Denny and Kyle recently interned at HumanGeo. This blog post represents the culmination of their efforts. If you're interested in interning or joining our team, let us know!Here at HumanGeo we make sense of large amounts of unstructured text data, but one of the constant challenges is being able to extract the key text details we're interested in across disparate data sources. An initial, low-effort approach is to create a keyword matcher that searches the document for words in various dictionaries, but there reaches a point where the keyword list gets too long and that process becomes too costly. Once one reaches this point, the method of attack needs to shift to a more powerful, more hands-off solution - Named Entity Recognition.