Name Recognition
Often schools share similar names and/or have shortened nicknames used in common parlance. For example Polytechnic is often shortened to “Poly,” resulting in 36 possible U.S. schools to choose from if used as the search criteria. “JFK” for John F. Kennedy is another example. There are about 32 John F. Kennedy high schools in the U.S.
Asking an applicant to choose their school out of 36, 32, or even more than a handful of possible choices isn’t a friendly experience. And as we’ve seen in our user tests, anything that gets in a student’s way impedes their ability to complete the application.
Additionally, today’s users are accustomed to their favorite search engines which make intelligent suggestions sometimes as quickly as typing only the first few characters of what they are looking for in spite of misspellings or common usage. It was our goal to provide the same experience for CEEB search, and our technology stack made it possible.
Elasticsearch
Element451 uses the ELK stack to analyze some of our statistics. ELK stands for Elasticsearch, Logstash, and Kibana. We applied the ELK stack to CEEB in order to make search results more relevant.
Elasticsearch is an extremely fast database that excels at search. (We’ll discuss the other two items in the stack in future posts.)
Elasticsearch brings the ability to search text, correlations, geo-points and more in big-data. If you didn’t catch that, read it again: GEO-POINTS! App451 and other Element451 modules are geo-aware. This means we can use an algorithm to suggest the closest schools to the physical location of a student that matching their search criteria.