Preparing your environment
This blog post series will focus on how to set up ones environment for doing work in Azure Machine Learning (Azure ML) using Python 3.7.3 and Jupyter notebooks for design and execution. Python is aided by the Anaconda distribution and we employ packages such as Pandas, Numpy, MatplotLib, and SciKit-Learn. This blog will focus on discussion points as well as explanatory anecdotes, the main knowledge source is in the form of a series of Videos as well as Jupyter notebooks which can be found on the accompanying GitHub repo. This blog post is Part 1 of a series
- Part 1: Setting up your Azure ML Local Environment Github | Video. Release Date: November 4th 2019 – Last Updated: N/A
- Part 2: Creating an Azure ML Environment and connecting with Python to your local environment Github | Video Release Date: N/A – Last Updated: N/A
- Part 3: Connecting to an Existing Azure ML Environment and conducting an Experiment Github | Video Release Date: N/A – Last Updated: N/A
- Part 4: Getting to know NumPy Github | Video
- Part 5: Getting to know Pandas Github | Video
- Part 6: Prediction or Classification – That is the Question Github | Video
- Part 7: Using the AutoML SDK Github | Video
- Part 8: Doing your own Algorithmic work Github | Video
- Part 9: Where to go next
The approach for this blog series
Logistically, the way I am approaching this is to “show by doing”; by recording my actions as I go through the process, I am working in real-time, this will include possibly fumbles, gotchas, and hopefully workaround and solutions. One of the result of the action especially when working in Python through Jupyter notebooks will include both markdown, code, and outputs, the video will show this in realtime and the GitHub repos will show it as a consumable asset that you can take and use in your own environment.
NB: I will obfuscate elemnents of the video and redact certain items where sensitive information and configuration files are exposed.
The main point here is that the Blog post is the Legend, the Video is the How-To and the Github Repo is the leave behind. This may also find its way into in person sessions and as such I will be treating the GitHub repo as a living document, so what you see today, may change over its lifetime.
What this IS and what this is NOT
This blog post series and all its accompaniments is my take on one approach of potentially many in what is a multi discipline technology. My background is that of a software engineer and solutions architect, with an emphasis for a very long time on the SharePoint Object Model a subset of the .Net Framework, within that ecosystem I specialized in back end and middleware development. I made a switch to Mobile and Cloud development back in 2014 still focusing on getting data from point a to point b, and of recent I am now focused on Intelligent Process Automation (IPA) which is a confluence of Digital Process Automation [aka workflows], Machine Learning and Artificial Intelligence. I say all that to say this, I am “not” a Data Scientist, I am “not” a Business Intelligence Analyst, I can write software in the C family but I am relatively new to Python, R, SaS. Packages such as Numpy, Pandas, etc is also new. What I do here ‘may’ work, but it may not be the best practice way of doing it, do not put what you see here in production.
What this IS, is a primer for folks like myself who have a natural inquisitiveness for learning something new, want to bridge a gap between what you know and where the technology is heading. As I plan to treat this as a living document, as I find newer and better ways of doing something I will update the posts while keeping the original so one can see the origin. In GitHub you can always see past commits for that.
I would like to treat this series as a conversation and I hope that we can all learn from each other here. Id like to ask the following:
- If you want to have a conversation about a topic in the bog post or video [especially video as I won’t be monitoring YouTube comments] use the comments here
- If you find something in code that needs revisiting, open a GitHub issue
Thank you, Cheers.