Blog 2

After reading some papers during the first week, I got the broad idea of the research and have some thought about it. I met with the professor for a few times. We have to know each other’s ideas and then make some more detailed plan for our future work. We all agreed that Random Forest package in R studio is very powerful and might be a good start for the research, and also narrowed down all the methods to Random Forest and Hierarchical Bayesian Model for start.

I continued to read more papers about Random Forest, and sometimes the more times I read through some particular papers, the more I can get from them, every time I could get something new. After our searching for papers from WorldPop, Flowminder, and google scholars, we decided to read much more detailed into five papers about Random Forest and continue our paper-searching process. One of the papers from WorldPop is very useful. It is about the basic idea about developing WorldPop and nearly the most original idea about Random Forest. After settling down the basic line of the research, we begin to conduct the work ourselves.

The first and also very crucial part of the research is data collecting process. Because of the privacy and security aspect of data online, it is kind of time-consuming to search for available datasets and download it. Also, since the research requires multiple data layers, we also have to control the year and resource of the data and make them corresponding to each other. This also might be a reason for the quality of the data. And this is the part we are going to start at the end of this week and next week.