ ​Hands-on with public datasets​

​14. Cloud Tools for DS (Optional Advanced)​

​●​ ​Google Colab​

​●​ ​BigQuery Basics​

​●​ ​AWS S3, SageMaker intro​

​ ow that you have an overview of the complete data science roadmap, let’s break down​
N
​each phase and concept into more detailed sections.​

​ his will include step-by-step learning paths, free hands-on resources, and brief​
T
​explanations of why each resource is valuable. This detailed breakdown will help you focus​
​on one topic at a time and make your learning journey clear and manageable.​

​Phase 1: Mathematics for Data Science​

​Topics Covered:​

​1.​ ​Linear Algebra​

​2.​ ​Probability​

​3.​ ​Statistics​

​4.​ ​Basic Calculus (optional but helpful)​

​1. Linear Algebra​

​Resource 1:​​Essence of Linear Algebra – 3Blue1Brown​​(YouTube)​

​●​ ​Platform: YouTube – 3Blue1Brown​

​●​ T
​ his visual playlist is​​famous​​for making abstract​​linear algebra intuitive through​
​animations. Covers vectors, matrices, linear transformations, etc.​

​●​ W
​ hy Learn:​​This builds your​​geometric understanding​​of matrix operations, a must​
​for understanding​​machine learning models​​and PCA.​

​Resource 2:​​Linear Algebra for Data Science – freeCodeCamp​​(YouTube)​

​●​ ​Platform: YouTube – freeCodeCamp​

​●​ A
​ long-form practical course for data science learners. Focuses on matrix math with​
​Python, NumPy operations, and real use cases.​

​●​ W
​ hy Learn:​​Great for​​hands-on learners​​who want to​​mix math with coding from​
​Day 1.​

​Resource 3:​​Khan Academy – Linear Algebra​

 ​●​ ​Platform: Khan Academy​

​●​ S
​ tructured lessons with quizzes, examples, and visual explanations. Very​
​beginner-friendly, focused on theory with occasional visuals.​

​●​ W
​ hy Learn:​​Best for those who prefer​​step-by-step​​foundations​​, and it's easy to​
​revisit topics at your own pace.​

​2. Probability​
​Resource 1:​​Probability Fundamentals – Khan Acadamy​​(YouTube)​

​●​ ​Platform: YouTube – Khan Acadamy​

​●​ S
​ imple explanations with cartoons. It breaks down difficult probability terms like​
​independence, Bayes’ Theorem, distributions, etc.​

​●​ W
​ hy Learn:​​Extremely clear if you're scared of math.​​Great to​​build confidence in​
​probability thinking​​.​

​Resource 2:​​Introduction to Probability – Harvard​​(EdX Audit)​

​●​ ​Platform: HarvardX (Free via self-study site)​

​●​ A
​ rigorous but approachable course built by Harvard for undergrads, now made free.​
​Includes PDFs, lecture notes, and problem sets.​

​●​ W
​ hy Learn:​​Industry-level depth. Perfect if you want​​to go​​beyond basic intuition​
​and into theory​​.​

​Resource 3:​​Khan Academy – Probability & Combinatorics​

​●​ ​Platform: Khan Academy​

​●​ B
​ ite-sized videos + hands-on quizzes. Explains coin flips, dice, conditional​
​probability, and combinations.​

​●​ ​Why Learn:​​Ideal if you're starting from scratch and​​want a​​structured, gentle path​​.​

​3. Statistics (Descriptive + Inferential)​

​Resource 1:​​CrashCourse – Statistics (YouTube)​
 ​●​ ​Platform: YouTube – CrashCourse​

​●​ F
​ ast-paced, visual and very engaging! Covers mean, median, variance, distribution​
​types, z-scores, etc.​

​●​ W
​ hy Learn:​​Great for​​quick foundational review​​and​​building familiarity with​
​statistical language.​

​Resource 2:​​Statistics for Data Science – freeCodeCamp​​(YouTube)​

​●​ ​Platform: YouTube – freeCodeCamp​

​●​ A
​ full 3-hour stats course specifically geared toward​​real-world data science​
​applications​​. Includes practical implementation with​​Python.​

​●​ W
​ hy Learn:​​Combines theory + coding, ideal for project-based​​learners who want​
​immediate application.​

​4. Basic Calculus (Optional but Useful)​

​Resource 1:​​Calculus for Machine Learning – YouTube​

​●​ ​Platform: YouTube – Jon Krohn​

​●​ ​Shows how derivatives and gradients apply in neural networks and optimization.​

​●​ W
​ hy Learn:​​If you plan to learn​​deep learning​​, this​​will help you understand how​
​models​​actually learn​​.​

​Resource 2:​​Khan Academy – Differential Calculus​

​●​ ​Platform: Khan Academy​

​●​ A
​ full series starting from “What is a derivative?” to applications like slope, rate of​
​change, etc.​

​●​ W
​ hy Learn: Best for​​zero background​​students — self-paced​​with visuals and​
​quizzes.​
​ hase 2: Programming with Python for​
P
​Data Science​
​Topics Covered:​

