"The First Node" of Your Data Place (Live Updated)

-

What I want - 6th of June, 2022

This is an article describing my progress in developing a data platform first for myself, but then for others. It will first address my common individual issues as a data engineer and pipeline manager, but also the groups I work with and the future work I end up doing. This platform is (so far) built on the following principles:

  • Code is the most expressive way of managing data.
  • Python is one of the easiest languages to express ones self with (at this time).
  • An online platform should have little to no compromises to runtime execution.
From that, we will build a data platform.

What I want - 7th of July, 2022 (31 days later)

After writing a lot of content on Your Data Place and publishing a few ideas, I've come to want a few more things out of the platform. These are not feature requests, but rather experience requests for how I'd like to operate this platform myself. See below:

  • A platform where I can write-out something I'm learning, then easily import it into old/new projects. For me, this is specifically for machine learning. I've got that many old half-baked ML project folders that sit somewhere on multiple of my PCs, that go from linear regression to attempting to use Keras for image classification - but as time goes on, I forget them. One of my goals this year will be to start learning Keras from the ground-up again (using this handy book I brought a while back) and make sure the Your Data Place runtime can cope with ML.
  • A Python SDK (module) to integrate cloud runtimes with local runtimes by using data. Using this logic, I'd be able to write local scripts that leverage (keyword) my YDP runtimes to qucikly experiment in new projects, if I don't have the time to setup a new YDP project.

Comments

Post a Comment

What did you think about this post? Do you have any ideas you think would be helpful?

Popular posts from this blog

Shareable data and code in YDP Projects

-
Image
Datasets, everywhere... One major ability I want in YDP is the ability to create datasets that can be interpreted as global, reproducable datasets which were created as part of a dataset generator. Although I presume datasets will primarily be used inside their own projects, I think there's great value in the ability to make these datasets global, and accessible by any other project. The Problem At the moment, the problem is in the implementation. This was no issue when YDP was a sandbox test project in my local filesystem, where I could easily WRITE a file here, then READ a file here under the same name. But, in a real production architecture, we're using Docker as of this moment to manage our independent filesystems. When it comes to being able to have a dataset be generated in Project 1's file system (FS) then wants to be used in Project 2's FS, what has to occur is something along the lines of the following: Flag a Python script as a generator for a dataset. Execute
Read more

Siesta of YDP

-
Unfortunately, I've had to take a siesta in Your Data Place. I'm still extremely passionate, but most of my spare time has been consumed by a half-work half-play project which is niche in my industry which will allow me some more future freedom if it ends up working correctly. I'm using some of the innovations I've outlined here in that project, to make it something completely unique and separate to the current competitors. It's in the pharmaceutical space, and I may start to think through concepts in this blog, as some of the ideas are quite abstract but need to be represented simply and easily to the users. In the time since I've last written, I have not learnt Rust to a proficcient level. I've learnt Firebase and become better at frontend web design as well as better at software application building using Python. I will return to Rust, and YDP, as these are my raw passions. I believe my experience in this next project will greatly inspire more ideas in YD
Read more

The start: a week of Rust

-
Image
I've taken a small break to focus on my primary work obligations, and in that time I've had lots of time to ponder. During this time, I flew to Tasmania (737-800) and read through a few preliminary chapters of Manning Publications'  "Math for Programmers"  (Paul Orland). This book is a real workhorse for programmers who want and know  they wish to be able to implement more mathematics into their code, but find the conventional teaching methods difficult to fast-track. Home now, I plan to spend a week giving Rust a real go from the other Manning Publications book, "Rust in Action"  (Tim McNamara). In this time, I'll learn everything with a noble aspiration: "How could this benefit the users of YDP?" (as well as my primary work, and other interests of mine like genetics). Rust is fast, but from what I've experienced, it's cracked and goated with the sauce. What I find interesting about Rust, is the author's intent is as clear as d
Read more