Network Science by Albert-László Barabási

Having close personal connection to the lab for a few years - He is very successful, very charismatic, and a consummate politician above all. He holds tons of sway over Northeastern admin (lucky to have such a big fish in their little pond), and of course at his Hungarian university he's basically god. Certain policies simply do not apply to him.

Anyway, in the lab there is very strong pressure for things to look right. Some students and postdocs play ball, and enjoy great success in their careers. Some get a little squicked out (use of funds, and the science, seemed to be possible areas of concern), wrap things up quietly and remain on good terms. One or two have major concerns and just kinda disappear. In addition to the treatment of foreigners and management style issues characteristic of academia (which tbh causes more anguish than any of the sketchier stuff).

There are of course a lot of other lazslo affiliated business ventures - startups, books, etc. Right now and for the past few years his big focus has been on producing fine art which can be pretty lucrative, and also .... A good medium for the things he likes to do with money/influence.

Anyway, very nice and charismatic guy, really successful, objectively speaking I admire him, but his name is also used as verb/improper noun in our household if that makes sense

Interesting critique post, but can you elaborate why you linked the wikipedia bio of a popular scientific podcaster, who apparently has nothing to do with this debate, other than a family connection to an author cited near the end of the post?

I would counter that many complex phenomena are best represented by graphs. There's a sort of elbow in the learning curve where you have a lot of tools but as your datasets grow in size the graphs become unmanageable and the signal:noise ratio goes way down. At this point you need to prune or chunk somehow, but it rapidly becomes more of an art than a science and reproducibility is elusive. Advanced techniques disseminate slowly and are not that intuitive, but there's still steady progress, eg deep learning models for sparsification while preserving structure.

Yeah, journals are too shy to post these criticisms. What the hell is the use of a paper if there is no public peer review?

It’s hard to say which book is better, as it depends a lot on what you’re looking for. Menczer’s book does a great job of introducing you to the basic concepts of network theory from a hands on perspective with Python code available through a companion GitHub repo. Barabasi’s book goes into a lot more depth and gives a bit more of an historical perspective of how things developed, but it is significantly more theoretical/mathematical in its presentation.

Full Disclosure I was a co-author of Menczer's, collaborated with his group for several years, and know all the authors personally. I met and spoke with Barabasi multiple times over the years but never worked with him directly.

<ShamelessPlug> you might also want to checkout my substack Graphs for Data Science: https://graphs4sci.substack.com/ where in each post I look at algorithm, implement it in Python, and explore it in a toy problem before showing you how to applying it to a real world dataset. </ShamelessPlug>