​1.​ ​Core Python Fundamentals​

​2.​ ​Python for Data Handling (NumPy & Pandas)​

​3.​ ​Hands-on Coding Practice​

​1. Core Python Fundamentals​

​Resource 1:​​Python for Beginners – Programming with​​Mosh (YouTube)​

​●​ ​Platform: YouTube – Programming with Mosh​

​●​ ​Covers variables, loops, functions, error handling, and OOP—all beginner-friendly.​

​●​ W
​ hy Learn: Mosh is known for​​clear, no-fluff teaching​​— this video (6+ hours) gives​
​a​​solid Python base in one sitting​​.​

​Resource 2:​​Python Basics – freeCodeCamp (YouTube)​

​●​ ​Platform: YouTube – freeCodeCamp​

​●​ S
​ tarts from zero: covers syntax, loops, functions, OOP, etc., and includes small​
​exercises throughout.​

​●​ W
​ hy Learn: Designed for​​complete beginners​​, and the​​examples are great for​​logic​
​building​​.​

​Resource 3:​​Python Practice – W3Schools (Interactive)​

​●​ ​Platform: W3Schools​

​●​ C
​ lick-to-run Python code with built-in exercises for each topic (functions, loops,​
​strings, etc.)​

​●​ W
​ hy Learn: Useful for​​hands-on practice instantly​​without needing to install​
​anything.​
​2. Python for Data Handling (NumPy & Pandas)​
​Resource 1:​​NumPy Tutorial – freeCodeCamp (YouTube)​

​●​ ​Platform: YouTube – freeCodeCamp​

​●​ A
​ full practical tutorial teaching array creation, indexing, broadcasting, and​
​mathematical ops using NumPy.​

​●​ W
​ hy Learn: NumPy is the​​foundation for all numerical​​computing in ML/AI​​— this​
​makes you job-ready in handling arrays.​

​Resource 2:​​Pandas Tutorial – Data School (YouTube​​Playlist)​

​●​ ​Platform: YouTube – Data School​

​●​ S
​ hort videos explaining concepts like Series, DataFrames, filtering, grouping,​
​merging, etc.​

​●​ W
​ hy Learn: Covers​​real-world use cases​​in short segments—perfect​​for revision​
​and project prep.​

​Resource 3:​​Pandas Official Tutorials – [Link]​

​●​ ​Platform: Pandas Official Site​

​●​ A
​ uthoritative documentation with code examples, ideal for deep diving into any​
​feature.​

​●​ W
​ hy Learn: Knowing how to​​read official docs​​is a​​long-term skill for any serious​
​coder/data scientist.​

​3. Hands-on Coding Practice​

​Resource 1:​​HackerRank – Python Challenges​

​●​ ​Platform: HackerRank​

​●​ 1
​ 0 Days of Python Practice – great for data science beginners to sharpen their​
​problem-solving.​

​●​ W
​ hy Learn: Ideal for​​developing logic-building and​​syntax recall​​—vital in​
​interviews.​
​Resource 2:​​Kaggle – Python Course (Free)​

​●​ ​Platform: Kaggle​

​●​ I​nteractive notebooks focused on writing and running Python code in the browser.​
​Focused on​​data-centric tasks​​.​

​●​ W
​ hy Learn: It’s​​hands-on, beginner-friendly​​, and​​directly​​aligned to data science​
​practice.​

​Resource 3:​​Exercism – Python Track​

​●​ ​Platform: Exercism​

​●​ O
​ ffers mentor-supported Python problems. Step-by-step and challenges that go​
​deeper than syntax.​

​●​ W
​ hy Learn: Encourages​​code quality and thought process​​— useful once you’ve​
​mastered basics.​

​Phase 3: Data Analysis & Visualization​

​Topics Covered:​

​1.​ ​Data Cleaning & Exploration with Pandas​

​2.​ ​Data Visualization with Matplotlib, Seaborn & Plotly​

​3.​ ​Working with Real Datasets (CSV, Excel, APIs)​

​1. Data Cleaning & Exploration with Pandas​

​Resource 1:​​Exploratory Data Analysis – Kaggle (Free)​

​●​ ​Platform: Kaggle​

​●​ H
​ ands-on notebook lessons covering missing data, string handling, type conversion,​
​and common EDA workflows.​
 ​●​ W
​ hy Learn: Helps you​​think like a data analyst​​— not just clean data, but​
​understand what's wrong with it.​

​Resource 2:​​Data Analysis with Python – freeCodeCamp​​(YouTube)​

​●​ ​Platform: YouTube – freeCodeCamp​

​●​ C
​ overs reading datasets, grouping, filtering, applying functions, and summarizing​
​data. Full project-based teaching.​

​●​ W
​ hy Learn: This​​mimics a real analysis scenario​​, good​​for beginners starting their​
​first mini-projects.​

​Resource 3:​​Real-Time EDA on Titanic Dataset – Krish​​Naik (YouTube)​

​●​ ​Platform: YouTube – Krish Naik​

​●​ S
​ tep-by-step analysis on the famous Titanic dataset. Shows how to draw meaning​
​from columns and relations.​

​●​ W
​ hy Learn: Gives you​​real project experience​​and teaches​​how to build a​
​storytelling mindset.​

