Qatar Foundation Annual Research Conference Proceedings Volume 2018 Issue 3

One of the main suppliers of the workforce for the engineering industry and the economy in general is the higher education sector. The higher education sector is consistently being challenged by the fast evolving industry and hence it is under constant pressure to fulfill the industry's ever-changing needs. It needs to adapt its academic curricula to supply the industry with students who have update-to-date and relevant competencies. Nevertheless, the gap still exists between what education curricula offer and the skills that are actually needed by the industry. Therefore, it is crucial to find an efficient solution to bridge the gap between the two worlds. Bridging the gap helps the industry to cut the costs of training university graduates and assists in higher education advancement. In response to these issues, competency-based education was developed. It was first developed in the United States, in response to the growing criticisms towards traditional education that was seen as more and more disconnected from the societal evolutions, especially changes within the industry. Despite some criticisms, the competency-based education pedagogical approach has been employed by several western countries to improve their upper-secondary vocational curricula. In recent times, it started to be more and more adapted to the higher education, as a way to update and improve academic courses. In this research work, a semantic ontological model is presented to model the competencies in the domains of education and the industry. It illustrates the use of ontologies for three identified end users: Employers, Educators, and Students. Ontologies are best used to solve problems of interoperability between different domains, provide a shared understating of terms across domains and are helpful in avoiding the wasted effort when translating terminology between domains. They also provide opportunities for domain knowledge reuse in different contexts and applications. Ontologies can also act as unifying framework between software systems and eliminates interoperability and communication issues that are faced when trying to translate concepts between systems. The scope of the research work is to build an ontology representing the domain concepts and validate it by building competency models. The competencies from the domains of education and industry are defined and classified in the ontology. Then by using a set of logical rules and a semantic reasoner, we would be able to analyze the gap between the different education and industry profiles. We will propose different scenarios on how the developed ontology could be used to build competency models. This paper describes how ontologies could be a relevant tool for an initial analysis and focus on the assessment of the competencies needed by the engineering market as a case study. The research questions this work investigates are: 1) Are semantic web ontologies are the best solution to model the domain and analyze the gap between industry needs and higher education? 2) What ontology design approaches are more suitable for representing the competencies model? 3) What are the limitations of ontology modeling? Two main limitations are discussed: The Open World Assumption and the Semantic Reasoner limitations. This work is part of the Qatar Foundation NPRP Pro-Skima research project.

Many statistics show the number of connected devices (Internet of Things; IoT) worldwide to grow drastically in the next few years, to exceed 30 billion just by 2020. All these devices need to be connected to the internet, to establish tow ways communication with the backend applications. The list of applications and services that can be enabled by using IoT can be endless, covering different areas such as smart cities, smart home, connected vehicles, intelligent transport systems, agriculture, air and weather monitoring, industry 4.0, etc. One of the fundamental requirements of an IoT device is to be connected and reachable at any time. But when we look at the different applications that run on top of IoT, many of them do not require a continuous connection to the internet. For example, the air monitoring IoT devices, normally they sense and report data only once every 15 to 60 minutes. Such devices would not require a continuous connection to the Internet, but only every 15 to 60 minutes. For such devices and use-cases, we propose to use a flying IoT gateway, that can come next to the sensor every e.g. 15 minutes, to take the data that has been collected by the sensor and carry it to the backend. In this contribution we present a prototype of a solution that uses unmanned aerial vehicles (UAVs), aka drones, to provide a delay tolerant data routing solution for IoT devices. In this solution, a drone flies over a set of deployed IoT devices, to retrieve the collected and stored data and then deliver it to the backed. Such a solution can be suitable for sensing devices that do not require a real-time communication, like traffic speed cams. Indeed, the speed cams can collect data and store it locally, until a drone comes to carry and transfer it to the backend.This solution helps does not only reduce the overall cost by eliminating the cost of the Internet connectivity at each IoT device, but it also reduce the security vulnerability as the devices are physically not connected to the internet all the time, nor directly.This work has been been conducted under the R&D project NPRP9-257-1-056 which is funded and supported by QNRF.

