Index page data awesome

[Swaagie]

Some ideas that came to mind, might not be all doable and workable. Some are simpler and more straightforward than others.

• geographic location of all mirros in one map (static mostly),
• moving average (n=3) of fullfat changes feed, e.g. changes per 5 minutes, 12px/hour
• topic map based on title/keywords/readme content, parsing with tokenizer/stemmer of some sort, combination of http://bl.ocks.org/mbostock/5944371 and http://www.jasondavies.com/coffee-wheel/
• graph holding collaborating connections between authors, quickly discover who works with whom on what, might need a dedicated neo4j graph database.
• For fun: tree structure showing dependency tree, see http://prcweb.co.uk/lab/d3-tree/, auto-completer to quickly change displayed module.

Layout could be vertical, box per data chart. Max horizontal space, easy to add or remove.

[3rd-Eden]

• The status of the replica, a lot of the current replica's are out of date or broken. So we should probably show if it's up, down or behind replication.