Usability participant searched "UVA", got poor results.

[maryannbrody]

This only came up once - but I saw a user search for "UVA" and get very irrelevant results. (definitely not anything for the university of Virginia!)

• This is a data issue, but just wanted to make sure it was known.

[rickleegit]

This is a known issue. Thanks for pointing it out from another perspective though. The ability to effect this type of change, en masse, will be enabled by the inclusion of the new VA-Certs API which will subsume the functionality of the current WEAMs public system. The API is scheduled for release in Jan 2017. Until then, EDU claims to not have the resources to cleanse the data appropriately. For reference the DB contains ~16,000 schools and more than double that number in OJT institutions. All of these are in various states of data disrepair. I will forward the UVA issue to the team, for their awareness and again encourage them to engage some level of data cleansing.

cc: @emilyville @staceylanger

[saintsoup52]

There is an easier solution to this issue.

ED in their IPEDS database collects school aliases that would include common variations like UVA, MIT and UCLA. Conveniently we already ingest the IPEDS data file that contains the alias.

On CFPB's comparison tool we allow users to search for the official school name and the school's alias when someone is selecting a school.

This is a real issue... one of the most common searched school names is MIT.

Additionally, I am pretty certain that VA Cert will never collect school aliases

[rickleegit]

@staceylanger,

Patrick's suggestion, above, is possible and an easier solution that manual data cleansing but would require additional development enhancements to both the CT application and data service.

Based on our current product strategy and resourcing, this seems like it would need to be added to a backlog of potential future enhancements for the time being.

If/when we choose to engage this enhancement, we would want to answer, at a minimum, the following questions:

• How many aliases per school do we permit? (i.e. Penn State, PSU, etc)
• How do we handle the hierarchical relationships with main and branch campuses? (i.e PSU - Altoona, Penn State - Altoona)

The following application services will be effected:

• Data Service - a field needs to be added at a minimum, possibly a new table and ingestion functionality.
• CT application - the application will also need to be updated to accept and parse this new information for each record, including the search functionality which will need to be adjusted to index this new data point.

[saintsoup52]

The way IPEDS does this is they put all the aliases in one field, separated by pipes. See below: "UCSF |UC San Francisco |University of California at San Francisco |University of California, San Francisco"

The search just looks for that text appearing anywhere in that field or the school's official name field, much like if someone where to type a random search word in our search box e.g., "massage".

When a user types in "UCSF" on College Navigator they are returned the official name of the school "University of California-San Francisco"

Our folks at CFPB have this worked out if you want to see their coding.

[rickleegit]

Moving to Backlog.