​2. Data Visualization (Matplotlib, Seaborn, Plotly)​

​ esource 1:​​Matplotlib & Seaborn Crash Course – freeCodeCamp​
R
​(YouTube)​

​●​ ​Platform: YouTube – freeCodeCamp​

​●​ F
​ ull 2-hour course showing how to plot bar charts, scatter plots, heatmaps,​
​histograms with real datasets.​

​●​ W
​ hy Learn: Helps you​​build dashboards and reports​​— a vital part of analyst and​
​DS roles.​

​ esource 2:​​Interactive Plots with Plotly – Harry’s​​data Journey​

R
​(YouTube)​

​●​ ​Platform: YouTube – Harry’s Data Journey​

 ​●​ I​ntroduces Plotly to create interactive plots, charts, and dashboards directly from​
​notebooks.​

​●​ W
​ hy Learn: Adds​​interactivity to your portfolio​​, especially​​useful in apps or​
​dashboards.​

​3. Working with Real Datasets​

​Resource 1:​​Kaggle Datasets – Start Exploring​

​●​ ​Platform: Kaggle​

​●​ A
​ library of 50,000+ real-world datasets across finance, health, education, games,​
​etc.​

​●​ W
​ hy Learn: Practicing with real data builds​​project-ready​​confidence​​and​
​problem-solving skills.​

​ esource 2:​​How to Import Data (CSV, Excel, APIs)​​– CodeBasics​

R
​(YouTube)​

​●​ ​Platform: YouTube – CodeBasics​

​●​ S
​ hows how to read CSVs, connect to APIs, extract Excel files, and clean data​
​automatically.​

​●​ W
​ hy Learn: Understanding​​multiple data sources​​prepares​​you for enterprise-level​
​projects.​

​Resource 3:​​Real-World Projects – YouTube (Codebasics)​

​●​ ​Platform: YouTube – CodeBasics​

​●​ H
​ ands-on projects like IPL data analysis, YouTube revenue prediction, and customer​
​segmentation.​

​●​ W
​ hy Learn: Helps you​​practice storytelling and presentation​​,​​which are core in​
​real interviews.​

​Phase 4: Machine Learning (ML)​

​ his is where your data starts turning into​​predictions​​and​​insights that drive action​​. You’ll​
T
​explore how machines learn patterns, make decisions, and how you can build models that​
​generalize well to new data.​

​Topics Covered:​
​1.​ ​Introduction to Machine Learning​

​2.​ ​Supervised Learning Algorithms​

​3.​ ​Unsupervised Learning Algorithms​

​4.​ ​Model Evaluation & Tuning​

​5.​ ​ML Projects (Hands-on)​

​1. Introduction to Machine Learning​

​Resource 1:​​Machine Learning for Beginners – freeCodeCamp​​(YouTube)​

​●​ ​Platform: YouTube – freeCodeCamp​

​●​ A
​ 4-hour beginner course by Simplilearn that starts from scratch and covers core ML​
​concepts.​

​●​ W
​ hy Learn: You’ll understand​​what ML is​​, types of​​learning​
​(supervised/unsupervised), model building pipeline, and basic terms like underfitting,​
​overfitting, bias, and variance.​

​Resource 2:​​Kaggle: Intro to Machine Learning (Free)​

​●​ ​Platform: Kaggle​

​●​ C
​ ode-along mini-course teaching how to predict house prices using Decision Trees​
​and Random Forests.​

​●​ W
​ hy Learn: Helps​​apply ML hands-on​​even if you’re​​new. Easy-to-understand​
​concepts + real dataset usage.​

​Resource 3:​​ML Basics – StatQuest (YouTube Playlist)​

​●​ ​Platform: YouTube – StatQuest​

 ​●​ S
​ hort, animated explainers that break down ML math intuitively (e.g., linear​
​regression, trees).​

​●​ W
​ hy Learn: StatQuest is​​gold for conceptual clarity​​with fun, cartoon-style​
​teaching.​

​2. Supervised Learning Algorithms​

​Resource 1:​​Supervised ML with Scikit-learn – freeCodeCamp​​(YouTube)​

​●​ ​Platform: YouTube – freeCodeCamp​

​●​ T
​ eaches classification, regression using Python’s Scikit-learn with projects like Iris​
​classification.​

​●​ W
​ hy Learn: Shows how to write ML pipelines with real data, train/test split, and​
​model interpretation.​

​ esource 2:​​Kaggle: Intermediate ML (Handling Missing​​Values,​

R
​Categorical Data)​

​●​ ​Platform: Kaggle​

​●​ E
​ xpands on ML intro by dealing with real-world mess: missing values, categorical​
​features, leakage.​

​●​ ​Why Learn: Takes you one level deeper to​​write professional-grade​​ML code​​.​

​Resource 3:​​ML Algorithms – Krish Naik Playlist (YouTube)​

​●​ ​Platform: YouTube – Krish Naik​

​●​ P
​ ractical explanation of Linear Regression, Logistic Regression, SVM, KNN, etc.,​
​with full implementation.​

​●​ W
​ hy Learn: Builds​​confidence in coding ML algorithms​​end-to-end, not just​
​importing them.​

​3. Unsupervised Learning Algorithms​

​Resource 1:​​Unsupervised ML – freeCodeCamp (YouTube)​
 ​●​ ​Platform: YouTube – freeCodeCamp​

