prempv

__we should not forget that this stacks many steps with implicit errors on top of each other__ - No words truer have been spoken.

I've had the same experience as well. I suspect if it's due to large presence of HTML in the training data as part of codebases and online content

Making the visualization a web-app is the fastest approach we could think of for prototyping. What other options are there for an interactive visualization? Quite curious..

I'm surprised by the amount of skepticism I'm noticing here. I have been following Suhail's work on Mighty for a few months and I was very much looking forward to it.

One use case that where I think this would make my life easier is in the big-data / ML space where I'm trying to visualize large quantities of data. JS, WebGL and other supporting tools are all available today, but it's quite painful to load a graph visualization with 1M nodes and make it responsive without spending a lot of time optimizing the JS code. As a data scientist when I'm simply hacking stuff and want a quick prototype it's nearly impossible.

Graphistry [https://www.graphistry.com/] has a decent setup for graphs viz, but it didn't quite fit my needs. I've also tried JS running on a large machine (with GPU) and VNCed to it. That experience was quite poor.