josrod/data-science-road-map

A roadmap for those looking to start or expand a career in the data community

A prioritized list of core skills, reading material, personal portfolio projects and practice assignments every new data scientist should have.

This repo is a companion site for the course 2023 CORE: Data Science and Machine Learning.

The goal here is very ambitious; to be the only reference site you need on your inital learning journey as a data scientist. Getting started in data science is hard. There is an overwhelming number of resources and suggestions. Many people give up!

The goal of this road map is to get you started (or re-directed) on your journey the right way with NO knowledge gaps. Everything from original content to linked resources has been vetted through experience and practice.

Data jobs come in three general flavors (but are often called different things in practice); Data Analyst, General Data Scientist, and Machine Learning Engineer. The skills, course, and resources are divided into these categories for ease of reference. Pay close attentioin to recommended books to add to you library, several are must reads.

Table of Contents

1. START LEARNING HERE - Combined Core Skills List
2. Certification Checklist
3. Portfolio Project Checklist
4. Foundations
   1. Library Recommendations
   2. Data Science Communities
   3. Public Data
   4. Resume Tools
   5. Interview Prep
5. Data Analyst
   1. Library Recommendations
   2. Cheatsheets
   3. Spreadsheets
   4. SQL
   5. Business Intelligence
6. General Data Scientist
   1. Library Recommendations
   2. Cheatsheets
   3. Probability and Statistics
   4. R Resources
   5. web-development
7. Machine Learning Engineer
   1. Library Recommendations
   2. Cheatsheets
   3. Python Resources
   4. Math for Machine Learning
   5. Model Deployment
   6. Deep Learning
8. Next Steps
9. Latest News

---

Combined Core Skills List

Start learning at the top of this list and check each skill until you are done. When done you will have gained all core skills required for data science!

This table is the set of skills that should be common to every data scientist. There are many things excluded that become important as individuals specialize. This list represents the absolute foundation.

