I've been working on book search at the Internet Archive, and I've been using Wikipedia article titles and redirects as entities and synonyms. I wanted to build autocomplete for this gizmo, so I downloaded 7 days of pageviews for the en Wikipedia, and wrote a tiny script to sum them up. It worked great!
Here's the demo (currently live, will disappear eventually). "number" is the pageviews count.
curl http://researcher3.fnf.archive.org:8080/autocomplete?q=Que | json_pp { "autocomplete" : [ { "number" : 68310, "label" : "Queen Victoria" }, { "number" : 53283, "label" : "Quentin Tarantino" }, { "number" : 29192, "label" : "Quebec" }, { "number" : 23717, "label" : "Queen Elizabeth The Queen Mother" }, { "number" : 20500, "label" : "Quetiapine" } ] }