A moment ago I received an email from my publisher saying my “Angular Services” book just published. As always it feels great. But I have to say this one was different. For this book I have created an open source code repository and called it “The Sherlock Project“. I used that project to explore the Angular framework concepts with some AI twists.
I was busy with my Machine Learning certificate (Uni. of Washington) at the time Google was wrestling with Angular release candidates (RC1 & RC2). At the same time I was dreaming about an idea. I was thinking: How can I use “Machine Learning” algorithms in Google’s Angular framework and create services to do some freaky stuff.
It was challenging
At that time it sounded interesting, I was excited to get up in the morning and create a bunch of TypeScript classes to:
- collect data from news agencies (CNN, BBC, Fox and so on)
- fetch hot keywords from Google trends in the past 24 hours
- dig into my private gold mine of collected news – which was persisted to Google’s Firebase Realtime Database – and look for insights
For example I wanted to:
- find if an allegation is true (Trump’s sex allegations was one of my case studies)
- predict the outcome of a decision (Living on Mars and the SpaceX project was/still is one of my favorites)
- and at the end I wanted to use data visualization techniques to show the results
It wasn’t easy all the way. In fact at some stages I had no idea how to find a solution for an impossible looking challenge. On top of that Google developers did a great job to publish breaking changes every 3~4 weeks. Can you believe that? They published 8 release candidates in 6 months before releasing the first Angular stable version. It was a pain in the neck. Every time a new RC was out, I had to make major changes in my project and I blame Google for that. Kidding! 🙂 You guys rock! Yes those regular updates were annoying but every time I upgraded my code accordingly, I have noticed how safer and faster it became. For that reason I thank you all.