Nowadays, thanks to the spread of mobile devices, people can easily access to the internet. News and other information can be retrieved by just submitting a web request using a browser or smartphone application. From a web request it is possible to extract the topics discussed in the requested page that, together with the geographic origin of the requests (even more and more accurate thanks to mobile sensors), represents a meaningful set of data to analyze. This paper will present a system able to analyze logs of web requests in order to extract the main topics for a specific geographic area taking care of both qualitative and performance aspect, in particular in avoiding costly re-computations.
Please open the report and the poster for more information.
You can find sample of data and pickle files here.