Week 1/2: Plunging headfirst into the program

Because of the Finals schedule at Oregon State University, two weeks worth of work was rolled into one. The background readings on the Columbia River and APNA, familiarization with current Matlab codes, and writing my own code, which was the agenda for week one and two was done in about four work days.  But it wasn't all work this week.  I had the opportunity to go to to Astoria with my mentor, Melissa Gilbert and graduate student Leslie Slasor to drop off equipment and pick up the APNA (Autonomous Profiling Nutrient Analyzer).  On the way back, we went on a hike up Saddleback Mountain.  At the top was a relaxing panoramic view of the Pacific Ocean, Columbia River Estuary and the Coastal Range.
Before I continue, I should probably explain what I'm doing here.
My purpose at CMOP is to improve the current Matlab data processing codes for analyzing output from an instrument called the APNA (Autonomous Profiling Nutrient Analyzer).  It a step forward in that it (theoretically) allows for the continuous, near real-time monitoring of biogeochemical variables in the Columbia River Estuary.  Right now the APNA measures Nitrate, Nitrite, Phosphorus, Silicate, Iron II and Ammonia at hourly intervals.  Because new technology is inherently touchy, deployments have been limited to about three week intervals.  Even in those limited deployments, large amounts of data have been gathered.  But most of that data  is in its raw form.  The current Matlab code works decently, but is plagued by the presences of "bubbles" in the data.  These appear as random spikes in the data.  My main task this week was to improve the current "debubble" program.  I used a moving average and standard deviation of n points around a certain point in the data to calculate limits of data to keep and or to reject.  After hours of debugging, I got a rudimentary version of this running.  I plan to improve the speed and efficiency of this program next week.  Hopefully, at the end of the summer, the APNA will be autonomous when it comes to collecting useful data.