Difference between revisions of "User:Streeter8"

From REU@MU
Jump to: navigation, search
(Weekly Progress Report)
 
m (1 revision)
 
(No difference)

Latest revision as of 06:25, 12 June 2015

Jackson Streeter

Personal Information

Summer 2013

Readings

Research Objectives, Weekly Log, Weekly Progress Report

Week 1

Research Objectives

  • Collect and begin initial readings
  • Begin learning and applying MATLAB
  • Meet GasDay team members and find suitable workspace
  • Observe Thesis Outline Form

Weekly Log

  • Tuesday, May 28
    • Orientation
    • Marquette University Campus Tour
    • Pre-Experience Survey
  • Wednesday, May 29
    • GasDay Orientation
    • Electronic Research Orientation
    • Meet with Dr. Corliss
      • Project Orientation and Initial Readings Assigned
    • Reading:
      • Test-Driven Development, Writing Scientific Software: A Guide to Good Style, An Introduction to Mathematical Modeling
  • Thursday, May 30
    • Practice MATLAB
    • GasDay Seminar
      • Where to Find GasDay Data Part I
    • Good Research Practices and the Art of Keeping Logs Seminar
      • Complete Week 1 REU Environmental Satisfaction Survey
    • GasDay Seminar
      • Where to Find GasDay Data Part II
    • Reading:
      • Test-Driven Development, Writing Scientific Software: A Guide to Good Style
  • Friday, May 31
    • Practice MATLAB
      • Mathematics Fundamentals
      • Creating Functions
      • Plotting Functions
      • Creating Script Files
        • Using Inputs and Menus
    • Reading:
      • A Concise Introduction to MATLAB
      • Flatland
  • Saturday, June 1
    • GasDay 20th Anniversary Celebration
  • Sunday, June 2
    • Reading:
      • Writing Scientific Software: A Guide to Good Style

Weekly Progress Report

  • Tasks Completed
    • Began learning MATLAB
    • Acquainted myself with Marquette University, the GasDay Lab, the GasDay team, Milwaukee, and my fellow REU students
    • Began introductory reading
  • Readings Completed
    • Test-Driven Development, Preface, Ch 1-10
    • Writing Scientific Software: A Guide to Good Style, Preface, Ch 1-2.3, 3.1-3.2
    • An Introduction to Mathematical Modeling, Preface
    • A Concise Introduction to MATLAB, Ch 1
    • Flatland, Introduction, Ch 1-2
  • Tasks for Next Week
    • "Master" MATLAB
    • Devise an approach to tackle the research problem
    • Continue my assigned readings
    • See what can be done about the missing furniture and other missing items that were supposed to be in our room
    • Inquire about the reimbursement for the parking pass
    • Remain positive, and continue to build my knowledge and skill base
  • Present Obstacles
    • I have no idea what to do
      • By that, I mean that I have much uncertainty about my problem. I cannot currently write well in MATLAB. The way to obtain the data I need seems quite difficult, and any mistake seems like it could leave serious damage on old data. And many of the assigned readings discuss difficult topics, namely the assigned papers (as opposed to the books.)
      • In light of these difficulties, I am not discouraged. Rather, I am encouraged, because I recognize that I am doing work that is significant. Also, this may be the most difficult academic problem that has ever been posed to me. An answer to this question would be one of my greatest academic achievements to date.
      • In summary, my problem is intimidating, but I am not afraid of being challenged.
    • The readings are difficult
      • When I struggle with readings, my mind tends to wander and I lose focus on what I am trying to read

Week 2

Research Objectives

  • "Master" MATLAB
  • Devise an approach to tackle the research problem
  • Continue my assigned readings

