MOOCs and e-learning are in the midst of an unexpected (if entirely logical) resurgence. At the same time, data science — famously described by the the Harvard Business Review as “the sexiest job of the 21st century — remains an attractive career-change option among those with a mind for stats. And finally, in the seemingly endless Python vs. R debate, Python appears to have put some clear distance between itself and its “rival” language.
All of which goes to say: There’s never been a better time to enroll in a Python-focused data science course. It’s where newcomers can grasp important statistical concepts, learn about popular machine learning algorithms and get exposed to must-know libraries — all through the lens of what is now, according to some rankings, the most popular programming language in use.
With that in mind, below are 21 Python data science bootcamps and courses to know and consider.
Note: Bootcamps that had offered in-person learning have since shifted to fully online/remote models. Check sites for updates.
This well-ranked, five-course intensive, available via Coursera, starts off with an introduction to basic programming before diving into dataviz plotting and charting, various machine learning methods, text mining and social network analysis — with each course focused on Python. Like most Coursera specializations, all courses are also available à la carte.
Cost: $49/month subscription
Massachusetts Institute of Technology computer science professors and lecturers dive into all the common models and machine-learning algorithms that working data scientists commonly utilize during this 15-week, edX-hosted course, covering everything from collaborative filtering in recommender systems to natural-language-processing RNNs. For projects, students design sentiment analysis classifiers (as used in product review studies) and neural networks for classifying handwritten digits. Basic understanding of Python programming and probability stats are prerequisites. You can bone up on both through MIT’s e-learning options, here and here.
Cost: Free/$300 with certificate
Students start this seven-course intensive (each course is three weeks) with foundational Python code structure and end using scikit-learn to build their own predictive models. In between, courses span visualization, data interpretation and data manipulation, including exposure to popular plotting packages, like Matplotlib, and must-know Python libraries for data analysis (NumPy and Pandas). Expect to devote between three and five hours of coursework each week.
Cost: $3,600
Most real-world data work is a lot more workaday than, say, fancy computer vision models or deep-learning neural networks. This eight-month, five-course deep dive, hosted on edX, covers how management analysts can get practical performance insights by learning to apply and evaluate predictive and statistical models with Python. Some familiarity with statistics and programming is required.
Cost: $1,350
CS50, a collection of Harvard University computer-science courses available via EdX, was a MOOC watershed when it debuted in 2012, and its reputation and popularity has only solidified since. “[F]ew, if any [digital university courses], combine the institutional credibility, the enormous reach, and the zealous engagement” that the CS50 collection — which includes this AI-with-Python intro — carries, according to a recent New Yorker profile. Why so credible? It’s hard! In fact, most don’t finish. But “those who stick with it often become diehards,” as the profile noted.
Cost: Free/$199 with certification
Emeritus partners with 15 different universities to bring higher-ed shine to the e-learning galaxy, including its Columbia University Engineering-affiliated data science sequence. It also offers this two-month stepping stone, covering need-to-knows like cleaning, plotting, visualizing and analyzing data with Python and its Pandas library. Video lectures (124 in total) are complemented by live, guided webinars. Students should expect to spend four to six hours per week on coursework. The course also includes career guidance sessions and access to 30-plus practice data sets.
Cost: $900
Metis is among the more renowned names in the burgeoning data-science bootcamp ecosystem, and its ACCET-accredited intensive has produced graduates that landed jobs at Spotify, Facebook and other tech powerhouses. But true newbies should also consider this six-week, Python-focused prep course. It lays all the bedrocks — linear algebra, statistics, calculus — while also introducing Jupyter Notebooks and all the foundational Python libraries. Metis also offers the three-week Python for Beginners, which focuses more on strict programming, less on data-science concepts.
Cost: $17,000
Similar to Metis, BrainStation offers both a full-time, instructor-led data science intensive and more narrowly focused, part-time online prep. Among them is this data-centric crash course in Python. Students cover essentials like Pandas and NumPy, how to read documentation and how to use metadata repositories.
Cost: $14,500
Related18 Free Data Sets for Learning New Data Science Skills
This 12-week immersive exposes students to advanced concepts and tools like clustering, neural network fitting and recommender systems, but it kicks off with a Python-basics bootcamp-within-a-bootcamp of sorts. The self-paced section includes essential programming concepts while also providing exposure to fundamental concepts like probability distributions and hypothesis testing.
Cost: $15,950
The data-science and machine-learning offering from this long-running bootcamp provider kicks off with a module devoted to Python fundamentals. There, you’ll hit on Jupyter Notebooks and basic libraries (Pandas, NumPy, Matplotlib/Seabo), along with key concepts like data structures, data cleaning, relational databases and tips for scraping for data and retrieving data with APIs. The module is available part-time (10 months) and full-time (five months). Physical campuses are located in Austin, Chicago, Denver, Houston, New York, San Francisco, Seattle and Washington, D.C. — for when in-person learning returns.
