Data assimilation and estimation theory

Class times: Fall 2021 – MonWedFri 10:00 PM - 10:50 AM

Class room: PE 105

Final exam: 5:00 PM Friday December 10th, see final exam policy

Vaccination sites

• List of vaccination sites in Washoe County
• Washoe County Health Department clinics
• Immunize Nevada list of clinics

Help & Questions

There will be a classroom discussion forum where you can seek help from the instructor and peers about the assignments. Some assignments will require that you post and discuss results in this forum.

The best way to contact the instructor is via email or message within Canvas. Messages received before 4:00 PM during business days will usually get a response the same day. Emails received after 4:00 PM or during the weekend are not guaranteed a same-day response. Longer questions will not be answered by email and will be directed to appointments or office hours.

The instructor will be available from 4:00 - 5:30 PM MonWed for public office hours and questions, shared with students from STAT 445 / 645. If you cannot attend office hours or if you need to discuss something privately you can set an appointment with the instructor directly.

Netiquette

When communicating online, you should always:

• Treat your instructor and peers with respect, even in email or in any other online communication.

• Use clear and concise language. Be respective of readers’ time and attention.

• Remember that all college level communication should have correct spelling and grammar.

• Avoid slang terms and texting abbreviations.

• Use standard fonts that are optimized for online reading (e.g., sans serif) along with a consistent and readable size (12 or 14 pt.)

• Avoid using the caps lock feature AS IT CAN BE INTERPRETED AS YELLING.

• Limit and possibly avoid the use of emoticons. Not everyone knows how to interpret them.

• Be cautious when using humor or sarcasm as tone is sometimes lost in an email or discussion post and your message might be taken literally or offensively.

• Be careful sharing personal information online (both yours and other’s).

• Always give proper credit when referencing or quoting another source. (Corollary: Don’t copy and paste another student’s post and claim it as original as that is essentially plagiarism.)

• Don’t repeat someone else’s post without adding something of your own to it. (See corollary above regarding reuse of someone else’s post.)

• Always be respectful of others’ opinions even when they differ from your own. When you disagree with someone, you should express your differing opinion in a respectful, non-critical way. (Corollary: Do not make personal or insulting remarks.)

Content

Data assimilation (DA) refers to techniques used to combine the data from physics-based, numerical models and real-world observations to produce an estimate for the state of a time-evolving random process and the parameters that govern its evolution. Owing to their history in numerical weather prediction, DA systems are designed to operate in an extremely large dimension of model variables and observations, often with sequential-in-time observational data. As a long-studied “big-data” problem, DA has benefited from the fusion of a variety of techniques, including methods from Bayesian inference, dynamical systems, numerical analysis, optimization, control theory and machine learning. DA techniques are widely used in many areas of geosciences, neurosciences, biology, autonomous vehicle guidance and various engineering applications requiring dynamic state estimation and control.

After introducing students to essential elements of linear estimation theory, this course will focus on the development and application of operational, statistical learning algorithms in nonlinear, physics-based models. The course will rely on the Python programming language for many exercises and for the final project, with the DAPPER library providing a framework for the class. The DAPPER library is suitable for small- to mid-scale DA and uncertainty quantification research and provides a pedagogical stepping stone for students to high-performance libraries and / or writing their own solvers. Students do not need to know Python programming in advance, but some familiarity with programming and computation generally will be beneficial. final exam policy.

Structure for online learning

This class will follow a flipped design in which students are expected to watch one or more short lectures before the class session. If a student is unable to or would prefer not to attend in-person, a HyFlex remote learning option will be granted. Video lectures will cover content that is relevant to the day’s activity in class. Class sessions will be held on Zoom and students are expected to attend the Zoom meeting to work on the day’s activities with their cohort. Students should have a laptop for working through activities during class in-person class in the Python programming language, or should be able to attend remotely with a computer that will accommodate this. Each week will require posting a discussion of about the results in the week’s activities. Programming experience in Python is not required and there will be bi-weekly assignments in DataCamp that will build on fundamental skills in the Python scientific programming environment.

Recommended background:

This course will treat some of the topics in the following courses. Students will benefit from taking these courses in advance, or concurrently, but these courses are not strictly required.

• MATH 285 (Differential equations) or equivalent
• STAT 461/462 (Probability and Stochastic Processes) or equivalent
• STAT 445 (Introduction to Statistical Computing) or equivalent
• STAT 446 (Introduction to Bayesian Statistics) or equivalent

Personal and University computers

This is a class that will involve programming, data manipulation and computation. You are recommended to use a personal device where you have privileges to install statistical software and store persistent data to, and to clone git repositories. Student laptops are available from the UNR library through their equipment lending program. Graduate students in the Statistics and Data Science program may be eligible to borrow a laptop computer from the Department of Mathematics and Statistics if they do not have a personal device, please contact the instructor for more information. Students may also use library computers and / or remote desktop services for attending Zoom sessions online and working on exercises and assignments. Students who attend class in PE 105 are encouraged to bring an audio headset for communication in Zoom group work during the class session

Students using the Windows Remote Desktop client will need to log into Windows Security with user name written in the form UNR\NetID. UNR should be followed by “backslash”, not “slash”. Mac users will need to additionally download the Microsoft Remote Desktop Application to use remote desktop services. UNR remote desktop clients has Anaconda and LaTeX installed already, and assignments may be completed in this environment if a student does not have a computer with these capabilities. It is recommended to save any data / assignment files in cloud storage as assignment progress may not be saved otherwise.

Canvas

Students are responsible for checking their email accounts and Canvas for announcements. Students are assumed to be aware of all information posted to these sources prior to each meeting. Announcements, homework and grades will be posted in Canvas.

Course archive page

This page includes an archive of lecture notes and activities from the class, along with the current schedule for the semester.

Video lectures

Each week there will usually be three video lectures to watch before the class activities during the normal class sessions.

Activities

In a normal class session, students will work through a class activity that uses the content from the video lectures that week. Participation credit requires a written summary and analysis at the end of the week according to the rubric in Canvas.

Homework

The following DataCamp courses will be required for completion credit in the fall:

• Introduction to Python
• Intermediate Python
• Importing data in Python
• Python Data Science Toolbox Part 1
• Python Data Science Toolbox Part 2
• Data Types for Data Science in Python
• Writing Efficient Python Code
• Writing functions in Python

Certificates of completion for these courses completed before the course start will be accepted for full credit.

Final exam policy

The final exam will consist of a final term report written on one of several available topics to choose from in the assignment. The final project must be submitted on Canvas by 5:00 PM on Friday, December 10th. Details on the assignment are in the final project description.