Weekly Log

  • Monday, June 3
    • Practice MATLAB, a lot
      • Installed MATLAB on personal laptop
      • Converted Alphametic solver written in Java to MATLAB
    • Meet with Dr. Corliss
      • Initial approach to problem
      • MATLAB as a programming language
      • Nature of Computation Notebook
    • Reading:
      • A Concise Introduction to MATLAB
      • Flatland
  • Tuesday, June 4
    • Practice MATLAB
      • Writing Functions
    • LaTeX and Source Citation Seminar by Dr. Corliss
    • Reading:
      • A Concise Introduction to MATLAB
      • Writing Scientific Software: A Guide to Good Style
      • Fitting a Continuous Profile to Hourly Natural Gas Flow Data
  • Wednesday, June 5
    • Accessed GasDay Wiki for first time
    • Extracted Data for Analysis for first time
    • Practice MATLAB
      • Overlay Plots of Temperature
      • Overlay Plots of "Error"
    • Meet with Dr. Corliss
      • GasDay Acronyms
      • What is to be plotted
      • How to find the data
      • GasDay Wiki
      • Graduate Thesis form for end of program paper
  • Thursday, June 6
    • GasDay Camp
      • GasDay Overview - Professor Thomas Quinn
      • GasDay Professional Etiquette - Dr. Ron Brown
      • New Product Initiative - Nathan Winninger, Nick Wilson
      • Where are we going: Research - Dr. Ron Brown, Dr. Richard Povinelli
    • Working Lunch
      • Cool/Difficult games created by programers with too much time
      • Differences in grading styles between universities
      • Completed Week 2 REU Environmental Satisfaction Survey
  • Friday, June 7
    • GasDay Camp
      • The Natural Gas Industry - Dr. Ron Brown
      • GasDay Workflow - James Lubow
      • Using GasDay - James Lubow
      • GasDay Models - Dr. Ron Brown
    • Meet with Dr. Corliss
      • Theses and journal articles
      • Model building and techniques
      • Project Presentation for GasDay Lab
      • End-of-program report
    • Brewers Baseball Game!!
  • Saturday, June 8
    • Installed TeXShop on personal laptop
  • Sunday, June 9
    • Draft project paper outline

Weekly Progress Report

  • Tasks Completed
    • GasDay Camp
    • Learned what I needed to know about how to write and run MATLAB
    • Rough Rough Draft Outline for paper
  • Readings Completed
    • A Concise Introduction to MATLAB, Ch 2-7
    • Writing Scientific Software: A Guide to Good Style, Ch 3.3-4.3
    • Flatland, Ch 3-4
  • Tasks for Next Week
    • Plot MLR, ANN, GasDay forecast, and SUM Errors
    • Discuss GasDay forecast with Dr. Brown and Jim
    • Presentation for GasDay Lab over my research Topic
    • Rough Draft of Introduction for my paper
    • Implement SUM Error into GasDay forecast
  • Present Obstacles
    • Still getting acclimated to Milwaukee
      • There were lots of people coming in and leaving Marquette and the GasDay lab
    • GasDay Camp was very interesting, but it took time and mental energy
    • Dr. Brown is very busy, so Jim and I may need to pester him a little aggressively to get to talk to him sooner, rather than later
    • The intimidation of working on something bigger than me
      • I am confident that I can complete my research, but there is a level of abstractness that is still a little overwhelming

Week 3

Research Objectives

  • Plot MLR, ANN, GasDay forecast, and SUM Errors
  • Discuss GasDay forecast with Dr. Brown and Jim
  • Presentation for GasDay Lab over my research Topic
  • Rough Draft of Introduction for my paper
  • Implement SUM Error into GasDay forecast

Weekly Log

  • Monday, June 10
    • Plot MLR, ANN, GasDay forecast and actual for Day 0
    • Meet with Dr. Corliss
      • Current progress on plotting
        • Improving plots
      • Interpreting hourly data
      • Extracting wanted data
      • MATLAB processes and commands
    • Drew out the hourly forecast data
      • Found holes in forecast
  • Tuesday, June 11
    • Meeting with Dr. Brown, Dr. Corliss, and Jim about GasDay forecast
    • Good Presentations, Good Technical Writing, and the Difference Between Them Seminar, Dr. Brylow
    • Developed hour cleaning, data cutting, and overlay plot methods
  • Wednesday, June 12
    • Issues with hourly data
    • Meet with Dr. Corliss
      • GasDay forecast
      • Issues in Hourly
      • Project Outline
    • Put together presentation for GasDay
    • Developed improved cleaning, and pulling methods for data
    • Reading:
      • Flatland
  • Thursday, June 13
    • GasDay Seminars
      • Presented GasDay != GasHour
      • GasDay Notation Issues, Tian Gao
      • Research Projects within GasDay, Dr. Brown
    • Working Lunch
      • Example Research Presentation
        • (i,j)-Step Competition Graphs, Dr. Factor
    • Updated documentation in my functions and scripts
    • Generalized Plot Function
      • Time, Forecast date
  • Friday, June 14
    • Developed improved cleaning function
    • Discussed GasDay forecast with Dr. Brown and Jim
    • Created Ratio, Ln Ratio, Error Ratio, and Ln Error Ratio Plots of GasDay forecast and Sum of Hourlys
    • Analyzed conditional and unconditional probabilities for errors of GasDay forecast and Sum of Hourlys

