How do I count occurrences of unique field values?

@nmpeterson's answer is a good one. If you don't want to create a table you can get this information as a SearchCursor. You can do it by making use of a dictionary to tally counts:

import arcpy

fc = r"D:\AVI_DAVID\zevel\zevel.gdb\Export_Output_miv_cur"
field = "USAGE"

#Create dictionary to store unique values
CountDi = {}

with arcpy.da.SearchCursor (fc, field) as cursor:
    for row in cursor:
        if not row[0] in CountDi.keys():
            CountDi[row[0]] = 1
        else:
            CountDi[row[0]] += 1

for key in CountDi.keys():
    print str(key) + ":", CountDi[key], "features"

An alternative would be using lists and sets:

import arcpy

fc = r"D:\AVI_DAVID\zevel\zevel.gdb\Export_Output_miv_cur"
field = "USAGE"

Occurances = []
with arcpy.da.SearchCursor (fc, field) as cursor:
    for row in cursor:
        Occurances.append (row[0])

for i in set(Occurances):
    icount = Occurances.count(i)
    print str(i) + ":", icount, "features"

Many ways to skin a python. Note that this makes use of the data analysis search cursor, which requires ArcGIS 10.1 or newer.

There is a tool in ArcGIS called "Frequency" (arcpy.Frequency_analysis()) that allows to to count the number of occurrences of each unique value in a specific field (or unique combinations of values in multiple fields). It will create a new table containing the original field name(s) with a row for each unique value/combination, and another "Frequency" field containing the number of rows in the original feature class with that particular value.

In your case, the arcpy call would look like this:

arcpy.Frequency_analysis(fc, PATH_TO_OUTPUT_TABLE, ["USAGE"])