Data Science Roadmap by Swapnil Microsoft

Roadmap to Data
Science
by
SWAPNIL NARAYAN
Microsoft | IIT | Hacker Cup

About the
Instructor
Hey there,
I’m Swapnil Narayan, a graduate from
IIT(ISM) Dhanbad with Computer
Science majors.
I’m a Software Engineer at Microsoft
India, and have also got offers from
Amazon and Oracle for Software
Engineering roles.
I’m a very passionate Competitive
Programming Instructor and have a
decent experience for the same at
various popular edtech platforms, and
have taken sessions with IITs, NITs, and
other engineering colleges .
I will be your mentor for this session and
will walk you through the topics the
following slides.

What is Data Science??
O Data Science as a multi-disciplinary subject encompasses the use
of mathematics, statistics, and computer science to study and
evaluate data. The key objective of Data Science is to extract
valuable information for use in strategic decision making,
product development, trend analysis and forecasting.
O A Data scientist is sort of 'jack-of-all-trades' for data crunching.
Basically, 3 main skills a data scientist needs to possess are
mathematics/statistics, computer programming literacy and
knowledge of particular business.

Data Science is a Broader Field

Comparison between Different Roles
in 2018

How to become a Data Scientist??
Math
Programming
Languages
Data Wrangling and
Management
Data Analysis and Visualization
Machine Learning
Deep Learning

Mathematics
O Linear Algebra: Matrix, Eigen, Tensor etc.
O Calculus: Differentiation and Integration.
O Probability: Bayes Theorem, Optimization etc.
O Statistics: Inferential Statistics, Descriptive Statistics, Chi
squared Testes, Random Variable, Gaussian And Normal
Distribution.

Programming Languages
O Python: It is the Bible.
→ Easy to understand, i.e., plane English
→ No semicolon
→ Simple and tons of libraries available
O Talk about Packages
→ Data visualization using ggplot2, tidy are extremely important

Data Wrangling and Management
O Data Mining
O Data Cleaning
O Data Management
Relevant Skills:
→ MySQL: RDBMS
→ NoSQL: Mongo DB, Cassandra etc.
JOIN

Data Analysis and Visualization
O Plotting libraries in programming languages, e.g.,
• plotly, matplotlib, seaborn → python
• ggplot2 → R
• Tableau and PowerBI is booming now.
[Pandas and Numpy for Data Analysis]

Machine Learning and Deep Learning
O Domain Knowledge???
HEALTHCARE, BUSINESS, FINANCE, SPORTS etc.
Supervised Unsupervised Reinforcement

Machine Learning Algorithms
O Topics: Regression, Decision Tree, Random Forest, Naïve
Bayes, Ensemble Learning, AdaBoost, Hierarchical
Clustering, Association, k-means Clustering, SVM, KNN,
Gradient Descent, Cross Validation, Entropy, Accuracy,
Precision, Collaborative Filtration, PCA, Markov model,
Boltzmann theorem etc.
Testing Evaluation and Validation of Models

Deep Learning Algorithms
O Neural Networks, Feed Forward NN, Fuzzy Logic,
Sequence Model, LSTM, RNN, CNN, CapsNet, Time Series
etc

Big Data
O Map Reduce
O Hadoop
O Apache
O Spark
O Hive
O Pig
O Mahout
O Yarn

Learning Outcomes
O Build artificial neural networks with Tensorflow and Keras
O Build Deep Learning networks to classify images with
Convolutional Neural Networks
O Implement machine learning, clustering, and search using TF/IDF
at massive scale with Apache Spark's MLLib
O Implement Sentiment Analysis with Recurrent Neural Networks
O Understand reinforcement learning - and how to build a Pac-Man
bot

O Make predictions using linear regression, polynomial
regression, and multivariate regression
O Implement Sentiment Analysis with Recurrent Neural
Networks
O Understand reinforcement learning - and how to build a
Pac-Man bot
O Classify medical test results with a wide variety of
supervised machine learning classification techniques
O Cluster data using K-Means clustering and Support Vector
Machines (SVM)

O Build a spam classifier using Naive Bayes
O Use decision trees to predict hiring decisions
O Apply dimensionality reduction with Principal Component
Analysis (PCA) to classify flowers
O Predict classifications using K-Nearest-Neighbor (KNN)
O Develop using iPython notebooks
O Understand statistical measures such as standard deviation
O Visualize data distributions, probability mass functions, and
probability density functions
O Visualize data with matplotlib

O Use covariance and correlation metrics
O Apply conditional probability for finding correlated
features
O Use Bayes' Theorem to identify false positives
O Understand complex multi-level models
O Use train/test and K-Fold cross validation to choose the
right model
O Build a movie recommender system using item-based and
user-based collaborative filtering
O Clean your input data to remove outliers
O Design and evaluate A/B tests using T-Tests and P-Values

Best Blogs and Open Source
Community
O Medium AI Community
O Official Documentations
O Github and Stackoverflow
O Kaggle- Spend 5 hours of a day here
O Cheat Sheets from Amazon aws

Best Books
For Machine/ Deep Learning Data Science
Beginners
Book
Statistics

Overview of Data Science Tools and
Packages

Data Science Roadmap by Swapnil Microsoft

More Related Content

Similar to Data Science Roadmap by Swapnil Microsoft (20)

Recently uploaded (20)

Data Science Roadmap by Swapnil Microsoft