I don't usually post on non-patent or other IP matters, but I'm making an exception for a valuable search engine that should be used when it's data rather than words that you are looking for.
It can be very hard to find data such as pie graphs, bar graphs and the like. Putting in words into Google so often means that you simply get lots of pages of varying quality and relevance which you have to work through. Putting in extra words, especially in the Advanced format, by the way, often does help.
I now routinely use the free site Zanran when I want to gather statistics on subjects such as poverty in London, or patents by companies in telecommunications. Just run the cursor down the PDF or Excel icons and an image from the document appears. It's a simple and easy way to identify useful documents. You can add limits as well, such as saying that the the site must be less than 2 years old, or that the site must come from a certain country (I would suggest that they add other countries to their list to increase usage).
I first heard about Zanran a few years ago when one of their development people asked if he could have a coffee with me to see if I had any thoughts about how an early version could be improved. There was an interesting discussion and some changes, they kindly told me, were incorporated as a result. I don't know how exactly the search engine works -- I'm guessing that it looks for captions or words around graphs and tables.
The importance of filtering the vast amounts of information on the Web so that you quickly find good material becomes ever more crucial. Zanran fills an important hole in that effort.