Join GitHub today
GitHub is home to over 28 million developers working together to host and review code, manage projects, and build software together.Sign up
When choosing facets for a Staged exhibit, it is important to understand that each facet value is transferred over the network every time a facet selection is made. This makes the uniqueness qualities of the data very important in determining how well your exhibit will perform.
For example, in a 100K item exhibit with a faceted field with 30K unique values, the data transfer for this field alone will take 100x the time to perform as a facet with just 300 unique values. If those field values are of considerable length (more than a few characters) - such as a list of organization names - the introduced latency could be on the order of seconds on some networks or on slower client machines (due to parsing/processing latency). When considered for all the facets in your exhibit, the issue can quickly compound into a significant usability problem.
For these reasons, facets should be chosen wisely, sticking to those fields with relatively low numbers of unique, short values. It also becomes more important for your exhibit, that the data be "clean" in the sense that identical facet values appear in the data as identical strings, i.e. that insignificant whitespace is removed, and capitalization and encodings normalized. Consider using a tool such as Google Refine for these purposes.