Weekly Progress Report

  • Tasks Completed
    • Presented GasDay Seminar
    • Found out how to run Why Analysis and GasDay forecast
  • Readings Completed
    • Flatland, Ch 5
  • Tasks for Next Week
    • Plot GasDay forecast and Sum of Hourlys Error
    • Study GasDay forecast Code
  • Present Obstacles
    • Nervous about talking to customer representatives
    • Complexity of GasDay forecast

Week 4

Research Objectives

  • Plot GasDay forecast and Sum of Hourlys Error
    • Ratio
    • Ln of Ratio
  • Study GasDay forecast Code
    • Discover where Sum of Hourlys can be used as input
    • Implement Hourly actuals as input

Weekly Log

  • Monday, June 17
    • Meet with Dr. Corliss
      • Probability Tables
      • Application of Bayesian Techniques
      • Markov Chain/Process
      • Error Plotting
      • Clustering
      • Discussion with customer
    • Developed a Markovesque Chain for probability of Error State for 2 days in future
  • Tuesday, June 18
    • Worked out errors in extraction
    • Study GasDay forecast
  • Wednesday, June 19
    • GasDay Data Security Seminar
    • Meet with Dr. Corliss
      • RMSE of GasDay and Gas Hour
      • Markov Process
      • 3D Graphs
      • Complexity of GasDay forecast
    • Improved Markov Process
    • Plotted Error to flow
    • Study GasDay forecast
  • Thursday, June 20
    • What's Important to Forecasting Seminar
    • Working Lunch
      • Example Presentation: Embedded Systems, Dr. Brylow
      • Completed Week 4 REU Environmental Satisfaction Survey
    • Meet with Dr. Corliss
      • 3D Plotting
    • Study GasDay forecast
  • Friday, June 21
    • Study GasDay forecast
    • Meet with Dr. Corliss
      • Progress in GasDay forecast

Weekly Progress Report

  • Tasks Completed
    • Mastered Database Extraction, Why Analysis, and GasDay forecast execution
    • Took Markov Processes as far as currently appropriate
    • Researched and edited GasDay forecast code, finding potential areas for where Hourly data can be entered
  • Readings Completed
    • Flatland, Ch 6
  • Tasks for Next Week
    • Implement Sum of Hourly Forecasts as an input to the GasDay forecast
    • Graph Error of GasDay forecast and Sum of Hourlies against flow, temperature, date
    • REU Presentation
    • Talk to Tom about talking to customer
  • Present Obstacles
    • GasDay forecast Code is complex and not well documented
    • Cannot quite get the 3D Plots to look how I want
    • Seminars, lunch breaks, and other meetings tend to break-up periods of focus
    • Going through the code, I often lose focus because I find myself going in a circle or the complexity of the code cause me to think of what else I could do that is easier.

Week 5

Research Objectives

  • Implement Sum of Hourly Forecasts as an input to the GasDay forecast
  • Graph Error of GasDay forecast and Sum of Hourlies against flow, temperature, date
  • REU Presentation
  • Prepare for talk with customer

Weekly Log

  • Monday, June 24
    • Meet with Dr. Corliss
      • 3D plot coloring
      • Preparation for discussion with customer
      • Preparation for REU presentation
    • GasDay Data Security Follow-Up Seminar
    • Worked on REU Presentation
    • MATLAB Test Driven Development
    • Created 3D plots modeling Hourly/GasDay forecast Error vs. different attributes
      • Such as temperature, wind speed, etc.
  • Tuesday, June 25
    • Study GasDay forecast
      • "Mastered" GasDay forecast
  • Wednesday, June 26
    • Meet with Dr. Corliss
      • Hourly actuals and daily actuals discrepancy?
      • Observed 3D plots
        • Linear Relationships
      • GasDay forecast Code
        • Day shift?
      • Discuss GasDay forecast with Jim
    • Create REU Presentation
  • Thursday, June 27
    • Refine Presentation
    • Mini Presentations
    • Work on implementing hourly data as replacement GasDay forecast input
  • Friday, June 28
    • Meet with Dr. Corliss
      • REU Presentation
      • Final Paper
      • Basic Increment Plots
    • Created basic increment plots

