Basic HTML Version
© 2014 Pacific Crest
245
locate information to help me try out a
data transformation in a specific data
analysis software package
P
lan
How to complete the activity
1. Review the different types of data transformations (in the Techniques).
2. Each member of the team should prepare to present/explain a different data transformation technique.
3. Review the Model.
4. Answer the Critical Thinking Questions using the model and the data transformations.
5. Complete the remainder of this activity (from Demonstrate Your Understanding through Assessing
Your Performance) on your own, or as directed by your instructor.
M
odel(s)
Exemplars and representations
S
imulating
P
oint
-S
ource
E
ffluent
L
oading
I
mpacts
T
o
T
he
B
eaufort
R
iver
DataSetsandDataPreparation:
Thedatausedforanalysis
and modeling consisted of continuous (1-hour interval)
tidal and water-quality data, daily total precipitation data,
and weekly effluent data. In 1999, BJWSA, in cooperation
with the USGS, established a network of seven gauging
stations (fig. 1) on the Beaufort River that monitor water
level (WL), water temperature (WT), specific conductance
(SC), and dissolved Oxygen (DO). Three of the stations
also record tidal stream flow. Precipitation data were
obtained from the National Weather Service and two of the
WRFs. Effluent data (sampled once a week) consisting of
5-day biochemical oxygen demand (BOD5) and ammonia
(NH3) also were obtained from the WRFs.
Two calculated variables were derived — tidal range
(XWL) and DO deficit (DOD). Tidal range is an important
variable for determining the flushing dynamics of the
tidal rivers. Tidal range, calculated from water level, is
defined as the water level at high tide minus the water
level at low tide for each semi-diurnal tidal cycle. The
DOD is the measure of the difference between actual DO
measurement and DO for fully saturated conditions. The
DOD was computed using an algorithm that assumes a
constant barometric pressure over the data collection
period (USGS, 1981). The DOD was adjusted for salinity.
From: TRANSFORMING LARGE DATABASES INTO CRITICAL KNOWLEDGE USING DATA MINING– THREE CASE
STUDIES IN SOUTH CAROLINA AND GEORGIA (source document available online) by Paul Conrads (USGS SC Water Science
Center Stephenson Center, Suite 129 Columbia, SC 29036) and Edwin A. Roehl Jr. (Advanced Data Mining, Greer, SC)
5.4 Transforming Data