​●​ ​Explains Clustering, K-Means, PCA, and Dimensionality Reduction with Python.​

​●​ W
​ hy Learn: Essential for​​segmentation tasks, recommendation​​systems​​, and​
​reducing large datasets.​

​Resource 2:​​KMeans Clustering – StatQuest (YouTube)​

​●​ ​Platform: YouTube – StatQuest​

​●​ ​Animated breakdown of how KMeans algorithm works.​

​●​ W
​ hy Learn: Understand how unsupervised learning​​really​​works​​without jumping​
​into code first.​

​4. Model Evaluation & Tuning​

​Resource 1:​​Model Evaluation Metrics – StatQuest (YouTube)​

​●​ ​Platform: YouTube – StatQuest​

​●​ ​Covers accuracy, precision, recall, confusion matrix, ROC curve, etc.​

​●​ W
​ hy Learn: These metrics are​​asked in every interview​​— crucial to justify your​
​model’s performance.​

​Resource 2:​​GridSearchCV​

​CorssValidation​

​●​ ​Platform: YouTube – 2 channels codebasics and stackQuest​

​●​ S
​ tep-by-step guide to hyperparameter tuning using GridSearchCV and​
​RandomizedSearchCV.​

​●​ W
​ hy Learn: Tuning models​​improves performance drastically​​,​​and this video​
​teaches it cleanly.​

​5. ML Projects (Hands-On)​

​Resource 1:​​Titanic ML Project – Kaggle Competition​
 ​●​ ​Platform: Kaggle​

​●​ ​Classic beginner-friendly classification project with starter notebooks and forums.​

​●​ W
​ hy Learn: First real ML project that teaches​​data​​prep, model training, and​
​evaluation.​

​Resource 2:​​End-to-End Projects – CodeBasics (YouTube)​

​●​ ​Platform: YouTube – CodeBasics​

​●​ ​Real-world projects (house price prediction, loan default prediction, HR analytics).​

​●​ W
​ hy Learn: Great for building​​portfolio-worthy case​​studies​​using structured​
​industry problems.​

​Phase 4.5: Deep Learning​

​ hat You’ll Learn:​
W
​Master the art of building models that can learn from​​images, text, and sequences​​. Deep​
​learning enables everything from​​image recognition​​to​​language translation​​and​
​chatbots​​.​

​1. Neural Network Fundamentals​

​ esource 1:​​Neural Networks & Deep Learning – 3Blue1Brown​

R
​Platform:​​YouTube​
​Visually intuitive series explaining how neurons, weights, and backpropagation work using​
​animations.​
​Why Learn:​​Makes abstract math like gradients and​​layers very simple with brilliant​
​animations. Great starting point.​

​ esource 2:​​Deep Learning Specialization – Andrew​​Ng (Audit Free)​

R
​Platform:​​Coursera​
​One of the most trusted beginner series covering NN architectures, loss functions, and​
​forward/backward pass.​
​Why Learn:​​Combines foundational math with hands-on​​coding exercises. Industry-relevant​
​content.​

​ esource 3:​​Neural Networks from Scratch – Sentdex​

R
​Platform:​​YouTube​
​Build a neural net from scratch in Python & NumPy — no libraries, just logic.​
​Why Learn:​​If you love learning the hard way, this​​gives true clarity on what’s happening​
​under the hood.​

​Practice Idea:​

​●​ ​Build a neural net from scratch to solve XOR problem or digit classification​

​●​ ​Visualize activation values layer by layer​

​2. Deep Learning Frameworks (TensorFlow & PyTorch)​

​ esource 1:​​TensorFlow in 100 Minutes – freeCodeCamp​

R
​Platform:​​YouTube​
​Crash course on TensorFlow & Keras with a small image classifier project.​
​Why Learn:​​Best intro if you want to go from zero​​to working model quickly.​

​ esource 2:​​PyTorch Tutorial for Beginners – Aladdin​​Persson​

R
​Platform:​​YouTube​
​Learn tensors, gradients, and building networks with PyTorch.​
​Why Learn:​​PyTorch is widely used in research. This​​helps you switch between frameworks​
​easily.​

​ esource 3:​​Official Keras Documentation​

R
​Platform:​​[Link]​
​Tutorials for building models — regression, classification, image, etc.​
​Why Learn:​​Clean examples and official guidance for​​beginners using Keras.​

​Practice Idea:​

​●​ ​Build a regression model in Keras (predict house prices)​

​●​ ​Rebuild the same in PyTorch​

​3. Convolutional Neural Networks (CNNs)​

​ esource 1:​​CNNs Explained – StatQuest​

R
​Platform:​​YouTube​
​Simple animated explanation of how filters, kernels, pooling layers work.​
​Why Learn:​​Turns complex CNN concepts into fun stories​​with visuals.​

​ esource 2:​​Deep Learning for Computer Vision – freeCodeCamp​

R
​Platform:​​YouTube​
​Build CNNs using TensorFlow on real datasets like MNIST.​
​Why Learn:​​Combines code + theory in one long-form tutorial.​