Extended AbstractPeople increasingly use social media such as Facebook and Twitter during disasters and emergencies. Research studies have demonstrated the usefulness of social media information for a number of humanitarian relief operations ranging from situational awareness to actionable information extraction. Moreover, the use of social media platforms during sudden-onset disasters could potentially bridge the information scarcity issue, especially in the early hours when few other information sources are available. In this work, we analyzed Twitter content (textual messages and images) posted during the recent devastating hurricanes namely Harvey and Maria. We employed state of the art artificial intelligence techniques to process millions of textual messages and images shared on Twitter to understand the types of information available on social media and how emergency response organizations can leverage this information to aid their relief operations. Furthermore, we employed deep neural networks techniques to analyze the imagery content to assess the severity of damage shown in the images. Damage severity assessment is one of the core tasks for many humanitarian organization.To perform data collection and analysis, we employed our Artificial Intelligence for Digital Response (AIDR) technology. AIDR combines human computation and machine learning techniques to train machine learning models specialized to fulfill specific information needs of humanitarian organizations. Many humanitarian organizations such as UN OCHA, UNICEF have used the AIDR technology during many major disasters in the past including the 2015 Nepal earthquake, the 2014 typhoon Hagupit and typhoon Ruby, among others. Next, we provide a brief overview of our analysis during the two aforementioned hurricanes.Hurricane Harvey Case StudyHurricane Harvey was an extremely devastating storm that made landfall to Port Aransas and Port O'Connor, Texas, in the United States on August 24-25, 2017. We collected and analyzed around 4 million Twitter messages to determine how many of these messages are, for example, reporting some kind of infrastructure damage, or reports of injured or dead people, missing or found people, displacements and evacuation, donation and volunteers reports. Furthermore, we also analyzed geotagged tweets to determine the types of information originate from the disaster-hit areas compared to neighboring areas. For instance, we generated maps of different cities in the US in and around the hurricane hit areas. Figure 1 shows the map of geotagged tweets reporting different types of useful information from Florida, USA. According to the results obtained from the AIDR classifiers, both caution and advice and sympathy and support categories are more prominent than other informational categories such as donation and volunteering. In addition to the textual content processing of the collected tweets, we perform automatic image processing to collect and analyze imagery content posted on Twitter during Hurricane Harvey. For this purpose, we employ state-of-the-art deep learning techniques. One of the classifiers deployed in this case was the damage-level assessment. The damage-level assessment task aims to predict the level of damage in one out of three damage levels i.e., SEVERE damage, MILD damage, and NO damage. Our analysis revealed that most of the images (∼86%) do not contain any damage signs or considered irrelevant containing advertisements, cartoons, banners, and other irrelevant content. Of the remaining set, 10% of the images contain MILD damage, and only ∼4% of them show SEVERE damage. However, finding these 10% (MILD) or 4% (SEVERE) useful images is like finding a needle in a giant haystack. Artificial intelligence techniques such as employed by the AIDR platform are hugely useful to overcome such information overload issues and help decision-makers to process large amounts of data in a timely manner.Fig. 1: Geotagged tweets from Florida, USA.Hurricane Maria Case StudyAn even more devastating hurricane than the Harvey that hit Puerto Rico and nearby areas was hurricane Maria. Damaged roofs, uprooted trees, widespread flooding were among the scenes on the path of Hurricane Maria, a Category 5 hurricane that slammed Dominica and Puerto Rico and has caused at least 78 deaths including 30 in Dominica and 34 in Puerto Rico, and many more left without homes, electricity, food, and drinking water.We activated AIDR on September 20, 2017 to collect tweets related to Hurricane Maria. More than 2 million tweets were collected. Figure 2 shows the distribution of the daily tweet counts. To understand what these tweets are about, we applied our tweet text classifier which was originally trained (F1 = 0.64) on more than 30k human-labeled tweets from a number of past disasters. AIDR's image processing pipeline was also activated to identify images that show infrastructure damage due to Hurricane Maria. Around 80k tweets contained images. However, ∼75% of these images were duplicate. The remaining 25% (∼20k) images were automatically classified by the AIDR's damage assessment classifier into three classes as before. Figure 2: Tweets count per dayWe believe that more information can be extracted from image about the devastation caused by the disaster than relying solely on the textual content provided by the users. Even though it is in the testing phase, our image processing pipeline does a decent job in identifying images that show MILD or SEVERE damage. Instead of trying to look at all the images, humanitarian organizations and emergency responders can simply take a look at the retained set of MILD or SEVERE damage images to get a quick sense of the level of destruction incurred by the disaster.