Cost: $15,000
This 13-week Python-based intensive claims a high graduate salary and placement rate at high-profile tech firms, with multi-quarter career services support. As for the curriculum, students move from Python and statistics fundamentals to machine learning and predictive models to NLP and recommendation algorithms, wrapped up with a capstone project. A part-time version is expected to launch in the fall, and there’s also currently a dedicated Python Fundamentals option for the greenest of the green. Campuses are located in Austin, Boulder, Denver, Los Angeles, New York City, Phoenix, San Francisco, Seattle and San Jose — for once in-person teaching resumes.
Cost: $17,980
This “bundle” offers a discounted rate for three multi-unit courses. The first focuses on syntax basics, list manipulation and data wrangling; the second dives deeper into analysis (with NumPy, SciPy and Pandas) and introduces visualization (with Matplotlib and Seaborn); and wraps up with a 20-hour course on machine learning algorithms. Each course is also available as a standalone option.
Cost: $4,732
A promising choice for those interested in a part-time route that also offers hands-on guidance, this five-week option introduces a new Python library and corresponding concept each virtual meeting: linear algebra and stats with NumPy, dataviz with Matplotlib and Seaborn, exploratory data analysis with Pandas, machine learning with scikit-learn and presentation with Bokeh. The course is taught by Isaac Faber, chief data scientist at the U.S. Army AI Task Force.
Cost: $2,795
Related14 Best Data Science Books to Read Right Now
There’s an abundance of information packed across this 88-hour, 23-course Python-focused plunge — from data cleaning to cluster analysis to supervised and unsupervised learning. (DataCamp also offers a similar track for R.) But don’t be overwhelmed; the affordable subscription model and self-paced video structure allows students to sidestep cram-a-lot intensity. The first chapter in each of the nearly two-dozen courses is free to view.
Cost: $25/month subscription
Dataquest’s track operates on a similar model as DataCamp, with a plethora of courses (nearly three dozen) that students can choose between at will with a yearlong subscription. Expect Python-focused looks at data cleaning, Kaggle competitions and even a bit of engineering, among much more. There’s a notable focus on real-world portfolio-building, with projects that include stock and home-sale predictions, bias analysis based on SAT scores and demographics, and a Bayesian approach to spam filtering.
Cost: $24.50/month subscription
This popular, well-rated sequence — made available through Coursera and taught by IBM data scientists — covers data analysis, visualization, machine learning and more — each through the prism of Python. It culminates with the so-called Battle of the Neighborhoods capstone, in which students retrieve and scrape data, then create a model to find housing that meets given criteria, locate the ideal spot to open a particular kind of business, or solve some other location optimization challenge.
Cost: $39/month subscription
This high-rated Udemy course covers the Python basics; outlines key statistical concepts, such as linear regression, logistic regression and bias-variance tradeoff; and introduces more complex systems like natural language processing, recommendation engines and neural networks. (There’s also a few hours devoted to TensorFlow and Keras in the advanced back end.) This course is one of the more well-known, respected options among the online video-lecture courses.
Cost: $109.99
Developers tend to have strong feelings about programming certifications in general — let your demonstration projects and open-source contributions speak for themselves, goes the refrain — and Python is no exception. But whether or not you opt to take the exam for Python Institute’s top-level certification (perhaps the most prominent one available), the non-profit does offer two free training modules worth considering. The emphasis is more on programming, rather than data science, but it’s a good primer of fundamentals regardless.
Cost: Free
INE’s “pass” gives students access to 18 courses, ranging from novice to professional, from quick-hit intros (the 39-minute Exceptions With Python) to in-depth immersives (seven-plus-hour Deep Learning With Python). All are taught in a Python-focused framework, with topics spanning data analysis, programming, data management, machine learning and automation.
Cost: $49/month
This four-day tutorial is a recorded, on-demand version of a recent live-online bootcamp session. Offered in partnership with Data Society, the course starts with an overview of the core responsibilities of data scientists, then introduces NumPy and Pandas applications, including how to summarize and reshape data. Attendees also get a 60-day trial of Skillsoft’s learning platform, which includes more data science- and Python-related video lessons plus several practice labs.
Cost: Free
After an intro to SQL, students dive into analysis, visualization, statistics, data cleaning and munging, and web scraping — each using Python and its corresponding libraries. From there: machine learning-driven techniques like recommender systems and NLP. There’s an explicit focus on portfolio-building and projects, which include a linear regression model of the collapsing honeybee population and plotting line graphs with lime sales data. Students complete 26 modules across an estimated 35-week schedule. Shorter, targeted “skill paths,” rather than “career,” are also available, including several Python data science-related options.
Cost: Available with pro subscription ($19.99/month)
Related4 Principles to Inspire a Truly Data-Driven Culture
