Feedback from ASMS2023

[bittremieux]

Points of improvement:

• Ensure proper balance between lectures, questions, and hands-on exercises.
• Link the concepts introduced during the lectures to concrete examples of how they are used on MS data in the scientific literature.
• Provide more information on how to get started with Python programming and getting familiar with Jupyter notebooks for people new to coding.
• Add comments to all lines of code in the notebooks to explain what they do.
• Additional coding exercises for more advanced learners; implementing some of the algorithms themselves.
• Include more hands-on exercises working with MS/real-life data, potentially re-analyzing data from publications.
• Provide more hands-on exercises, e.g. as take-home tasks.
• Include a session on which ML models to use for specific types of data.

Notes from ourselves:

• Sessions with minimal time for the hands-on exercises: Getting Started, Dimensionality Reduction.
• Clustering lab: use falcon to cluster MS data.

[bittremieux]

Feedback from the ASMS survey.

Lectures:

• Ensure a balance between math fundamentals and real-world examples.
• Include a flow chart to decide which ML algorithm to use based on data characteristics.

Labs:

• Clearly advertise that Python is used for coding tasks.
• Optional pre-course assignment to get familiar with Python programming and the Jupyter notebook environment (lab 1?).
• Add comments explaining what every line of code does to improve understanding for novice coders.
• Dedicate enough time to getting to know ML software libraries and teach people how to consult the documentation.
• Divide the labs into beginner and advanced tasks to cater to different levels of experience. Advanced users want to implement algorithm fundamentals
• More hands-on activities learning how to evaluate ML results in the literature (when is it satisfactory vs when did it fail) and how to get started analyzing your own data.