Recently I upgraded my monitor from a Dell P2319H to a Samsung UR550 (LU28R55) to introduce myself into the 4K world. My current setup consists into a Mac Mini and company-issued Dell Latitude both through HDMI 2.0+. And it was thanks to the Windows machine that I started to notice the screen looked way better (color, sharpness) there than when I was using the Mac. Took me sometime, but I was able to fix it.

During the last 5 months since my last post, Brazil kept working poorly to control COVID-19 spread througout the contry but - helped by virus seasonality and some counter measures (masks, distancing) - we slowly go out the worse moment of the first wave. However, things are changing again. And not for better.

During last week WWDC, it was presented a simplified lifecycle scheme for SwiftUI apps. During the presentation if was showed how to create a new app using the new lifecycle – which is pretty easy – but how about updating an existing app? I’ve been working in a project for a new iOS app, so let’s try to update it to the new lifecycle.

Besides the actual data and the historical trends, I wanted to have some kind of prediction too and a natural choice was to look into Tensorflow framework which is built for Python and provide a lot of handy tools and algorithms. In this post I’ll talk a bit about the technical stuff and some analysis from the results.

I really got excited working with Streamlit, so I decided to add a trend chart using some of the curves I usually use when analyzing the stock market. As expected, it was a piece of cake and the only thing I had to learn a bit was using Pandas to create moving averages.

Recently I run into this article https://medium.com/@ahmasoh/the-end-of-flask-in-data-science-738920090ed9 in my Medium’s Daily Digest and gave it a try as a professional Flask developer. For my surprise, it seems that Flask is often considered for building Data Science/Machine Learning reports – though Flask is good for web applications, I have a feeling it is too much for whom is not a web developer, Jupyter seems a better fit here.

It appears the latest Xcode build 11E608c (20-May-2020) introduced a new behavior (or a bug) for property observers (didSet, willSet). Before the update, I had no problem in my project in have a function for a didSet observer of an object that updated its own value.

Being inserted into Apple’s environment for a while, I always felt kinda obligated to start developing for either iOS or MacOS. Though there was one problem: Objective-C. The language and XCode seemed to have a high learning curve, and as I was still on college and more focused on my research, I never dig further on it. Then, in 2014 Swift was announced and the future looked brighter.

I was writing a little bot in Python to automate some build handling – we have large builds that are stored remotely and, as our network is not as fast as it needed, the bot downloads them as they get available – and I wanted to integrate it with Microsoft Teams. One easy way available is to configure a team channel to receive e-mails, so I needed to send them from the Python bot.

In 2014 I tried to start a blog using GitHub pages but it lasted only one post. Now I’m trying bring it back to life, but the old code wasn’t work with new libraries.