testing a processing algorithm that takes in data from twitter [tweets] and analyzes the text strings [written my nate smith and me].  the result is an approximation of a flesch reading ease test roughly based off the content.  limiting rules: from the 100 most recent picks, take those that have 15 words or more [smaller groupings of words really skew the test results], dont count certain words [@**, #**, http:**].  note: lower test results are supposed to indicate a higher reading level.
search: "sustainable"  test average: 52

search: "dog" test average: 88

search: "hungry" test average: 80

 search: "judicious" test average: 65