Weekly Progress Report

  • Tasks Completed
    • "Matered" 3D plot coloring and creation
    • "Mastered" GasDay forecast
      • Implemented Sum of Hourly Forecasts as an input to the GasDay forecast
    • REU Mini Presentations
    • Created basic increment plots
  • Readings Completed
  • Tasks for Next Week
    • Cubic Splines of incremented data
    • Begin rough draft of final paper
    • Regression Analysis of Hourly/GasDay forecast Error vs. different attributes
  • Present Obstacles
    • The GasDay forecast is very complex. I spent a lot of time studying it, taking away time that I could spend completing other objectives.
    • Staying Focused
      • The lab is an awesome place to work, but can get distracting. Birthdays, lunches, and friendly conversations are often moments that take longer than planned, and taking time away from research. Though well and good, these moments do slow down progress.
      • My own focus. When I come across something that is difficult or that I do not know how to do, I will find train of thought wandering into other ideas and projects that seem easier or more fun. Reeling myself back in and focusing can be hard at times.

Week 6

Research Objectives

  • Cubic Splines of incremented data
  • Begin rough draft of final paper
  • Regression Analysis of Hourly/GasDay forecast Error vs. different attributes
  • Prepare for talk with customer?

Weekly Log

  • Monday, July 1
    • Meet with Dr. Corliss
      • Plots in preparation for cubic spline distribution of error
      • How to create cubic splines in MATLAB
    • Discuss GasDay forecast with Jim
    • Histogram of hour defects
  • Tuesday, July 2
    • Meet with Dr. Corliss
      • Creating cubic splines in MATLAB
      • fconmin command
      • Cubic Splines
    • What Makes a Good Poster Seminar by Dr. Magiera
  • Wednesday, July 3
    • Meet with Dr. Corliss
      • Creating cubic splines in MATLAB
      • fconmin
      • Writing
      • What I'm doing the next week and a half
  • Thursday, July 4
    • Independence Day
  • Friday, July 5
    • Practice and Worked in LaTex
    • Began Text Outline and placing current word in document
    • Worked on Saving relevant graphs in a usable format for LaTex

Weekly Progress Report

  • Tasks Completed
    • Broke hourly adjustments into a cubic spline
    • Began processing paper in LaTeX
  • Readings Completed
    • Modern Control Theory, 6.8
    • Probability and Statistical Inference, 6.7-6.8
  • Tasks for Next Week
    • Make significant progress on paper
    • Start on Poster
    • Start on GasDay Presentation
  • Present Obstacles
    • Cubic Splines appear no more accurate than naive, flat distribution of error
    • I have never used LaTeX before

Week 7

Research Objectives

  • Make significant progress on paper
  • Start on Poster
  • Start on GasDay Presentation

Weekly Log

  • Monday, July 8
    • GasDay LDC Meeting
    • Writing Final Paper
    • Preparing Graphs for paper
  • Tuesday, July 9
    • Continue Working on Paper
    • Continue looking at Cubic Spline code
      • There is an error in here somewhere
  • Wednesday, July 10
    • Continue Working on Paper
    • Continue looking at Cubic Spline code
      • Discovered error
  • Thursday, July 11
    • GasDay Seminars
      • Paul Kaefer
      • Greg
    • Working Lunch
      • Why You Should Apply to Graduate School Discussion - Dr. Brylow
    • Continue looking at Cubic Spline code
      • Discovered that what I thought was the error was in fact not the error
      • Discovered what was actually the error
    • Added a weighted naive error distribution.
  • Friday, July 12
    • Continued writing paper
      • Created and imported many figures/tables
    • Created new method of hourly incrementation with linear/step transitions.

Weekly Progress Report

  • Tasks Completed
    • Completed Cubic Spline forecast fit
    • Created weighted naive forecast fit
    • Began naive with linear transitions fit
  • Readings Completed
    • Principles of Forecasting: A Handbook for Researchers and Practitioners, Sections over Combining Forecasts
  • Tasks for Next Week
    • Improve Naive method transitions
    • Prepare and give presentation of work at GasDay Seminar
      • Begin working on Final REU presentation
    • Continue writing paper
    • Begin working on poster
  • Present Obstacles
    • Some of the topics I am looking over, such as ANN, Cubic Splines, and Markov Chains, are very detailed, if not at least difficult.
    • MATLAB, like any language has a learning curve, and I'm still learning how to use all of its tricks.
    • LaTeX also has a learning curve. I have never used it before, so I am copying and pasting lots of code, not really sure how it all is working quite yet.
    • Making sure that my graphs have all necessary details, while hiding proprietary information is tedious, but necessary.

