[box style="rounded" border="full"]Quick Links: This post is the "Why" and background on my motivation for this project. Separate posts breakdown the implementation. Links coming as published.

• Part 0: Why Guitar Tab? (this post, genius)
• Part 1: Text Tab to "RiffCode"
• Part 2: Riff Storage and Querying in ElasticSearch
• Part 3: Simple UI display with PHP-FatFree and Twitter Bootstrap

[/box]

What is it?

A fun data/hack project that I've been working on for the past few weeks. Basically put...

Riffbank is a "reverse search engine" for guitar tabs.

Give it a simple section of guitar tab (a "riff"), and it will tell you what songs it could be (plus any other metadata it knows). Give it a try and let me know! [ still a work in progress, being a small side-project ]

Why?

Anyone who has seen my crazy shit on this blog has probably stumbled upon my 'Enter Sandman Tableau 7 viz', Basically, I set out to see if I could process guitar tab in such a way so it could be re-displayed in an analytical visualization engine like Tableau, and become a bit of a "tab learning app". Obviously this was a proof of concept and just a cool hack to use Tableau for - but I never completely stopped thinking about the concept of using guitar tab acsii files as a "source of data". [ some riffJSON action ] If you've learned guitar at some point in life and been into tech and computers - you've probably used text-file based guitar tablature before. Hell, I can remember using it back in the early 90's for god sake. I would print it out on a dot-matrix printer and keep a mess of it in my guitar case... Ah, memories. Anyways, it's been around for a long time - and there is NO shortage of it on the internet these days. While writing the sandman viz, I remember thinking "could I do this with a whole shitload of tab files - and if so, what insights (or not) could I gain from that kind 'data' repository?". [the riffbank elasticsearch cluster] More to come...