In this post I collect together projects and ideas from the class and from myself. My criteria is that (a) a project should be a good complement to the content of this course, (b) can contribute to your career development, (c) are well defined and can be well managed and (d) can be finished in 4 weeks with 5 hours per week.

- Review data sets that track the locations and proximity of people with mobile phones and other types of sensors: Where can those data sets be downloaded? What are the works on each data set? What are the interesting findings, innovative methodology and significant applications of these data sets? Sum of 10-based logarithm of the data size (in bytes) >100, references > 50 from most cited works to least cited works. Tell a cohesive story about these data sets.
- Review social media data sets: Where can those data sets be downloaded? What are the works on each data set? What are the interesting findings, innovative methodology and significant applications of these data sets? Sum of 10-based logarithm of the data size (in bytes) >100, references > 50 from most cited works to least cited works. Tell a cohesive story about these data sets.
- Review one of the open source simulation softwares: Demonstrate how to use the software by walking through one example, show the architecture of the software (modules and their interactions, entry points, key modules, etc.), show how to dump events or system events from the software. Those software could include UrbanSim, MATSim, TranSIMS, MetroPolis, STEM, etc.
- Define the dynamics of network evolution through a set of events, simulate network evolution, make inferences from observations using either variation method or MCMC. Hint: we can follow the examples in the problem sets, work with fewer events, no need to work with real world data.
- Define the dynamics of diffusion through a set of events, simulate network evolution, make inferences from observations using either variation method or MCMC. Hint we can follow the examples in the problem sets. Hint: we can follow the examples in the problem sets, work with fewer events, no need to work with real world data.
- Identify a real world data, postulate the dynamics, and conduct inference using one of the algorithms discussed in the lectures.
- more to come….