Add Epigames datasets to netdiffuseR

[aoliveram]

Hi all — quick note following our email thread with Tom, Andres (@codeanticode), and @gvegayon .

Andres has shared the Epigames datasets (simulations of disease outbreaks in a multiplayer game environment), which are tailor-made to demonstrate our standard metrics. A bit about the data:
These are dynamic temporal networks capturing user contacts and infection transmission events over time. The datasets vary significantly in length and temporal resolution (e.g., dataset 195 covers a few hours, while dataset 165 covers about 15 days). The raw data is highly detailed, exporting network states frame-by-frame (sometimes over 20,000 frames) as .mtx files.

What we'd like:

• Include a curated subset of the Epigames datasets (e.g., IDs 165) into the package's data repertoire.
• Add an internal parsing routine to easily load these frame-by-frame .mtx files into R-compatible formats (CsparseMatrix) and convert them to diffnet objects.
• Allow dynamic time discretization, e.g., setting nat_time_step_min = 24*60 to aggregate frames into daily networks, or 60 for hourly slices. (Optional)
• Update package documentation to describe the origin, variables, and structure of the Epigames data.

This way we can easily call this data to run the models for the upcoming R03 NIH proposal.

I'll open a branch and start working on this!

[gvegayon]

Sounds good, I dig the idea of having a function to collapse timeframes. Regarding the format, is that a standard format? Ensure you avoid adding extra dependencies. It'd be great if you could open a draft PR while you are working on it, so I can start tracking it as well.

[aoliveram]

hey @gvegayon, I started a draft PR on this. The data in his diffnet and raw (non-diffnet) classes are added, but there are .csv that need to be deleted. Also, I want to add the code that generated the csv data in the first place. I need to figure out how to do that. More on that later.

The documentation still needs to be added, so failures are expected.