Roadmap to Crack Data

Science/ML Interviews
This roadmap, constructed from
information solely from the sources
you provided, offers a detailed plan
to guide you on your journey to
becoming a data scientist.

Anwar Haq
Sr. Data Scientist @ Cisco
Ex-Meta | Ex-Wells Fargo
 02

Introduction
Entering the World of Data Science

Diverse Roles in the

Data Science Family
The field of data science encompasses a range of roles
with varying responsibilities and skill requirements.
Some key roles include:

Data Scientist:

Analyses complex datasets to identify patterns,

What is Data Science? build predictive models, and deliver actionable
Data science is a insights for business decisions.

multidisciplinary field that

Data Analyst:
extracts valuable insights
Interprets data to produce reports and
from data. It encompasses visualizations, focusing on descriptive statistics
various techniques, from to answer specific business questions.
simple data analysis using
SQL queries to building Machine Learning Engineer:
Develops, tests, and deploys machine learning
complex predictive models
models that can learn from data and make
using machine learning predictions. They focus on algorithm
algorithms. The demand for development and optimization.
data scientists is high and
Data Engineer:
shows resilience even during
Designs and maintains the architecture for large-
economic downturns, making scale data processing systems, ensuring
it a promising career path. efficient, scalable, and reliable data pipelines.

Data Architect:
Designs and manages the organization's data,
ensuring data accessibility and well-organised
metadata.

Roadmap to Crack Data Science/ML Interviews

 03

Choosing Your Industry Needs:

Research industries that interest you and

Path see which data science roles are in

demand. For example, healthcare needs
data scientists knowledgeable in
Interests: healthcare data and regulations, while
Are you passionate about building models finance prefers professionals skilled in
(consider Machine Learning Engineer), risk modelling and quantitative analysis.
analysing trends (Data Analyst or Data
Scientist), or data infrastructure (Data Career Prospects:
Engineer)? Investigate the growth potential, average
salaries, and job availability for the roles
Skills: you're considering.
Different roles require specific skillsets.
Assess your strengths and identify areas
for development.

Roadmap to Crack Data Science/ML Interviews

 04

Research and
Networking
Laying the Groundwork
Identifying Target Companies

Strategically selecting target companies is

crucial for a focused job search. Consider:

Define Career Goals: What do you want to

achieve in your career? Specific industry? Startup
or large corporation? Research institution?

Research Industry Leaders: Identify companies

known for innovation, data-driven approaches,
and culture.

Consider Company Size and Growth Stage:

Startup (varied experience) vs. larger corporation
(stability).

Evaluate Company Values and Culture: Does the

company's mission and values align with yours?
Check employee reviews on platforms like
Glassdoor.

Check Job Listings and Career Pages: Get a

sense of roles and valued skills.
Geographic Preferences: Does the location match
your preferences?

Roadmap to Crack Data Science/ML Interviews

 05

Analyzing Job Descriptions

Carefully examine job descriptions to understand required skills and
qualifications:

Identify Key Skills and Qualifications: Recurring technical skills (Python, SQL, machine
learning) and soft skills (communication, problem-solving).

Understand Role Responsibilities: What does the daily work entail? Does it align with
your interests and skills?

Look for Required Experience: Years of experience needed.

Check for Preferred Skills: Can give you an edge.

Analyse Job Titles: Different titles for similar roles ("Data Analyst" vs. "Business
Analyst").

Evaluate Job Requirements: Tailor your resume and cover letter accordingly.

Note Industry-Specific Requirements: E.g., healthcare data science roles may require
knowledge of HIPAA regulations.

Roadmap to Crack Data Science/ML Interviews

 06

Building Your
Network
Networking is vital for your job search and
career development.

Use LinkedIn Effectively:

Build a Strong Profile: Highlight your data

science skills, experiences, and achievements.
Connect with Industry Professionals: Data
scientists, hiring managers, and professionals
at your target companies.
Follow Relevant Companies and Influencers:
Stay updated on industry trends and job
openings.

Attend Industry Events: Conferences, meetups,

and workshops provide opportunities to learn,
network, and connect with potential employers.

Engage in Online Communities: Participate in data

science forums and contribute to open-source
projects. Platforms like Kaggle and DataCamp offer
communities for sharing knowledge and
collaborating on projects.

Seek Mentorship: Find experienced data scientists

for guidance and support. Leverage company
mentorship programs.

