Before upgrading my current site to use Umbraco 8, I decided to revisit my experiences of moving sites from version 6 to version 7. When doing that I dabbled with the idea of starting with a brand new site and migrating the content separately. Here I take a look at applying that method to version 8 and from that my first Umbraco Package was born.
Before making a site live I tend to perform some tests to make the site as secure as possible. This usually involves using a tool like Qualys Web Application Scanning or OWASP ZAP to probe the site and then address any problems. In this entry I detail some of the more common changes that I make.
One of my pet projects is to be able to scan in images of my old diary and convert the written words to text. As this is a bit too ambitious for an early machine learning project, I decided to first perform a more basic analysis on the images: classify them into left and right pages.
In this post I detail how I set up my code so that it automatically updates Tumblr when changes are made to the content within Umbraco. When I update a Cinema Cento post on the CMS, it now gets created or updated on my Tumblr blog.
In this entry I explain how I set up OAuth authentication for the Tumblr API. Eventually I want to be able to write posts to the API, but for now I explain how to retrieve user information as that also requires the extra steps.
I have been keeping a blog on Tumblr for a number of years, and I was keen to replicate and (eventually) post to it from my Umbraco website. This post describes how I went about using the Tumblr API to retrieve the current posts on the blog and re-create them as content within Umbraco.
As my machine learning studying has progressed I have now learnt about TensorFlow, a Python Library that provides lots of functionality for implementing neural networks. I decided to apply it to the Titanic data I have used previously.