​ esource 3:​​Intro to Deep Learning (CNNs) – Kaggle​

R
​Platform:​​Kaggle​
​Interactive notebook with CNN hands-on practice on MNIST.​
​Why Learn:​​Directly practice in-browser, no setup​​needed.​

​Practice Idea:​

​●​ ​Build a CNN to classify fashion images (Fashion MNIST)​

​●​ ​Try visualizing feature maps and learned filters​

​4. RNNs & LSTMs (For Sequences and Text)​

​ esource 1:​​RNNs and LSTMs – StatQuest​

R
​Platform:​​YouTube​
​Explains how RNNs and LSTMs work with sequences.​
​Why Learn:​​Makes time-based learning (stocks, text)​​easy to understand.​

​ esource 2:​​NLP with Deep Learning – freeCodeCamp​

R
​Platform:​​YouTube​
​Sentiment analysis using LSTMs in Keras.​
​Why Learn:​​Combines NLP theory with model training​​and evaluation.​

​ esource 3:​​PyTorch NLP Tutorials​

R
​Platform:​​[Link]​
​Step-by-step text generation and classification using PyTorch.​
​Why Learn:​​Use real PyTorch code for NLP pipelines.​

​Practice Idea:​

​●​ ​Predict next word or generate text (e.g., Shakespeare-style text)​

​●​ ​Train an LSTM model to forecast stock or temperature​

​5. Transfer Learning​

​ esource 1:​​Hugging Face Transformers Crash Course​

R
​Platform:​​YouTube​
​Intro to transformer-based NLP models like BERT.​
​Why Learn:​​Hugging Face is the standard in NLP — must-have for language projects.​

​ esource 2:​​Kaggle Transfer Learning – Cats vs Dogs​

R
​Platform:​​Kaggle​
​Fine-tune a pre-trained CNN to classify cat vs dog images.​
​Why Learn:​​Real-world practice with minimal data​​using pre-trained models.​

​Practice Idea:​

​●​ ​Fine-tune ResNet for custom flower image classification​

​●​ ​Use BERT for text classification (positive/negative reviews)​

​6. Hands-on Deep Learning Projects​

​ esource 1:​​CNN Image Classification – Kaggle​

R
​Platform:​​Kaggle​
​Step-by-step project using CNNs to classify dog vs cat images.​
​Why Learn:​​Combines preprocessing, model building,​​training, and submission.​

​ esource 2:​​Sentiment Analysis with LSTM – Towards​​Data Science​

R
​Platform:​​Medium​
​Uses LSTM model to analyze text sentiment.​
​Why Learn:​​Great starter NLP project with deep learning​​integration.​

​ esource 3:​​Time Series Forecasting with LSTMs – Krish​​Naik​

R
​Platform:​​YouTube​
​Forecast stock prices or similar time series using deep learning.​
​Why Learn:​​Real-world sequence modeling in a simple,​​clear format.​
​Search:​​“Krish Naik LSTM Time Series”​

​Practice Idea:​

​●​ ​Build a Streamlit app to classify X-rays or flowers​

​●​ ​Deploy a Hugging Face transformer for tweet sentiment​

​●​ ​Forecast electricity usage with LSTMs​

​ hase 5: SQL & Databases for Data​
P
​Science​
​ QL is​​non-negotiable​​for data scientists. Whether​​you're pulling customer data, joining​
S
​transactional logs, or filtering millions of rows — SQL is the language of data access. It’s a​
​must-have for both interviews and real jobs.​

​Topics Covered:​
​1.​ ​SQL Basics: SELECT, WHERE, ORDER BY​

​2.​ ​Aggregations: GROUP BY, COUNT, SUM, AVG​

​3.​ ​Joins: INNER, LEFT, RIGHT, FULL​

​4.​ ​Subqueries, Window Functions, CTEs​

​5.​ ​Hands-on SQL Projects​

​1. SQL Basics​

​Resource 1:​​SQL for Data Science – freeCodeCamp (YouTube)​

​●​ ​Platform: YouTube – freeCodeCamp​

​●​ ​Covers SQL syntax, filtering, sorting, aliasing, and basic aggregation using MySQL.​

​●​ W
​ hy Learn: This is your​​go-to crash course​​, taught​​like a bootcamp from zero to​
​real-world usage.​

​Resource 2:​​SQL Bolt – Interactive Lessons​

​●​ ​Platform: SQLBolt​

​●​ C
​ lick-and-run interactive tutorials — from basic SELECTs to JOINs — with​
​explanations and practice.​

​●​ W
​ hy Learn: Great for​​absolute beginners​​who want to​​learn by doing in-browser​
​without setup.​

​Resource 3:​​Mode SQL Tutorial – Business Focused​

 ​●​ ​Platform: Mode Analytics​

​●​ L
​ essons on querying real business datasets. Explains not just syntax, but​​why​​you​
​write a query.​

​●​ W
​ hy Learn: Helps you​​think like an analyst​​, writing​​SQL for product, business, or​
​growth teams.​

​2. Aggregations & Filtering​

​Resource 1:​​LeetCode SQL Easy–Medium Questions​

​●​ ​Platform: LeetCode​

​●​ H
​ ands-on problems that test your use of COUNT, GROUP BY, and filtering with​
​real-world context.​

