Fair use is the copyright exception with which people are often most familiar. Whether a use is "fair" depends on the facts of a particular case. Four factors must be evaluated when such decisions are made. The first factor is the purpose and character of the use. Among the considerations is whether the use is for commercial or for nonprofit educational purposes. Works that transform the original by adding new creative authorship are more likely to be considered fair use than those that do not; however, even a reproduction can be considered a fair use in some circumstances. The second factor is the nature of the copyrighted work. The scope of fair use is generally broader for fact-based works than it is for fanciful works, and broader for published works than for unpublished ones. The third fair use factor is the amount and substantiality of the portion used. Generally, the more that is taken, the less likely it is to be fair use, but there are situations in which making complete copies is considered fair. The fourth factor is the effect on the potential market for or value of the copyrighted work. A use that supplants the market for the original is unlikely to qualify as fair.
From "Copyright Issues Relevant to the Creation of a Digital Archive: A Preliminary Assessment", by June Besek
So you can take facts from one source (as long as it's just excerpts, not the whole thing) and republish them (as long as you add valuable commentary) for non-profit public use (as long as the market value of the original source is not affected). A major point of contention will be whether reproducing part of a work detracts from the value of the original source (as readers may feel they don't need the whole work), or improves the value of the original source (by publicising the work and improving their reputation). Also, does transforming data from one form to another count as adding new creative authorship?