anvaka

I think the book is truly awesome. So I've made a visualization to see where it stands in Amazon’s “also-bought” graph.

I used Amazon’s product advertising api to fetch the graph.

Visualization was created with my own algorithm that tries to embed a graph into city-like grid. It routes paths between nodes, so that paths that were taken before are more preferable for new paths.

The thickness of a road gives rough estimate of how many paths went there.

Please let me know if you have any suggestions!

I tried to exclude copyright lines as much as I could. I used "license markers" for that, but I might have missed something.

Here is more information about it: https://github.com/anvaka/common-words#how

Thanks! The data comes form GitHub snapshot, stored on BigQuery. Here is more details about it: https://github.com/anvaka/common-words#how

The copyrights are filtered out, because indeed there was a lot of them.

It uses it as long as there is enough space. Once it fails to find a rectangle to fit a new word, it tries to reduce the size of the word.

In general, word clouds are bad for comparing sizes. For that reason I used plain list in the sidebar on the left (or at the bottom if you are on mobile)

Thank you :)

Thanks!

I'm using file extension to differentiate between extensions.

You can request other languages here: https://github.com/anvaka/common-words/issues/4

As long as language's extension is unique, I think I can make a visualization of it.

Added: https://anvaka.github.io/common-words/#?lang=hs

Unfortunately the symbols will not show up, because I'm ignoring them: https://github.com/anvaka/common-words/blob/master/data-extr...

I've being collecting number of dependencies for the most popular packages last month. This visualization shows the growth rate of those packages.

Make sure to drop an mp3 file on this visualization to see an Easter egg :).

this is done now.

They are not standardized. However if you'd like to render your graphs in this tool refer to this doc: https://github.com/anvaka/pm#your-own-graphs

I observed people using the tool before it was bound to the scroll. Almost everybody used scroll in attempt to fly forward, while WASD navigation wasn't obvious. So I decided to show help on scroll...

Will bind it to h and ? as well

Thanks!

This is just a hobby project, which I've built in my spare time: http://anvaka.github.io/allnpmviz3d/

It gives you a spaceship to explore the universe of npm, the largest package manager for javascript.

Use WASD keys to move around. Or if you have a smartphone - rotate your device around to control the camera.

The deprecated yeoman package was huge: http://npm.anvaka.com/#/view/2d/yeoman

Thank you :)!

I used to be worried by amount of dependencies too, when I was transitioning from .NET stack to js. Until I realized that each module is usually very small, doing only one thing, and most of the time doing it well.

Now I'm thinking of them in terms of code reuse. Code reuse is good

Very interesting idea!

What would be a connection between two maintainers? If their packages depend on each other, or if they co-maintain a package?

Thank you! The size in bytes or something else?

Use mouse wheel to zoom :)

I'm working on graph algorithms and this is all work in progress. This example shows how to render graphs from both node.js and a browser, using the same code. Would really love your feedback

Yep, this is it :)

I thought GitHub is really missing this feature, so as a hobby project I built this.

The source code is here: https://github.com/anvaka/gazer/

Looking for your feedback :)!

It's all open source: https://github.com/anvaka/jsruntime . Allows you to find objects by their property names/values in runtime. It crawls the entire reachable javascript objects graph. So be careful in what you are searching - might take a while.

It might be slightly more convenient than a regular code search, since it finds dynamic functions too and let you set breakpoints there. Hope you find it useful too. Link to plugin in chrome store: https://chrome.google.com/webstore/detail/js-runtime-inspect...

Some time ago I showed visualization of reddit in parts, using SVG.

I add WebGL support to my library (https://github.com/anvaka/VivaGraphJS ) and wanted to visualize full graph of indexed reddit.

A blog post describing how it works: http://blog.yasiv.com/2012/07/visualizing-communities-of-red...

Would love to hear what you think...

Ported this guy from Flash (http://wonderfl.net/c/7au5 ) to JavaScript. Microsoft is really missing out with WebGL support, but I know why people are saying that plugins are dying...

I keep working on my hobby. This time I visualized Facebook. Every line is friendship, and every image is your friend.

It might be slow for huge graphs (more than 2K), but hey it's open-source - please join me - https://github.com/anvaka/VivaGraphJS and help make it better :). Would love your feedback, hackers.

Thank you. These matrices cover wide spectrum of domains (such as structural engineering, computational fluid dynamics, model reduction, electromagnetics and so forth). They are available as a public data set on Amazon cloud: http://aws.amazon.com/datasets/2379?_encoding=UTF8&jiveR...

As for not connected dots - it might be a bug in my transformation code, but more likely they are just isolated components of the graph. E.g. http://www.yasiv.com/graphs#HB/bcsstm26 - big bang :)

I was bored last weekend, decided to create a demo of my graph drawing library (https://github.com/anvaka/VivaGraphJS ) and visualize collection of sparse matrices of University of Florida. I added WebGL renderers to the library recently and wanted to test its performance.

If your browser has disabled WebGL the site will fall back to SVG renderer, but it's much slower. It's best viewed in Chrome or Firefox. If you still unable to see visualization I made a video for you: http://www.youtube.com/watch?v=L5KrtGB8oA4

Would love to know what you think :)

Thank you for the reference. I read they are using Barnes-Hut optimization for larger graphs. Do you know what kind of optimization?

This is interesting and challenging problem. Interesting because it always leads to unexpected discoveries. Challenging because it's hard to show large clusters of friends in accessible manner.

LinkedIn has InMaps app ( http://inmaps.linkedinlabs.com/ ) which shows your network. It's cool, but really hard to read it...