Roadmap to Crack Data Science/ML Interviews

 07

Building
Foundational
Skills
Statistics and Mathematics
1. Descriptive Statistics
Summarizing and describing data using measures like:

Mean: The average of a dataset.

Median: The middle value in a sorted dataset.
Mode: The most frequent value in a dataset.
Standard Deviation: A measure of how spread out the data is.

2. Inferential Statistics
Making inferences and predictions about a population based on a sample.
Key concepts:

Hypothesis Testing: Testing assumptions about a population based on sample data.

Confidence Intervals: Estimating a range of values for a population parameter.

3. Statistical Tests

T-Tests: Comparing means of two groups.

ANOVA (Analysis of Variance): Comparing means of multiple groups.
Regression Analysis: Modelling the relationship between variables.
Time Series Analysis: Analysing data over time.

The depth of your statistical knowledge will depend on your chosen data
science role.

Roadmap to Crack Data Science/ML Interviews

 08

Resources
Books: "Practical Statistics for Data Scientists" provides a beginner-friendly introduction
with Python examples.

Online Courses: Platforms like Coursera and edX offer statistics courses tailored for data
science, such as "Statistics with R" and "Inferential Statistics."

YouTube Channels: StatQuest with Josh Starmer offers clear and engaging explanations
of statistical concepts.

Time allocation:
Allocate approximately 3 weeks to build a solid foundation in statistics.

Roadmap to Crack Data Science/ML Interviews

 09

Building Foundational Skills

Exploratory Data Analysis (EDA)
EDA is the crucial first step in any data science project, enabling you to understand your
data through cleaning, preprocessing, and visualisation techniques.

1. Data Cleaning and Preprocessing

Handling Missing Values: Imputing missing values or removing rows with missing
data.
Detecting and Treating Outliers: Identifying and addressing extreme values that can
skew your analysis.
Data Transformation: Converting data types, scaling variables, and creating new
features.

2. Univariate Analysis
Exploring individual variables using:

Histograms: Visualizing the distribution of numerical data.

Box Plots: Displaying the distribution and identifying outliers.
Summary Statistics: Calculating measures like mean, median, and standard deviation.

3. Bivariate Analysis
Examining relationships between pairs of variables using:

Scatter Plots: Visualizing the relationship between two numerical variables.

Correlation Coefficients: Measuring the strength and direction of the relationship.
Cross-Tabulations: Analysing relationships between categorical variables.

Roadmap to Crack Data Science/ML Interviews

 10

4. Multivariate Analysis
Investigating interactions among multiple variables using techniques like:

Principal Component Analysis (PCA): Reducing the dimensionality of data while

retaining important information.

Time allocation:
Allocate 3 weeks to develop proficiency in EDA.

Roadmap to Crack Data Science/ML Interviews

 11

Building
Foundational
Skills
Data Visualization
Data visualization is crucial for communicating insights and findings to both technical and
non-technical audiences. You will need to:

Select appropriate visualization techniques: Choose the best visualization to represent

your data and insights effectively, considering the type of data and the message you want
to convey. For example:

Bar charts and line graphs are suitable for comparing categories or showing trends over
time.
Scatter plots are useful for visualizing relationships between two variables.
Heatmaps are effective for displaying patterns in large datasets.

Use effective design principles: Create visually appealing and easy-to-understand

visualizations by considering factors like color choice, labelling, and chart layout.

Tell a story with your data: Use visualizations to support a narrative, highlighting key
insights and trends.

Resources
Python Libraries: Matplotlib and Seaborn offer powerful tools for creating static and
interactive visualizations.

R Packages: Ggplot2 is a popular package in R for creating aesthetically pleasing and

informative visualizations.

Data Visualisation Tools: Tableau, Power BI, and even Excel can be used for creating
interactive dashboards and reports.

Books: "Storytelling with Data" by Cole Nussbaumer Knaflic provides a comprehensive

guide to creating impactful visualizations and telling stories with data.

Roadmap to Crack Data Science/ML Interviews

 12

Online Courses: DataCamp's "Understanding Data Visualization" course helps deepen

your data visualization competencies.

Time allocation:
Allocate 3 weeks to develop your data visualisation skills, focusing on both technical skills
and design principles

Roadmap to Crack Data Science/ML Interviews

 13

Acquiring Essential Coding Skills

Python

Python is the dominant language in data science due to its versatility, simplicity,
and extensive libraries.

