Computational Social Science

(For new reader and those who request 好友请求, please read my 公告栏 first)One of the advantages of being retired is the freedom from any kind of obligation. You do something not because you had to but because you wanted to. Duringyour working years you may be interested in many things, but you do not have thetime to attend to them since your job duties come first. In retirement , you don’t have this difficulty and are free to pursue anything you are intellectually curiousabout. Between hard sciences such as physics and mathematics and humanities such as literature and history lie the disciplines often disparagingly termed as “soft sciences” such as social science. Last week there was a conference on Computational Social Science at Harvard. I decide to go and see just what the field is all about and what people in the field are interested in. Starting from essentially zero knowledge, below is what I learned.Traditionally, social sciences have three avenues of research:1.Design of surveys2.Analysis of government data (e.g. census)3.Study of specific problem in depth (e.g., the spread of xxx disease in xxx province)Because of privacy concern and expense of collecting data, social sciences were often data-poor. However, with increasing digitalization and networking of ourlives, the situation is drastically changing. Instead, according to peoples in the field, data are no longer a problem. In fact the situation has become data-rich and information-poor. We are being overwhelmed with data from all sources but don’t know what to do with them. Example: there are over 100 million blog sites on the Internet. There is a project at Harvard that collects every page ever written on these sites daily. But no one has the time or budget to analyze the data (yes, big brother are watching all you bloggers). We have still quite a ways to go along the data-information-knowledge-wisdom path.Examples of data that are currently available:1.Cell phone traffic data all over the world2.Detail election data3.Geophysical information4.Biological and genetic information on the population5.Purchase records of individuals6.RFID data on everythingThese data are multidimensional and inter-related. Making sense of them via visualization, statistical analysis, and providing security at the same time are the hard challenges of current social sciences. In time social scientists feel they will be able to understand and predict spread of social trends such as obesity, happiness., etc. The discipline is on the cusp of a revolution.Note added 4/2/2017 Readers might be interested to read http://blog.sciencenet.cn/blog-1565-866637.html about "differential privacy" and the danger of too much data