Data Science Bootcamp

“Jack of all trades, master of none, though oft times better than master of one.”

An Intensive Bootcamp to build your Data Science Portfolio

The availability of data has provided a rich playground to build data-driven products to help take business decisions. Whether you want to predict the resale value of an second hand car, classify whether a customer will default on a loan product, or recommend which product a user is likely to buy next - they all use data science and machine learning in the process. The ability to take a business problem, frame it as an analytical problem and then provide a solution to the business to take decisions on it, has become an important skill to learn.

So how do you learn data science and get started on your journey to build data-driven product. Over the last three years, we have helped multiple organisations and professionals get started on learning data science - and have also written and talked about learning data science. There are two basic tenets on which we wanted to design the bootcamp. First, we want to cover the fundamental topics in data science through in-person structured sessions so that you can grok the concepts. Second, we want to provide enough elapsed time in between these sessions, so that the concepts learned can be consolidated by practice and allow you to start building your own data science portfolio.

In the five structured in-class days, you will learn how to solve business problem using data science (The Art of Data Science), the principle and application of data visualisation (Data Visualisation for Data Science), the math behind machine learning in a hacker’s way (HackerMath for ML), using machine learning in an applied context (Applied ML) and finally, how to create a data-driven product (FullStack Data Science). Between these classes, you will be working on a different data-set and applying what you have learned in the class. This way during the bootcamp, you would have started to build your own personal data science portfolio. Support for answering your queries as well as peer-to-peer learning will be provided through the use of messaging platform like Slack.

What will you learn?

Day 1: The Art of Data Science (Sunday, 20th Aug)

Learning data science involves understanding the process of approaching a business problem and going through a series of structured steps to find a decision solution. There is both a science and an art to the whole process. The goal of this day is to enable you to understand the end-to-end data science process through a case-driven approach.

• Frame: Problem Definition
  • Hypothesis-driven approach
  • Types: descriptive, inquisitive, predictive, causal
• Acquire: Data Ingestion
  • From flat files and databases
  • From third party clients / APIs
• Refine: Data Wrangling
  • Check for quality and completeness
  • Treatment of missing values and outlier
• Transform: Feature Creation
  • Tidy data
  • Split-Apply-Combine
  • Derive, parse, calculate
  • Filter, sample, summarize
• Explore: Feature Selection
  • Single and Dual variable visualisation
  • Multi-dimensional Visualisation
• Model: Model Creation
  • Regression: Linear
  • Classification: Logistics
• Insight: Model Selection
  • Accuracy metrics
  • Selection criteria
• Deploy: Model Deployment
  • Making an model api
  • Using the API
• Build: Decision Application
  • Dashboard vs Decision Tools
  • Narrative communications

GitHub Repo: https://github.com/amitkaps/art-data-science

Day 2: Data Visualisation for Data Science (Sunday, 27th Aug)

Visualisation plays a key component during the entire data science process - in exploratory data analysis, model visualisation as well as in communicating the results through a narrative or a dashboard. The goal of this day is to help you gain a deeper understanding on the art and science of data visualisation.

• Intro to visualization
  • Understand Exploratory Data Analysis (EDA)
  • Role of model visualisation
  • Communicating insights using visualisation
• The Grammar of Graphics
  • Visualizing single & dual variables
  • Visualizing categorical data
• Visualizing Multi-Dimensional Data
  • Using aesthetics and facets for more than two variables
  • Using matrix view, parallel coordinates for more than two variables
• Communicating with graphs
  • Graphical perception and critique
  • Understand color, scales, labeling and annotation
  • Theming and publication ready graphics
• Interactive graphics
  • Creating interactive graphs for the web
  • Allowing interactive data-model manipulation
• Visualizing geo-spatial data
• Creating Interactive Data Dashboards

Day 3: HackerMath for Machine Learning (Sunday, 3rd Sept)

Math literacy, including proficiency in Linear Algebra and Statistics, is a must for anyone learning data science. The goal of this day is to introduce the key concepts from these domains that get used repeatedly in data science applications. Our approach is what we call the “Hacker’s way”. Instead of going back to formulae and proofs, we will teach the concepts by writing code and in practical applications. Concepts don’t remain sticky if the usage is not made apparent.