Role	Skill	Type	Tool
Foundational	Define data science	Soft	Any
	Explain why data science is important	Soft	Any
	Give examples of data science projects	Soft	Any
	Know how to get public datasets	Soft	Any
	Participate in the data science community	Soft	Any
	Build a project portfolio	Soft	Any
Data Analyst	Explain what a data analyst is	Soft	Any
	Understand summary statistics: location, shape, spread, and dependence	Math	Any
	Mathematical modeling (linear programming)	Math	Any
	Setup MS Excel on Desktop and Cloud	Spreadsheet	Excel
	Use operators	Spreadsheet	Excel
	Use built-in functions	Spreadsheet	Excel
	Import a text file	Spreadsheet	Excel
	Use data tables w/ summary stats	Spreadsheet	Excel
	Import data from various sources	Spreadsheet	Excel
	Lookups and Matches	Spreadsheet	Excel
	Understand data visualization concepts	Soft	Any
	Data visualization	Spreadsheet	Excel
	Build a dashboard w/ KPIs	Spreadsheet	Excel
	Import data with Power Query	Spreadsheet	Excel
	Use pivot tables	Spreadsheet	Excel
	Use the analysis tool pack	Spreadsheet	Excel
	Use VBA and macros to automate tasks	Spreadsheet	Excel
	Explain what a database is	Database	Any
	Understand what tools are required to write SQL	Database	Text Editor
	Understand SQL Syntax	Database	Text Editor
	Build a SQLite database from scratch	Database	Text Editor
	Use SQL Statements: SELECT, FROM, WHERE	Database	Text Editor
	Use SQL Statements: BETWEEN, LIKE	Database	Text Editor
	Use SQL Statements: AND, OR, NOT, EXISTS, NULL	Database	Text Editor
	Use SQL Statements: ORDER BY, DISTINCT	Database	Text Editor
	Use SQL Aggregate Functions	Database	Text Editor
	Use SQL WITH statement and subqueries	Database	Text Editor
	Use SQL for modifying data with inserting, updating and deleting	Database	Text Editor
	Understand SQL views	Database	Text Editor
	Connect Excel to SQLite and execute SQL from within Excel	Database	Excel
	Explain what Business Intelligence is	Soft	Any
	Install Tableau	Business Intelligence	Tableau
	Use Tableau data types	Business Intelligence	Tableau
	Build Tableau visualizations	Business Intelligence	Tableau
	Create Tableau filters	Business Intelligence	Tableau
	Connect Tableau to external data sources	Business Intelligence	Tableau
	Join data in Tableau	Business Intelligence	Tableau
	Understand Tableau dates	Business Intelligence	Tableau
	Build Tableau visualizations for comparisons	Business Intelligence	Tableau
	Build Tableau visualizations for distributions	Business Intelligence	Tableau
	Build Tableau visualizations for multiple axis	Business Intelligence	Tableau
	Understand Tableau formatting	Business Intelligence	Tableau
	Build calculations and parameters in Tableau	Business Intelligence	Tableau
	Understand data story telling concepts	Soft	Any
	Build Tableau dashboards and stories	Business Intelligence	Tableau
	Share Tableau dashboards and stories (with Tableau Public)	Business Intelligence	Tableau
	Understand the difference between Tableau Public and Pro	Business Intelligence	Tableau
Data Scientist	Explain what a data scientist is	Soft	Any
	Explain why using a scripting language is important	Soft	Any
	Explain what R, CRAN, and RStudio are	Soft	Any
	Install base R	R	base R
	Install RStudio	R	RStudio
	Use base R calculations	R	RStudio
	Understand objects in R	R	RStudio
	Understand functions in R	R	RStudio
	Understand what an R script is	R	RStudio
	Use base R datasets	R	RStudio
	Use the help functions in R	R	RStudio
	Use base R plots	R	RStudio
	Install R packages	R	RStudio
	Understand atomic vectors	R	RStudio
	Understand object attributes	R	RStudio
	Use matrix and array objects	R	RStudio
	Understand classes	R	RStudio
	Understand factors	R	RStudio
	Understand coercion	R	RStudio
	Use lists	R	RStudio
	Use data frames	R	RStudio
	Load and save data	R	RStudio
	Select values from a data frame	R	RStudio
	Change values in a data frame	R	RStudio
	Subset a data frame	R	RStudio
	Deal with missing values	R	RStudio
	Understand control flow	R	RStudio
	Conduct an Exploratory Data Analysis (EDA) using summary stats, and viz	R	RStudio
	Explain the difference between base R and the Tidyverse	R	RStudio
	Use ggplot mapping aesthetics	R	RStudio
	Use ggplot facets	R	RStudio
	Use ggplot multiple geom	R	RStudio
	Use ggplot stat transforms	R	RStudio
	Use ggplot position adjustments	R	RStudio
	Use ggplot coord systems	R	RStudio
	Use dplyr filter	R	RStudio
	Use dplyr arrange and select	R	RStudio
	Use dplyr mutate	R	RStudio
	Use dplyr pipes, group_by, and summaries	R	RStudio
	Use stringer for text manipulation	R	RStudio
	Explain what Markdown and RMarkdown are	Soft	RStudio
	Build and share an EDA using RMarkdown	R	RStudio
	Understand useful probability concepts	Math	RStudio
	Understand probability distributions	Math	RStudio
	Understand statistical hypothesis testing (comparison on means)	Math	RStudio
	Understand A-B testing	Math	RStudio
	Understand bootstrap statistical methods	Math	RStudio
	Understand the difference between frequentists and Bayesian stats	Math	RStudio
	Understand conjugate priors and Thompson sampling	Math	RStudio
	Understand Monte Carlo simulations	Math	RStudio
	Understand simple and multiple linear regression for inference	Math	RStudio
	Understand timeseries modeling	Math	RStudio
	Use web hosting tools to share analysis	Web Development	GitHub
	Use Git version control to manage code (GitHub)	Git	GitHub
	Create interactive analysis web hosted tools	R	R Shiny
Machine Learning Engineer	Explain what a Machine Learning Engineer is	Soft	Any
	Understand what the cloud and cloud service providers are	Soft	Any
	Create a cloud hosted virtual machine	Cloud	AWS
	Use a Command Line Interface (CLI)	CLI	Ubuntu/Terminal
	Understand what docker is	Containers	Docker
	Deploy a docker container on a cloud VM	Containers	Docker
	Explain project jupyter, jupyterlab, and the docker stacks	Containers	Docker
	Explain what python is	Soft	base Python
	Understand what a Jupyter Notebook is	Soft	Jupyter
	Use basic math operations	Python	Jupyterlab
	Use basic data types	Python	Jupyterlab
	Use variables	Python	Jupyterlab
	Use built-in functions	Python	Jupyterlab
	Use comparison operators	Python	Jupyterlab
	Use Boolean operators	Python	Jupyterlab
	Combine comparison and Boolean operators	Python	Jupyterlab
	Understand control flow and code chunks	Python	Jupyterlab
	Import modules	Python	Jupyterlab
	Create functions	Python	Jupyterlab
	Understand the difference between local and global variables	Python	Jupyterlab
	Use lists	Python	Jupyterlab
	Use additive operators	Python	Jupyterlab
	Use methods on lists	Python	Jupyterlab
	Use dictionaries	Python	Jupyterlab
	Understand classes and methods	Python	Jupyterlab
	Interact with files	Python	Jupyterlab
	Explain why python is good for data science	Soft	base Python
	Use matrix operations and linear algebra	Math	Jupyterlab
	Explain what numpy is	Soft	base Python
	numpy for matrix operations	Python	Jupyterlab
	numpy indexing and slicing	Python	Jupyterlab
	numpy Boolean indexing	Python	Jupyterlab
	numpy reshape and transpose	Python	Jupyterlab
	numpy pseudorandom numbers	Python	Jupyterlab
	numpy unary and binary functions	Python	Jupyterlab
	numpy aggregate functions	Python	Jupyterlab
	numpy saving and loading data	Python	Jupyterlab
	Explain what pandas is	Soft	base Python
	pandas read data	Python	Jupyterlab
	pandas for basic data exploration	Python	Jupyterlab
	pandas at and iat	Python	Jupyterlab
	pandas reshaping data	Python	Jupyterlab
	pandas subsetting	Python	Jupyterlab
	pandas summarizing	Python	Jupyterlab
	pandas group_by	Python	Jupyterlab
	pandas handling missing data	Python	Jupyterlab
	pandas and plotting	Python	Jupyterlab
	Explain what matplotlib and seaborn are	Soft	base Python
	Use matplotlib for data viz	Python	Jupyterlab
	Use seaborn for data viz	Python	Jupyterlab
	Use pandas with seaborn	Python	Jupyterlab
	Explain they various types of machine learning	Soft	Any
	Understand why training data is so important	Soft	Any
	Understand trade-offs in model selection	Soft	Any
	Understand what hyperparameters are	Soft	Any
	Understand over/under fitting	Soft	Any
	Understand bias-variance trade-off	Soft	Any
	Understand how and way training data is split	Soft	Any
	Understand how supervised models are evaluated for quality	Soft	Any
	Calculate regression measures of quality	Math	Any
	Calculate classification measures of quality	Math	Any
	Use a heuristic to create a model	Python	Jupyterlab
	Understand the supervised model training paradigm of improvement through iteration	Soft	Any
	Understand role of a cost function for optimizing parameter selection	Soft	Any
	Use linear regression and the OLS cost function	Math	Jupyterlab
	Use logistic regression and the cross-entropy cost function	Math	Jupyterlab
	Use CART models for regression and classification	Math	Jupyterlab
	Use ensemble models - random forest	Math	Jupyterlab
	Use ensemble models - xgboost	Math	Jupyterlab
	Conduct feature engineering using unsupervised learning	Math	Jupyterlab
	Explain what deep learning is	Soft	Any
	Use deep learning APIs	Soft	OpenAI/AWS
	Package an ML model as a microservice	Containers	Docker

