Interesting analysis! I've been dabbling in the space for some time now and I feel like the data analysis and data visualization aspects of AI are often overlooked in Java.
The tooling is not well consolidated, for instance there are at least 5 different Java Jupyter kernel competing and no charting library that comes close to what exists in julia, python and R. It is similar for DataFrame libraries, even though tablesaw seems to dominate. The AI developments are also a matter of community and it feels like the ones who build and innovate in AI simply don't find the tooling they need to create and experiment in Java.
I've recently been building an alternative to Streamlit for Java that I think is a missing piece in the Java ecosystem: https://github.com/jeamlit/jeamlit. Quarkus and Spring are not designed for "data work" per se and require frontend work for anything that touches data viz. Curious what you think of such tools and the ecosystem in general in the context of quick experimentation, data viz and building data apps that consume AI / AI results, and that are easy to use for AI builders.
Hey Artur, there's a subtle typo "simplicfication", thanks :)
Interesting analysis! I've been dabbling in the space for some time now and I feel like the data analysis and data visualization aspects of AI are often overlooked in Java.
The tooling is not well consolidated, for instance there are at least 5 different Java Jupyter kernel competing and no charting library that comes close to what exists in julia, python and R. It is similar for DataFrame libraries, even though tablesaw seems to dominate. The AI developments are also a matter of community and it feels like the ones who build and innovate in AI simply don't find the tooling they need to create and experiment in Java.
I've recently been building an alternative to Streamlit for Java that I think is a missing piece in the Java ecosystem: https://github.com/jeamlit/jeamlit. Quarkus and Spring are not designed for "data work" per se and require frontend work for anything that touches data viz. Curious what you think of such tools and the ecosystem in general in the context of quick experimentation, data viz and building data apps that consume AI / AI results, and that are easy to use for AI builders.
Hey Cyril - thank you for mention 😁 I for sure include it in the "The Rest of the Month" edition
Doesn't Embabel deserve a few words in this context?
It will! But I overall left for later topic of high-level agentic tools 😊
Embabel, Brook, Kooq... a lot is happening in the space
As always excellent post! I got my hands on Bob couple days ago and it’s pretty cool!