​●​ W
​ hy Learn: Good for​​interview-style questions​​and​​building muscle memory in​
​query writing.​

​Resource 2:​​StrataScratch SQL Practice​

​●​ ​Platform: StrataScratch​

​●​ R
​ eal company datasets (Google, Amazon, Uber) and tasks like top customers, avg​
​spend, churn prediction.​

​●​ W
​ hy Learn: You learn​​how SQL is used in the wild​​—​​perfect for preparing for job​
​assessments.​

​3. Joins, Subqueries & Window Functions​

​Resource 1:​​Joins Explained Visually – Khan Academy​

​●​ ​Platform: Khan Academy​

​●​ ​Visual diagrams explaining INNER, LEFT, RIGHT, FULL joins and how they work.​

​●​ ​Why Learn: Makes​​join logic crystal clear​​, especially​​for visual learners.​

​Resource 2:​​Window Functions & CTEs – Mode SQL Advanced​

 ​●​ ​Platform: Mode Analytics​

​●​ L
​ earn how to use advanced SQL to rank users, calculate moving averages, and​
​segment customers.​

​●​ W
​ hy Learn: These tools are​​used in advanced analytics​​and product metrics —​
​boosts your skillset fast.​

​4. Hands-on SQL Projects​

​Resource 1:​​Kaggle: SQL Datasets (Google BigQuery)​

​●​ ​Platform: Kaggle​

​●​ P
​ ractice SQL queries on datasets like Stack Overflow, US Census, and Google​
​Analytics using BigQuery.​

​●​ ​Why Learn: Write​​live queries​​on massive real datasets​​for portfolio building.​

​Resource 2:​​8 SQL Case Studies – DataLemur​

​●​ ​Platform: DataLemur​

​●​ R
​ eal-world company questions turned into SQL challenges: average order value,​
​active user count, etc.​

​●​ ​Why Learn: Excellent​​mock interview practice​​with​​solutions and walkthroughs.​

​ hase 6: Statistics & Probability for Data​

P
​Science​
​ tatistics is the​​foundation​​of Data Science. Without​​it, you're just plotting numbers. It helps​
S
​you​​analyze data​​,​​validate models​​, and​​draw actionable​​conclusions​​from experiments​
​or trends.​

​Topics Covered:​
​1.​ ​Descriptive Statistics​

​2.​ ​Probability Basics​

 ​3.​ ​Bayes’ Theorem & Conditional Probability​

​4.​ ​Distributions (Normal, Binomial, etc.)​

​5.​ ​Hypothesis Testing​

​6.​ ​Hands-on Stats Projects​

​1. Descriptive Statistics​

​Resource 1:​​Statistics for Data Science – freeCodeCamp​​(YouTube)​

​●​ ​Platform: YouTube – freeCodeCamp​

​●​ F
​ ull course on mean, median, mode, variance, standard deviation, data spread, and​
​distributions.​

​●​ W
​ hy Learn: Covers​​real examples​​to explain concepts​​in the context of data science​
​and analytics.​

​Resource 2:​​Descriptive Stats – Khan Academy​

​●​ ​Platform: Khan Academy​

​●​ ​Bite-sized videos and quizzes on central tendency, box plots, and histograms.​

​●​ ​Why Learn:​​Super beginner-friendly​​with animations​​and interactive practice.​

​Resource 3:​​Basic Stats Using Python – W3Schools​

​●​ ​Platform: W3Schools​

​●​ ​Teaches mean, median, mode, and std dev using Python step-by-step.​

​●​ W
​ hy Learn: Bridges stats theory with​​Python code instantly​​— great for applied​
​learners.​

​2. Probability & Distributions​

​Resource 1:​​Probability Basics – 3Blue1Brown (YouTube)​

​●​ ​Platform: YouTube – 3Blue1Brown​

 ​●​ ​Visual explanation of probability, combinations, permutations, and random variables.​

​●​ W
​ hy Learn: Stunning​​intuitive animations​​for understanding​​“why probability​
​works.”​

​Resource 2:​​StatQuest Probability + Distributions​

​●​ ​Platform: YouTube – Zedstaticstics​

​●​ ​Covers normal, binomial, Poisson, and uniform distributions clearly and with fun.​

​●​ W
​ hy Learn: Josh makes​​scary concepts fun and friendly​​.​​Perfect for foundational​
​clarity.​

​Resource 3:​​Probability in Python – PythonLikeYouMeanIt​

​●​ ​Platform: Data camp​

NumPy​
​●​ ​Shows how to simulate and visualize probabilities in Python using​​ ​.​

​●​ W
​ hy Learn: Teaches​​how to code probabilities​​— necessary​​for simulations, ML,​
​and stats modeling.​

​3. Hypothesis Testing & Confidence Intervals​

​Resource 1:​​Hypothesis Testing – Codebasics (YouTube)​

​●​ ​Platform: YouTube – Codebasics​

​●​ ​Explains p-value, z-test, t-test using Excel and Python.​

​●​ ​Why Learn: It’s​​project-based and practical​​, with​​relatable business examples.​

​Resource 2:​​Hypothesis Testing – StatQuest​

​●​ ​Platform: YouTube – StatQuest​

​●​ ​Crisp and cartoon-style breakdown of null/alternate hypotheses, Type I/II errors.​

