Disclaimer: I am also currently a PhD student.
My research lies at the intersection(s) of HCI, privacy, location based social networks, mobility theory and spatial statistics. This means that the work in this area gets published from computer science-y journals like ACM CACM, IEEE Privacy & Security to communication-y journals like New Media and Society to conferences like CHI, SOUPS and MobileHCI.
I adopt 3 strategies to organize my own work. If I do not do this then the existing literature and new critical thinking about my area will spiral out of control for me.
I maintain an annotated bibliography for any relevant work in this area. As Suresh pointed out, its generally up-to-date and quite liberal in nature. This is not public but I always email it to folks who ask for it. I find that maintaining a bibliography in Mendeley, exporting as a BibTeX file and writing a couple of sentences about each article works for me. I make sure that each article is filed under some loose sub-heading. For instance, one such general sub-heading in my area could be "Location Privacy and Surveillance."
I write scripts in order to extract, manipulate and analyze data and I always publish them on github. This is public. Usually, I write scripts in php, python or R. Its great to see other folks forking my projects or following them and making them better than I could. :)
I write short blog posts about my impressions on certain topics and also post some simplistic visualizations and analyses of my work there. I take special care in making sure that these are short but to the point. Nobody likes to read long, rambling blog posts. :P