Strict Standards: Non-static method MjFunctions::mj_captcha_init_sessions() should not be called statically in /var/www/vhosts/ on line 140
About SEN2SOC experiment

About SEN2SOC experiment

SEN2SOC: bridging SENsor measurements and SOCial networks interactions via natural language generation for supporting smart city services

The SEN2SOC experiment utilizes the SmartSantander testbed in Santander, Spain, which constitutes a wireless sensor network of both fixed sensor nodes (e.g., at streetlights, facades, bus stops) and mobile sensor nodes (e.g., on board buses). Our experiment promotes the interaction between sensor networking platforms and humans in an effort to engage the Santander community around the SmartSantander framework. The sensor-to-social interaction is established through the combination of both sensor and social data into meaningful services or functions. On the one hand, social network behavior regarding the city of Santander is analyzed and the respective results are presented to users of the SEN2SOC applications. On the other hand, environmental sensor measurements of SmartSantander are processed and displayed in a simple and understandable way, as well as alerts regarding extreme environmental conditions are generated and reported. However, users of the SEN2SOC applications are enabled to express how they sense the environment (i.e., “users as sensors”) or to share environmental alerts in real time on the social network of their preference. Thus, along with sensor or social information provision, the SEN2SOC experiment accommodates input from the community.

The SEN2SOC experiment comprises a Web and a mobile application. The Web application essentially constitutes a monitoring tool for the SmartSantander sensor network and offers functions such as: environmental conditions’ monitoring; visualization of current or historic sensor data; comparison of data graphs for nearby sensors; and statistical analysis results related to sensor measurements. The mobile application presents to its users chromatic maps of various environmental parameters; suggests routes based on favorable environmental conditions; shows alerts regarding extreme environmental values; and informs about trending topics around Santander resulting from social media content analysis.

SEN2SOC Platform


SEN2SOC Architecture

The SEN2SOC architecture is component-based and is organized around the following components: the Sensor Data Monitoring, the Social Data Observer, the Interface, the Statistical Analysis, the Web Application, and the Mobile Application.

SEN2SOC Architecture

The Sensor Data Monitoring component retrieves sensor data from the SmartSantander platform. Its primary tasks include: data retrieval, data cleansing, data aggregation, and data storage. Sensor data is spatially aggregated in order to construct the so-called virtual nodes, that correspond to various geographic areas that Santander is divided into. Also, the Sensor Data Monitoring component is responsible for generation of alerts whenever environmental sensor readings exceed certain predefined thresholds (i.e., detection of extreme environmental conditions). Alerts are published to the SEN2SOC social media accounts and are forwarded to the SEN2SOC Mobile Application users in real time.

The Social Data Observer component collects social media content geolocated within Santander and performs user-generated content mining on various social media networks, such as Twitter, Flickr, and Foursquare. The Social Data Observer aggregates social media posts and photos based on time and geographic location, and also identifies popular topics and photo clusters based on semantic similarities and geographic proximity.

The Interface essentially constitutes the intermediary module of the SEN2SOC platform, which: lies in between all components; specifies services that other components can utilize; and allows data communication among SEN2SOC components.

The Statistical Analysis component correlates and analyzes sensor data. This component supports sensor data mining, performs statistical analysis, detects sensor data anomalies, and reports results to the SEN2SOC Web Application. Statistical analysis methods or models applied pertain to: data smoothing; prediction; correlation between two or more sensors; and autocorrelation for detection of repeating patterns.

The Web Application forms a monitoring tool for the SmartSantander sensor network and supports the visualization of real-time and historic sensor data. Other features include the following: detection of closest sensors to the sensor node selected by the user and comparison of the respective data using line charts; prediction models for various environmental parameters; chromatic maps of various environmental parameters; and presentation of social media content geolocated within Santander.

The Mobile Application offers various services such as: chromatic maps illustrating Santander environmental conditions in real time; route recommendations based on favorable environmental conditions; alerts on extreme environmental conditions; suggestion of areas and points of interest to city citizens and visitors; analysis results of social media users’ activity. Furthermore, the Mobile Application implements a mechanism that allows users to express their perception on the present environmental parameters of their current location, thus in a way “validating” SmartSantander sensor measurements.

The control flow diagram of the SEN2SOC platform is shown in the following figure.

SEN2SOC ControlFlow

Virtual Nodes

The SEN2SOC experiment bases some of its services on spatially aggregated sensor data. More specifically, Santander is divided into geographic areas and each geographic area forms a so-called “Virtual Sensor”. For every sensed environmental parameter, we calculate one aggregated value (virtual value) per area in the following way: a virtual value for an area equals the average value of all measurements recorded by nodes in that area at a specific time interval. Both fixed and mobile sensor nodes contribute to calculation of virtual values. Apparently, a mobile sensor contributes to various virtual nodes (i.e., Santander areas) depending on the current location of the mobile sensor on the map, whereas a fixed sensor always contributes to the same virtual node that corresponds to the area containing the sensor.

With regard to geographic division, Santander is divided into 148 geographic areas according to the shapefile provided by the National Statistics Institute of Spain (Instituto Nacional de Estadistica – INE). We note that a shapefile is a digital vector storage format for storing geometric location and associated attribute information. The afore-mentioned geographic division is reviewed every 10 years and sections of Santander are adapted according to various operational criteria, such as the number inhabitants per section. We are using the latest Santander shapefile available that was issued on May 2013 and the respective Santander geographic division can be seen in the following figure.