oa Expanding The Humanitarian Computing Library Using Automatic Tools

Humanitarian computing has become a critical field to support the response to natural and man-made disasters all over the world. The Humanitarian Computing Library (http://humanitariancomp.referata.com/) is a collaboratively edited ("wiki") site that includes a compilation of publications related to humanitarian computing. Each publication is represented by its meta-data including type (e.g. conference presentation or journal article), authors, date, abstract, published location as well as its URL. Currently, 770 publications exist in the library and are categorized into a few dozen categories to provide easy access for researchers. The aim of the project is to augment the contents of the library to include up-to-date publications, by automating the search for related publications instead of using the previous, manual search method. In addition, we extend the publications' meta-data by adding information about their citation count, in order to highlight the most important ones. The derived solution is a publicly-available web service, that allows any editor to use it to expand the library. The web services work by performing a programmatic query to Google Scholar and analyzing the result. The query contains a sample obtained from the list of papers on a given category of the library (e.g. "Social Media and Crises"). For each of the papers currently in the library, we obtain new papers citing it. We then remove duplicate titles and sort by decreasing frequency, removing papers that are already present in the library. As the Humanitarian Computing research field is quite new, new publications are highly dependent on previous ones, and thus related publications can be identified by three indicators: they cite an existing publication from the library, have common phrases/keywords with existing previous publications, and are highly cited. The developed web service made it easy to add recent publications to the library with much less effort than before. In conclusion, the deployed solution acts as a tool for collecting research papers on humanitarian computing, and helps expanding this knowledge base. Such a publicly-available web service can help advance the state of the art of Humanitarian Computing, leading to an improved understanding of this field, and potentially help minimize the loss of lives and property during disasters by the application of new technologies.