• Linear Algebra
  • Matrix: Basics, Inverse
  • Solve for Ax=b for nxn
  • Solve for Ax=b for nxp+1
  • Application: Linear Regression
• Calculus and Numerical Optimisation
  • Cost Function
  • Gradient Descent
  • Application: Classification
• Statistics
  • Bootstrapping
  • Direct Simulation
  • Shuffling
  • Application: A/B Testing

Github Repo: https://github.com/amitkaps/hackermath

Day 4: Applied Machine Learning (Sunday, 10th Sept)

The challenge for many beginners is how do I navigate the landscape of possible ML models and then how do I choose the right model. The goal of this day is to go deeper in to the model building, evaluation and selection process. Real-life case studies are used to teach the various algorithms and techniques. The focus will be on applications, rather than on exposition of the various algorithms.

• Model Building
  • Decision Trees
  • Bagging and Boosting
  • Random Forest
  • Gradient Boosting Machines
  • Feature Importance
• Model Evaluation
  • Training and Validation
  • Model Evaluation Metrics - Accuracy, ROC, AUC, Confusion Matrix etc.
  • Overfitting and Bias-Variance trade-off
  • Regularization (L1/L2)
  • K-fold Cross Validation
• Model Selection
  • Model Pipelines
  • Feature Engineering
  • Ensemble Models (Advanced)

Github Repo: https://github.com/amitkaps/applied-machine-learning

Day 5: Full Stack Data Science (Sunday, 17th Sept)

One of the common use case in building a data-driven app is to create an API and provide seamless integration with other business applications like a dashboard. The goal of this day is on building a basic understanding of server-side programming and front-end application, and allowing you to start creating a data-product from your ML models.

• Overview of the case
• Build Simple ML model (linear/logistic regression)
• Creating RESTful API
• Integrating model output to DB
• Updating the model as more data comes in (batch only - no streaming)
• A simple webpage front-end to visualise the results and interact with the API.
• Creating a simple application that accomplishes this end-to-end

Github Repo: https://github.com/amitkaps/full-stack-data-science

Who is it for?

• A programmer but not a data science practioner: A programmer with experience in server-side or front-end development and maybe has some familiarity with doing data analysis. You could be looking to transition in to building data driven products or a create a richer product experience with data.
• A data science practioner but not a programmer: A data science newbie with some experience in doing data analysis, preferably in a scripting language (R/Python/Scala), but wants to get a deeper and a more richer experience in data science.

Testimonial

“The instant Amit starts to talk, his attention to detail and clarity of thought is unmissable. Having learnt from him, I’ve always been astounded by the amount of effort that he puts into his content. And when he presents this content, it is understandable, relatable and the delivery is on point. I wouldn’t think twice about attending a workshop that he conducts. A total value for money and time.” – Shrayas R, Head of engineering at Logic Soft

“Enjoyed the workshop overall and really appreciate Amit’s smooth coordinating skills, alongside actual Data science skill sets in teaching.” – Vijay Kumar, Lead data scientist at GE Digital

“Wonderful session. People usually just teach how to use a library. But, Amit and Bargava taught how to approach the problem.” – Dhilipsiva, Full stack engineer at AppKnox

Pre-requisites

• Programming knowledge is mandatory. Attendee should be able to write conditional statements, use loops, be comfortable writing functions and be able to understand code snippets and come up with programming logic.
• Participants should have a basic familiarity of Python. Specifically, we expect participants to know the first three sections from this: http://anandology.com/python-practice-book/

Software Requirements

We will be using Python data stack for the workshop. Please install Ananconda for Python 3.5 for the workshop. Additional requirement will be communicated to participants.

Facilitators’ Profile

Amit Kapoor teaches the craft of telling visual stories with data. He conducts workshops and trainings on Data Science in Python and R, as well as on Data Visualisation topics. His background is in strategy consulting having worked with AT Kearney in India, then with Booz & Company in Europe and more recently for startups in Bangalore. He did his B.Tech in Mechanical Engineering from IIT, Delhi and PGDM (MBA) from IIM, Ahmedabad. You can find more about him at http://amitkaps.com/ and tweet him at @amitkaps.

Bargava Subramanian is a practicing Data Scientist. He has 14 years of experience delivering business analytics solutions to Investment Banks, Entertainment Studios and High-Tech companies. He has given talks and conducted workshops on Data Science, Machine Learning, Deep Learning and Optimization in Python and R. He has a Masters in Statistics from University of Maryland, College Park, USA. He is an ardent NBA fan. You can tweet to him at @bargava.

Anand Chitipothu is a software consultant and trainer based in Visakhapatnam. He has over 13 years of experience in architecting and developing variety of software applications. He is co-author of web.py, a micro web framework in Python. He has worked at Strand Life Sciences and Internet Archive. You can tweet him at @anandology