1. Core Python

Syntax: Understanding the basic rules of Python code.

Data Types: Working with numbers, strings, booleans, and lists.
Control Flow: Using conditional statements and loops to control program
execution.
Functions: Creating reusable blocks of code to perform specific tasks.

2. Essential Libraries

NumPy: For numerical computing and array manipulation.

Pandas: For data manipulation and analysis.
Scikit-learn: For machine learning algorithms and model building.
Matplotlib: For data visualization.

3. Advanced Techniques

Object-Oriented Programming (OOP): For creating reusable and modular code.

Exception Handling: For gracefully managing errors in your code.
File Input/Output: For reading and writing data from files.
Regular Expressions: For pattern matching and text manipulation.

Roadmap to Crack Data Science/ML Interviews

 14

Resources
Tutorials: The official Python documentation and Real Python offer comprehensive
resources.

Online Courses: Coursera's "Python for Everybody" and Udemy's "Complete Python
Bootcamp" are popular choices.

Projects: Participate in Kaggle competitions and contribute to GitHub repositories to

practice your Python skills.

Time allocation:
Allocate approximately 3 weeks to gain proficiency in the basics of Python and an
additional 7 weeks to enhance your skills while working on projects.

Roadmap to Crack Data Science/ML Interviews

 15

Acquiring Essential
Coding Skills
SQL
SQL (Structured Query Language) is
essential for interacting with databases,
enabling you to extract, manipulate, and
analyse data.

SQL (Structured Query Language) is essential for interacting with databases, enabling
you to extract, manipulate, and analyse data.

1. Fundamental Concepts

SELECT Statement: Retrieving data from tables.

WHERE Clause: Filtering data based on conditions.
ORDER BY Clause: Sorting data in ascending or descending order.
JOIN Operations: Combining data from multiple tables based on relationships.
Aggregate Functions: Performing calculations like SUM, AVG, COUNT, MIN, and
MAX.
GROUP BY Clause: Grouping data for aggregate calculations.

2. Advanced Techniques

Subqueries: Queries nested within other queries to perform more complex

operations.
Analytical Functions: Providing advanced data analysis capabilities, like ranking and
windowing.
Common Table Expressions (CTEs): Simplifying complex queries.

Roadmap to Crack Data Science/ML Interviews

 16

Resources
Tutorials: W3Schools SQL Tutorial and Mode Analytics SQL Tutorial offer practical
lessons.

Online Courses: Coursera's "SQL for Data Science" and DataCamp's "Introduction to SQL"
are recommended.

Projects: Practice SQL problems on platforms like LeetCode and SQLZoo.

Time allocation:
Allocate approximately 3 weeks to achieve job readiness in SQL.

Roadmap to Crack Data Science/ML Interviews

 17

Machine Learning
The Heart of Data Science

Introduction to Machine Learning

Machine learning, a core component of data science, involves training algorithms to learn
patterns and make decisions based on data.

1. Supervised Learning
Training models on labelled data, where the outcome is known. Key techniques:

Regression: Predicting a continuous target variable.

Linear Regression
Polynomial Regression
Classification: Predicting a categorical target variable.
Logistic Regression
Decision Trees
Support Vector Machines (SVMs)

2. Unsupervised Learning
Exploring unlabeled data to discover patterns. Key techniques:

Clustering: Grouping data points based on similarity.

K-Means Clustering
Hierarchical Clustering
Dimensionality Reduction: Reducing the number of features while preserving important
information.
Principal Component Analysis (PCA)

Roadmap to Crack Data Science/ML Interviews

 18

3. Deep Learning
Using artificial neural networks to analyse complex data and solve challenging tasks. Key
concepts:

Neural Networks: Interconnected nodes that process and transmit information.

Activation Functions: Introducing non-linearity into the network.
Backpropagation: Adjusting weights to improve model performance.
Convolutional Neural Networks (CNNs): Specialized for image and video processing.
Recurrent Neural Networks (RNNs): Designed for sequential data like time series and
text.

4. AI Feedback Loops
Incorporating feedback mechanisms to enable models to learn from their output.
Examples:

Reinforcement Learning: Training agents to make decisions in an environment

through rewards and penalties.

Resources
Books: "Hands-On Machine Learning with Scikit-Learn, Keras & TensorFlow" provides a
practical guide.