Week 8

Research Objectives

    • Improve Naive method transitions
    • Prepare and give presentation of work at GasDay Seminar
      • Begin working on Final REU presentation
    • Continue writing paper
    • Begin working on poster

Weekly Log

  • Monday, July 15
    • Implemented linear transitions into the naive method.
  • Tuesday, July 16
    • Meet with Dr. Corliss
      • Reference Texts
      • Final Paper
      • Adjustment of final 10 hours
      • Artificial Neural Networks
      • Improving Graphs
    • Improved Graphs
  • Wednesday, July 17
    • Meet with Dr. Corliss
      • Troubleshooting Techniques
      • Writing Exceptions
    • Attended Dr. Corliss Lecture at UWM
    • Finished Presentation for GasDay Seminar
    • I've been dealing with a date/forecast issue that I just can't seem to fix for a long-term database extract
  • Thursday, July 18
    • GasDay Seminar Presentation
      • Presented on GasDay, GasHour Inequality
      • Snow Cover Relations to Gas Demand
        • Spencer
    • Meet with Dr. Corliss
      • Presentation/Suggestions
    • Working Lunch
  • Friday, July 19
    • Developed long term error analysis
    • Meet with Dr. Corliss
      • Poster Presentation
      • After REU
      • Paper

Weekly Progress Report

  • Tasks Completed
    • Implemented linear transitions into the naive method.
    • GasDay Seminar Presentation
  • Readings Completed
  • Tasks for Next Week
    • Complete Poster
    • Put in Significant Work on Paper
    • Begin Final Presentation
    • Develop Long Error Analysis for each Hour
  • Present Obstacles
    • I've been putting a lot of work into developing my research, and not as much time writing it up.
      • More time needs to be dedicated to write up.
      • Time is short, so I cannot expect to get to much more done.

Week 9

Research Objectives

  • Complete Poster
  • Put in Significant Work on Paper
  • Begin Final Presentation
  • Develop Long Error Analysis for each Hour

Weekly Log

  • Monday, July 22
    • Meet with Dr. Corliss
      • Poster
      • Importance of Paper
    • Edited Poster
    • Writing Paper
  • Tuesday, July 23
    • Poster
    • Paper
    • Developed long-term error analysis of hourly forecast adjustments
  • Wednesday, July 24
    • Meet with Dr. Corliss
      • Poster
      • Paper
    • Electronic Version of Poster Turned In
  • Thursday, July 25
    • Edited Paper
    • Working Lunch
  • Friday, July 26
    • Edited Paper
    • Recalculation of RMSEs
  • Sunday, July 28
    • Edited paper
    • Wrote final presentation abstract

Weekly Progress Report

  • Tasks Completed
    • Completed poster for poster session
    • Ran long-term analysis of hourly forecast adjustments
  • Readings Completed
  • Tasks for Next Week
    • Give Poster Presentation
    • Complete Presentation
    • Wrap up MATLAB functions
    • Complete Paper
  • Present Obstacles
    • There is a lot to do, but not so much time to do it.
    • Writing papers is not my best attribute
    • The graphs take time to create correctly

Week 10

Research Objectives

  • Give Poster Presentation
  • Complete Presentation
  • Wrap up MATLAB functions
  • Complete Paper

Weekly Log

  • Monday, July 29
    • Meet with Dr. Corliss
      • Paper Edits
      • Abstract
      • GasDay Presentation
    • Edit Paper
    • Cleaned Code
  • Tuesday, July 30
    • Edit Paper
    • Poster Session
  • Wednesday, July 31
    • GasDay Presentation
    • Edit Paper
  • Thursday, August 1
    • Formal Presentations
    • Edit Paper
  • Friday, August 2
    • Formal Presentations
    • Final Edits of Paper
    • Electronic Version of Paper Turned In

Weekly Progress Report

  • Tasks Completed
    • Poster Session Completed
    • Code Completed and Given to GasDay
    • Final Presentation Completed
    • Paper Completed
  • Readings Completed
  • Potential Future Work
    • Stay in touch with lab
    • Prepare paper for publication
    • Summer 2014
      • Jackson obtained an internship to work at Cigna-Healthspring for this summer
    • Graduate School
      • Jackson is still considering his options for after undergraduate school. These options include graduate school, as well as entering the workforce.
  • Present Obstacles
    • We answered the GasDay-GasHour Inequality, but in doing so, we raised so many more questions. There simply was not enough time to answer all of the questions we asked.
    • Future communication will occur via email.

Reference

Links for professors and other important items are generally located at their first mention. However, for quick reference, important reference links have been made available in this section.