CAS_ADS

This project is maintained by mariezufferey

Module 5: consolidations and electives

[Back to main page]

Electives

Python: Python for programmers - Data science with Python and Pandas - Data analytics with NumPy and Pandas - Deeper dive into Python - Statistics with Python - Scalable analytics with Python (DASK) - Image processing with Python - Machine Learning with Python

Machine learning: Machine Learning with Python - Machine learning with Dataiku - Mathematical methods for data Science and machine Learning - Machine Learning review - Deep networks - Deep learning: a practical approach in MATLAB

Other: Introduction to MATLAB - LaTeX: introduction and further type setting techniques - Working with containers - Advanced HPC topics

Python for programmers

[Back to top]

Data science with Python and Pandas

[Back to top]

Data analytics with NumPy and Pandas

[Back to top]

Deeper dive into Python

[Back to top]

  1. Beyond notebooks
    • Python (Jupyter) notebooks - What are notebooks good at? - What are notebooks bad at? - What is needed to run Python code? - How to run python code? - Setting up VSC Live Share
  2. Objects and scope
    • Everything is objects - Object identity - The mystery of small integers - Object type - Object value - Mutable vs. immutable objects - Accessing objects
    • Scopes (local, non-local, global, builtin) - Scope gotchas
    • What can you do with objects: attributes (or properties) - method - standard functions and operations
  3. Classes (doc)
    • What are classes? (instance) - Class syntax - Simplest class - Non-declared goods
    • __dict__ property to Store instance properties; __slots__ to enforce a specific set of properties
    • Adding a method - Aside: f-strings (python >= 3.6) - Class methods (1st argument automatically receives the reference to the instance itself)
    • self to store data - hasattr(object, name) to check for attributes - __init__ to always have data and meaningful constructor
    • Aside: “dunder” or “magic” methods (fixed names for objects interfacing with specific language functions: surrounded by double underscores; do not create your own new names of this format !)
    • __str__(self) used whenever call print on the object
    • Class variables (property shared by all instances)
    • Parameter defaults gotchas - Class variables gotchas
    • Usage counters - Example: inventory system - Container class - Implementation details
    • Private attributes (convention: _)
    • Implementing the container - Implementing an iterator - Implementing other things
    • Class inheritance (super() to access the base class method if overriden)
    • Multiple inheritance
    • Aside: getters (functions returning a value as a property; @property) and setters (function that receives the value when a property is written to; @<funcname>.setter)
  4. Functional programming
    • “Pure” functions: do not depend on external or internal state; no “side effects”
    • Functional programming and immutability (best way: use immutable data structures)
    • Aside: shallow and deepcopies (<object>.copy(); for potentially mutable objects, a recursive “deep” copy is needed: deepcopy(<obj>) from the copy module)
    • “Higher-order” functions: functions that accept functions as parameters (in Python, functions are treated as values)
    • Standard primitives for functional programming: some built-in, some in the functools module (e.g. reduce)
    • Generators and lazy evaluation (e.g. map returns a generator (= an iterable which is evaluated one value at a time wheniterating), an example of lazy evaluation)
  5. Decorators
    • e.g. @mydec on the line before myfunc() equivalent to mydec(myfunc)
    • Dealing with arguments and return values - Decorator parameters
    • A practical example
    • Stacking decorators (the function is wrapped in the inverse order of decorators)
    • Decorators and function identity (after being wrapped by a decorator, the name of the function changes; if not desirable, use from functools import wraps)
    • Tracing recursive calls
    • Memoization (optimize the code by caching intermediate results)
  6. Modules and packages
    • What are modules and packages? (modules = mechanism for sharing code across multiple Python source files; packages = a way to organize multiple modules together in a tree-like structure of submodules)
    • Importing code (import and star-import; this later is discouraged as it modifies the global namespace in an unpredictable way)
    • Module example
    • What happens on import? - What else happens on import? (__pycache__ folder)
    • __name__: at runtime, contains the name of the current module; set to __main__ if executed directly
    • Where does Python search for modules? sys.path from sys module (directory containing the input script or current directory; PYTHONPATH environment variable; the installation-dependent default), can be modified at runtime
    • Packages (organized in subfolder; when published, contain additional metadata)
  7. Python environments
    • a specific Pythonversion plus a collection of extra packages
    • What problem do environments solve? (working with different package versions) - What’s needed for a virtual environment? (a package manager to install packages in the environment (e.g. pip ); a virtual environment manager (e.g. virtualenv )
    • Virtual environment workflow: creating, activating, installing dependencies, running code - Virtual environment example
    • Tricky parts of Python package management (non-python code and libraries; pip supports a binary format called “wheels”)
    • What about (Ana)conda? (conda = package / virtual environment manager part of the Anaconda distribution; can manage Python packages; has more support for binary packages)
Statistics with Python

[Back to top]

Scalable analytics with Python (DASK)

[Back to top]

Image processing with Python

[Back to top]

Machine Learning with Python

[Back to top]

Machine learning with Dataiku

[Back to top]

Mathematical methods for data science and machine learning

[Back to top]

Machine learning review

[Back to top]

Deep networks

[Back to top]

Introduction to MATLAB

[Back to top]

Deep learning: a practical approach in MATLAB

[Back to top]

LaTeX: introduction and further type setting techniques

[Back to top]

Working with containers

[Back to top]

Advanced HPC topics

[Back to top]

Contact

Questions, comments or suggestions ? Don’t hesitate to contact me !