​●​ ​Why Learn: Makes statistical testing​​easy to visualize​​and remember.​

​Resource 3:​​T-Tests & Confidence Intervals – Medium(Blog)​

 ​●​ ​Platform: Medium​

​●​ ​Text-based, clean guide to hypothesis testing with Python examples.​

​●​ ​Why Learn: Helps understand how​​real data scientists​​apply testing​​in notebooks.​

​4. Hands-on Statistics Practice​

​Resource 1:​​Kaggle: Statistics Course​

​●​ ​Platform: Kaggle​

​●​ ​Teaches descriptive stats, sampling, distributions, and statistical testing.​

​●​ ​Why Learn: Code + concept combo with​​real datasets​​and in-browser exercises.​

​Resource 2:​​Mini Statistics Projects – DataCamp (Free​​Chapters)​

​●​ ​Platform: DataCamp (first chapters free)​

​●​ ​Uses Python to explore datasets with confidence intervals and histograms.​

​●​ ​Why Learn: Project-oriented and​​Pythonic application​​of stats.​

​Phase 7: Data Visualization​

​ nce you've gathered insights from your data,​​how​​you present it matters​​. Data​
O
​visualization helps you​​tell stories​​,​​explain trends​​,​​and​​support decisions​​. Good​
​visualizations can speak louder than tables of numbers.​

​Topics Covered:​
​1.​ ​Matplotlib – The Foundation​

​2.​ ​Seaborn – Advanced Statistical Plots​

​3.​ ​Plotly – Interactive Dashboards​

​4.​ ​Tableau & Power BI (Free Tools)​

 ​5.​ ​Data Storytelling & Real-World Dashboards​

​1. Matplotlib – The Foundation of Python Visualizations​

​Resource 1:​​Matplotlib Crash Course – Corey Schafer​​(YouTube)​

​●​ ​Platform: YouTube – Corey Schafer​

​●​ ​Teaches the fundamentals: line plots, bar charts, labels, legends, saving figures.​

​●​ W
​ hy Learn: Matplotlib is the​​core library​​for all​​Python plots — essential for any​
​visualization work.​

​Resource 2:​​W3Schools Matplotlib​

​●​ ​Platform: W3Schools​

​●​ ​Simple and clear tutorial with live coding blocks.​

​●​ ​Why Learn: Great for​​hands-on trial and error​​without​​setting up Jupyter.​

​Resource 3:​​Official Matplotlib Tutorials​

​●​ ​Platform: [Link]​

​●​ ​Detailed documentation with examples and use cases.​

​●​ ​Why Learn: Go here when you're ready to​​customize​​and master​​Matplotlib deeply.​

​2. Seaborn – Statistical Plots Made Easy​

​Resource 1:​​Seaborn Tutorial – KGP talkie (YouTube)​

​●​ ​Platform: YouTube – KGP talkie​

​●​ C
​ overs line plots, scatter plots, boxplots, histograms, heatmaps — all with real​
​datasets.​

​●​ W
​ hy Learn: Seaborn builds on Matplotlib, making​​beautiful​​plots with minimal​
​code​​.​

​Resource 2:​​Seaborn Documentation + Gallery​

 ​●​ ​Platform: Seaborn​

​●​ ​Visual gallery with code for each plot type.​

​●​ W
​ hy Learn: Perfect for​​inspiration and copying working​​code​​when building​
​dashboards.​

​Resource 3:​​Kaggle: Data Visualization Course​

​●​ ​Platform: Kaggle​

​●​ ​Covers Seaborn + pandas viz with hands-on code challenges.​

​●​ W
​ hy Learn: You can​​code and submit instantly​​inside​​the browser — no setup​
​needed.​

​3. Plotly – For Interactive Visualizations​

​Resource 1:​​Plotly Crash Course – Tech with Tim (YouTube)​

​●​ ​Platform: YouTube – Tech With Tim​

​●​ ​Shows bar plots, line plots, interactive tooltips, and web-based dashboards.​

​●​ ​Why Learn: Plotly lets you create​​interactive, zoomable​​charts​​easily with Python.​

​Resource 2:​​Official Plotly Docs (Python)​

​●​ ​Platform: Plotly​

​●​ ​Step-by-step tutorials for each chart: area, scatter, 3D, map plots, etc.​

​●​ ​Why Learn: Master Plotly​​directly from the source​​with working notebooks.​

​Resource 3:​​Build Dashboards with Plotly Dash – YouTube​

​●​ ​Platform: YouTube – Charming Data​

​●​ ​Full tutorial on using Dash to build dashboards and web apps using Plotly charts.​

​●​ ​Why Learn: Lets you​​build live data dashboards​​for​​your resume and projects.​
​4. Power BI & Tableau (Free Options)​
​Resource 1:​​Power BI Full Course – Avi Singh powerBIPro​​(YouTube)​

​●​ ​Platform: YouTube – AVI singh PowerBIPor​

​●​ ​Learn Power BI basics, DAX, data modeling, and visualizations from scratch.​

​●​ W
​ hy Learn: Power BI is a​​top business intelligence​​tool​​used by companies across​
​industries.​

