Project¶

In which you will propose and carry out an original inference of some kind.

Description¶

This is your opportunity to apply what you've learned in a new context, perhaps using more challenging or realistic data (for your particular area), or otherwise explore beyond the scope of the existing notes/tutorials. There's a wide landscape of possible projects, and no absolutely fixed requirements.

At the simpler end, you might find a data set and a model to fit to it in the scientific literature, constrain the posterior distribution of the model parameters, find credible regions, and assess the goodness of fit as usual... all without straying beyond the methodology we've already covered. Provided that the solution is not too duplicative of a tutorial that already exists, this would be fine.

Projects could, however, also explore new methodology, either something we've covered in the notes but not provided tutorials on, or something entirely new.

Rarely, a project might focus primarily on exposition of a new topic (i.e. resemble notes more than a tutorial), although there still must be some form of analysis/demonstration component.

To make sure that the content and scope are within reason, you will have to propose you project and have it approved; this will include pitching your idea to the class and is an excellent way to find others to collaborate with.

Getting inspiration¶

Of course, one way to get inspired is to look at papers in your field (or some other field) - lots of them probably present data and fit models to them!

Some time before the project pitch phase, we will gather a list of previous projects for this course to give you a better idea of what's possible.

You are permitted to do a project related to your own research, provided it is not something you have have already done (i.e. the project represents original work). Projects should not duplicate effort for some other class.

Collaboration¶

Working in pairs or groups is strongly enouraged, since it enables more interesting and involved projects. Collaborators on a project will share the same grade for it, and our expectations will be proportional to the number of collaborators.

Data and intellectual property issues¶

We encourage the use of real data when possible, since wrestling with the real-life problems it represents is an important learning opportunity. Sometimes, however, the complexity or computational challenge associated with this is too impractical given the limited time we have. Often this issue can be resolved be using "realistic" rather than real data and/or simplified models, without sacrificing the essence of the project.

Although exceptions are possible, we generally discourage the use of proprietary data in class projects. Sharing such data with project collaborators and/or sharing your results with the class may not be permitted by whatever group obtained/produced the data, and finding out and getting permission can be time consuming. Often, similar and publicly available data or mock data are sufficient for our purposes.

Keep in mind that any collaborators on the project may also share intellectual rights to the work, should it eventually lead to publication.

Timeline and deliverables¶

We will devote some class time to the following activities to provide some structure. Canvas will have the specific dates associated with these.

  • Brainstorm ideas, discuss with peers in class
  • Formally (that is, with a slide deck) propose project ideas, get peer and instructor feedback
  • Present a progress report in class, get peer and instructor feedback

The final products of the project will be

  • Presentation of the results in class (during the final exam period)
  • A journal-style written report

Generally, these are expected to

  • introduce to the data being analyzed;
  • motivate and explain the model being investigated;
  • discuss how the inference was accomplished, including any required approximations;
  • visualize and interpret the results, including checks for numerical convergence and similar; and
  • discuss how the analysis might be extended, including conclusions about what did or didn't work well.

Grading will be broadly based on this rubric, with adjustments as appropriate to any particular project.