First, JavaScript and modern Java (Spring Boot etc.): To me those are still the most important technologies in B2B software.
Docker and microservices-related patterns and technologies in general: While not the silver bullet they're sometimes depicted to be microservices have some interesting characteristics beyond scalability alone.
Finally, Blockchain-related technologies (and decentralisation at large): There are interesting potential real-world applications beyond finance (supply chain management, for instance).
(re)learning modern javascript properly with a focus on writing desktop like applications and data visualization.
Take my machine learning knowledge from the realm of the statistical/theoretical and toy problems to the realm of solving real world problems for paying customers.
I also want to get more into making things and doing things like wood working, CNC work and interacting with the real world via electronics and microcontrollers, but that is purely for fun.
I'very always been a software guy but recently I've been getting into hardware recently, particularly with the raspberry pi. I built a barbell tilt sensor using berryIMU and raspberry pi to track my squat technique and analyze it.
For the coming year I'm going to learn more in the IoT space, particularly getting devices to talk to each other and using machine learning on the firehose of data.
For a small, personal project I'm working on: I'll learn about XMPP, ejabberd, firebase.
For another small personal project: visualizations (D3.js, especially treemaps). I'll take data from an ecommerce website and display something like what MIT's Observatory of Economic Complexity does to give a big picture of what is selling on that website (categories, products, etc).
My initial idea was for people buying and selling cars having to go through each posting. Instead of doing that, it would be cool to show a scatter plot of postings which would make it easier to detect a good deal.. I'll probably add features like setting alerts for certain parameters and receiving an SMS/email notification if a car matching that shows up.
The site doesn't have an API, so I'll have to get the data by crawling it (Scrapy and Beautifulsoup).. They do use schema.org/Product, though. So it'll simplify things. What would complicate is that many people don't know how to use currency properly (we use cents in usual parlance, and the unit in formal postings. Many can't do the transition).
Other than that, I'm learning React and looking into options to do a custom app for a friend using Bootstrap's dashboard stuff and React instead of Odoo which he finds cumbersome (he has a very small business).
I've been looking at Learning Management Systems because it'd be cool to propose a course in the local language for high-schoolers here according to the national program. Most don't speak English. I'm hesitating between this and doing something like Duolingo but for maths where you could solve problems by writing equations which would be interpreted and evaluated by the system (thus looking into SymPy).
I'm currently unemployed and there are no jobs so I have time, but it's structure and discipline that is the hardest to achieve. Maybe I could use the React/Bootstrap dash/Electron to make custom apps for local businesses. If I had a bit of money, I'd really want to experiment with Hydroponics.
I'm also in contact with local model airplane enthusiasts and looking into ways to streamline the process of creation (build an app they'd use to make it easier for them to make airplanes, maybe add simulations).
Docker and microservices-related patterns and technologies in general: While not the silver bullet they're sometimes depicted to be microservices have some interesting characteristics beyond scalability alone.
Finally, Blockchain-related technologies (and decentralisation at large): There are interesting potential real-world applications beyond finance (supply chain management, for instance).