Using the R Language to Manage and Show Statistical Information in the Cloud

Abstract

We present a methodology to enable users to interact with the statistical information (survey data) of a public opinion institute which is stored in Cloud infrastructure. Mainly using R, this approach was developed following the open-data philosophy. Also, as we used R, the implementation is mainly based on open-source software. R has several advantages from the point of view of data management and acquisition, as it is becoming a common framework that can be used to structure the processes involved in any statistical operation. It further simplifies the access to data and enables the use of all the power of R in Cloud infrastructure. This methodology was applied successfully to develop a tool to manage the data of the Centre d'Estudis d'Opinió, but it can be applied to other institutions to enable open access to their data. The infrastructure was also deployed to Cloud infrastructure to assure its scalability and 24/7 access. improve access to Federal data and expand creative use of those data beyond the walls of government by encouraging innovative ideas (e.g., web applications)".There are several websites and tools to access public statistical information as well as several applications that allow the reuse and sharing of code related to the access of public information, such as [8] or [9]. The next step is to allow users without technical knowledge to access the information and perform easy statistical tasks with it. To do this, the user must be able to execute those tasks on a remote server that stores both remote information and certain statistical functions.The possibilities of allowing end-users to execute certain statistical functions to obtain new information from the data were described by [10]. Currently, there are different tools that to show information over the web and allow the execution of statistical functions by the end users, e.g., the NESSTAR system [11]. In parallel with these proprietary solutions, several efforts are focused on developing APIs (Application Protocol Interfaces) to access statistical information. As an example, Data.org is preparing an API that allows users to interact with the system data to build their own applications and mash-ups; in [12], an API has also been implemented to interact with its data. However, the question of how to develop and use these APIs remains open. Every infrastructure that develops this type of solution implements a new API, and the developers must be able to address all of them.Another problem is related to the data preparation; there are different alternatives already available to define the surveys, e.g., [13] or [14]. These tools allow the user to export the data to various formats and perform posterior analyses (a well-known format is the Triple-S, an XML for survey software that enables the user to import and export surveys between different software). The main issue with this approach is that manual operations are required to process the data. In our proposed approximation, once the surveys are completed, their datasets can easily be upload...