Certification Checklist

Certifications are a tricky thing. They don't really demonstrate mastery but can make the difference on getting an interview. Here are our minimum recommended certifications. However, if you cannot afford to complete these certifications don't worry! Use the Kaggle courses and LinkedIn Assessments instead and let your project portfolio show your competence!

Category	Name	Link	Notes
All	2023 CORE: Data Science and Machine Learning	Link
Data Analyst	LinkedIn Excel	Link
Data Analyst	Kaggle SQL	Link and Link
Data Analyst	Tableau Data Analyst	Link
General Data Scientist	LinkedIn R Assessment	Link
Machine Learning Engineer	Andrew Ng's Intro ML Course	Link
Cloud - ML	AWS Certified Machine Learning - Specialty	Link	Only need 1 of 3
Cloud - ML	Google Professional Machine Learning Engineer	Link	Only need 1 of 3
Cloud - ML	Azure Data Scientist Associate	Link	Only need 1 of 3

Portfolio Project Checklist

We recommend you use GitHub Pages and blogdown to host your protfolio as shown in the course. Recommended minimal list of hosted pojects:

1. 2x MS Excel dashboards - hosted as webpages
2. 1x Tableau Public dashboard
3. 1x Tableau Public story
4. 2x EDA of a dataset using RMarkdown - published on Kaggle as well
5. 2x EDA of a dataset and ML modeel development using Python - published on Kaggle as well
6. 1x deploy an ML model to the clouding using AWS (or similar) EC2 and a docker container

The course walks you through or gives resources needed to complete each of these. Make sure you use novel datasets in your portfolio! If you only use the data from the course it will be very similar to everyone else...

Next Steps

If you have completed the certification checklist, built a resume and hosted project protfolio you are ready to start work! The next step in your learning journy should be to decide which of the job types you want to dive deeper into. Here are the recommended next learning resources for each:

1. Data Analyst - Work to become one of the Tableau Visionaries
2. General Data Scientist - Create and publish an R Package to CRAN
3. Machine Learning Engineer - Complete the fast.ai course 'Deep learning for coders'
4. Everyone - compete in a competition on Kaggle

Latest News