Online Courses: Coursera's "Machine Learning" by Andrew Ng and DataCamp's "Machine

Learning Scientist with Python" career track are highly recommended.

YouTube Channels: StatQuest with Josh Starmer and 3Blue1Brown offer intuitive
explanations of machine learning concepts.

Time allocation:
Allocate 3 weeks for learning fundamental concepts in machine learning and an additional
7 weeks for delving deeper into different areas and building portfolio projects.

Roadmap to Crack Data Science/ML Interviews

 19

Mastering Data Science

Tools
Essential Data Science Tools
Proficiency in various tools is essential for coding, version control, visualization, and
deployment of data science projects.

1. Notebooks and IDEs

Jupyter Notebook: An interactive environment for running code, visualizing data, and
creating documentation.
Google Colab: A cloud-based platform similar to Jupyter Notebook, offering free access to
GPUs for deep learning.
Integrated Development Environments (IDEs): Specialized software for code editing,
debugging, and project management. Popular IDEs include:

PyCharm
VS Code
Spyder

2. Version Control with Git

Git: A distributed version control system for tracking changes in code and collaborating
with others.
GitHub: A popular platform for hosting Git repositories and collaborating on projects

Key Concepts:
Branching
Merging
Pull Requests

Roadmap to Crack Data Science/ML Interviews

 20

3. Data Visualisation Platforms

Tableau: A powerful tool for creating interactive dashboards and visualizations.

Power BI: A business intelligence platform for data analysis and reporting.
Plotly: A library for creating interactive and web-based visualizations.

4. Cloud Computing Platforms

Amazon Web Services (AWS): A comprehensive cloud platform offering various services
for storage, computing, and machine learning.
Microsoft Azure: A cloud platform with services for data storage, analytics, and AI.
Google Cloud Platform (GCP): A suite of cloud computing services for data storage,
analysis, and machine learning.

Time allocation:
You can incorporate these tools into your workflow as you progress through the other
steps, allocating dedicated time as needed.

Roadmap to Crack Data Science/ML Interviews

 21

Building Your Portfolio

A well-crafted portfolio showcases your skills and practical experience, making you a more
competitive candidate.

1. Project Selection
Choose projects that cover diverse aspects of data science, demonstrating a broad skillset

Data Cleaning and Preprocessing: Showcase your ability to handle real-world data
with missing values, inconsistencies, and outliers.
Exploratory Data Analysis (EDA): Demonstrate your skills in data exploration,
visualization, and identifying patterns.
Machine Learning Model Building: Include projects involving various algorithms and
techniques, such as classification, regression, clustering, and deep learning.
Domain-Specific Projects: If you have a specific industry in mind, build projects that
address problems in that domain. For example, you could create a project on customer
churn prediction for the telecom industry or a fraud detection model for the financial
sector.

2. Project Documentation
Clearly document each project, making it easy for others to understand your work

Introduction: Describe the problem you addressed and the project's objectives.
Data: Explain the datasets used, their source, and any preprocessing steps taken.
Methodology: Detail the techniques and algorithms used for analysis and model
building.
Results: Present your findings using clear visualizations and insightful interpretations.
Conclusion: Summarize your key takeaways and any limitations of your approach.
Code: Include well-commented code snippets to demonstrate your coding skills.

Roadmap to Crack Data Science/ML Interviews

 22

3. Creating an Online Portfolio

Host your projects on platforms like:

GitHub: A popular platform for sharing code and collaborating on projects. Create a
well-organized repository for each project, including your code, documentation, and
any supporting files.
Personal Website: Build a website to showcase your projects and skills professionally.
Include a portfolio section with links to your GitHub repositories or detailed project
descriptions.

4. Highlighting Key Skills

Ensure your portfolio highlights your proficiency in essential data science skills:

Data Preprocessing
Statistical Analysis
Machine Learning
Data Visualization
Tools and Technologies Used (Python, R, SQL, TensorFlow, Tableau)

Resources
Books: "Data Science Projects with Python" by Stephen Klosterman and "Python for Data
Analysis" by Wes McKinney offer guidance on building practical projects.

Online Platforms: Kaggle's "Learn" courses provide project-based learning opportunities.

Portfolio Examples: Explore portfolios of experienced data scientists for inspiration and
ideas.

Time allocation:
Allocate approximately 7 weeks to develop 7 portfolio projects, working on documentation
and website creation alongside your project development.

Roadmap to Crack Data Science/ML Interviews