​Resource 2:​​Tableau Public + Full Tutorial – Tableau​​Tim (YouTube)​

​●​ ​Platform: YouTube – Tableau Tim​

​●​ L
​ earn how to use Tableau Public (free version) to create charts, dashboards, and​
​storyboards.​

​●​ W
​ hy Learn: Tableau is great for​​drag-and-drop storytelling​​,​​useful for non-coders​
​too.​

​Resource 3:​​Google Data Studio (Looker Studio)​​+​​Tutorial​

​●​ ​Platform: Google​

​●​ ​Build live dashboards with Google Sheets or BigQuery for free.​

​●​ W
​ hy Learn:​​Cloud-powered dashboards​​with zero code​​— ideal for portfolio work​
​and reporting.​

​5. Hands-on Visualization Projects​

​Resource 1:​​Kaggle Visualization Projects​

​●​ ​Platform: Kaggle​

​●​ ​Explore 1,000s of public notebooks analyzing Titanic, COVID-19, sales data, etc.​

​●​ W
​ hy Learn: Hands-on projects help you​​learn by doing​​,​​and you can fork others’​
​work.​

​Resource 2:​​Makeover Monday Dataset + Tableau Challenges​

 ​●​ ​Platform: Makeover Monday​

​●​ ​Weekly dataset challenges with a dashboard design theme.​

​●​ W
​ hy Learn: Improve​​data storytelling skills​​and post​​weekly on LinkedIn for​
​visibility.​

​Resource 3:​​Power BI Sample Dashboards – Microsoft​

​●​ ​Platform: Microsoft Docs​

​●​ ​Downloadable datasets for retail, finance, HR, and marketing to practice dashboards.​

​●​ ​Why Learn: Practice​​real-world scenarios​​with enterprise-ready​​datasets.​

​ hase 8: End-to-End Data Science​

P
​Projects (Mini to Major)​
​ ow it’s time to​​apply everything you’ve learned​​—​​data cleaning, exploration, statistics,​
N
​visualization, and Python — into real-world projects.​

​Doing projects helps you:​

​●​ ​Build a​​portfolio​​for your resume or GitHub.​

​●​ ​Learn how data flows from​​raw to insights​​.​

​●​ ​Prepare for interviews (most DS interviews ask for project explanations).​

​What You’ll Do in This Phase:​

​1.​ ​Mini Projects (Basic to Intermediate)​

​2.​ ​Major Projects (End-to-End with Models)​

​3.​ ​Kaggle Competitions / Hackathons​

​4.​ ​Portfolio Building & Documentation​

​1. Mini Projects (Perfect for Beginners)​

​Resource 1:​​Kaggle Titanic Survival Prediction​

​●​ ​Platform: Kaggle​

​●​ C
​ lassic dataset where you clean data, handle missing values, and build a basic ML​
​model.​

​●​ W
​ hy Learn: It’s the​​“hello world”​​of data science​​— great intro to structured​
​projects.​

​Resource 2:​​Data Cleaning & EDA with Netflix Dataset​​– GitHub​

​●​ ​Platform: GitHub​

​●​ ​Use pandas, seaborn, and matplotlib to clean and visualize Netflix content data.​

​●​ ​Why Learn: Teaches real-life​​exploratory data analysis​​with plots and insights.​

​2. Major Projects (Intermediate to Advanced)​

​Resource 1:​​Retail Sales Forecasting – GitHub Project​

​●​ ​Platform: YouTube – GitHub​

​●​ ​Full series with Excel → Python → ML → dashboard reporting.​

​●​ ​Why Learn: Gives a​​realistic end-to-end business project​​feel​​with KPIs.​

​Resource 2:​​HR Analytics & Attrition Prediction –​​GitHub Project​

​●​ ​Platform: GitHub​

​●​ ​Use classification models to predict which employees might leave.​

​●​ ​Why Learn: It covers​​feature engineering + classification​​on real HR data.​

​Resource 3:​​Zomato Data Analysis & Visualization –​​YouTube​

​●​ ​Platform: YouTube – The iScale​

 ​●​ C
​ omplete analysis + dashboard on Zomato restaurant data using pandas &​
​matplotlib.​

​●​ W
​ hy Learn: You learn​​data storytelling + dashboard​​building​​on real urban​
​dataset.​

​3. Kaggle Competitions + Datasets​

​Resource 1:​​Kaggle Competitions Page​

​●​ ​Platform: Kaggle​

​●​ ​Join live challenges like credit scoring, fraud detection, sales predictions.​

​●​ W
​ hy Learn: Lets you​​compete, learn from others, and​​get ranked​​in the​
​community.​

​Resource 2:​​Awesome Public Datasets – GitHub​

​●​ ​Platform: GitHub​

​●​ ​100s of datasets across domains like finance, sports, agriculture, and education.​

​●​ ​Why Learn: Find​​unique datasets​​for personalized projects​​and ideas.​

​Resource 3:​​Data Science Hackathons – Analytics Vidhya​

​●​ ​Platform: Analytics Vidhya​

​●​ ​Timed hackathons and long-term ML contests with cash prizes.​

​●​ ​Why Learn: Simulates​​real-world competition-style​​thinking​​, great for practice.​

​-​ ​With ❤️ Harsha Verse​