According to Energy Technology Perspectives 2014, there are approximately 900 million light duty vehicles (not counting two- and three-wheelers) today, and that number will be doubled by 2050 to reach 2 billion. Such a considerable increase will bring-in further challenges for road safety and traffic efficiency. Motor vehicle crashes are the leading cause of death for children and young adults in the United States, with an annual death toll of 33,000 and over 2.3 million people injured. Those figures are representative of the challenge not only in the US but also in other regions including Qatar and the golf region. Vehicle to Vehicle and Vehicle to Infrastructure (V2X) communication technology, which is heading our roads in few years, is seen as a great solution to reduce number of accidents on the road, and it is considered as an enabler of next generation of road safety and Intelligent Transport System (ITS). V2X communication is not limited to vehicle-to-vehicle and vehicle-to-infrastructure, it is also meant to be used for vehicle-to-bicycle and even vehicle-to-pedestrian communication. Indeed, by enabling real-time and fast communication between vehicles and vulnerable road users such as bicycles and road users, we may make the road much safer for those vulnerable road users. This is one of the use cases we would like to enable and test in the Qatar V2X Field Operational Field (Qatar V2X FOT) which is supported and funded by Qatar National Research Fund (QNRF). Enabling vulnerable road-users such as bicycles and pedestrians with V2X capabilities has many challenges. The main challenge is the energy consumption, as V2X operates in 5.9GHz radio band, which consumes relatively high energy. Therefore, to operate V2X on those vulnerable users, especially the pedestrian, the user needs to hold a battery which requires to be regularly recharged. We came out with a solution to this problem, which reduces the energy consumption of the V2X device to a level it can operate a small battery suitable to be a wearable device. This poster will expose the challenges when using V2X on vulnerable road users, especially pedestrians, and it will present solution and the related prototype that is developed and tested within the NPRP project. The solution and prototype presented in this poster are the outcomes of the research project NPRP 8-2459-1-482 which is supported and funded by QNRF.

Twitter has been developed as an immense information creation and sharing network through which users post information. Information could vary from the world»s breaking news to other topics such as sports, science, religion, and even personal daily updates. Although a user would regularly check her Twitter timeline to stay up-to-date on her topics of interest, it is impossible to cope with manual tracking of those topics while tackling the challenges that emerge from the Twitter timeline nature. Among these challenges are the big volume of posted tweets (about 500M tweets are posted daily), noise (e.g., spam), redundant information (e.g., tweets of similar content), and the rapid development of topics over time. This necessitates the development of real-time summarization systems (RTS) that automatically track a set of predefined interest profiles (representing the users» topics of interest) and summarize the stream while considering the relevance, novelty, and freshness of the selected tweets. For instance, if a user is interested in following the updates on the “GCC crises», the system should efficiently monitor the stream and capture the on-topic tweets including all aspects of the topic (e.g., official statements, interviews and new claims against Qatar) which change over time. Accordingly, real-time summarization approaches should use simple and efficient approaches that can scale to follow multiple interest profiles simultaneously. In this work, we tackle such problem by proposing RTS system that adopts a lightweight and conservative filtering strategy. Given a set of user interest profiles, the system tracks those profiles over Twitter continuous live stream in a scalable manner in a pipeline of multiple phases including pre-qualification, preprocessing, indexing, relevance filtering, novelty filtering, and tweets nomination. In prequalification phase, the system filters out non-English and low-quality tweets (i.e., tweets that are too short or including many hashtags). Once a tweet is qualified, the system preprocesses it in a series of steps (e.g., removing special characters) that aim at preparing the tweet for relevance and novelty filters. The system adopts a vector space model where both interest profiles and incoming tweets are represented as vectors constructed using idf-based term weighting. An incoming tweet is scored for relevance against the interest profiles using the standard cosine similarity. If the relevance score of a tweet exceeds a predefined threshold, the system adds the tweet to the potentially-relevant tweets for the corresponding profile. The system then measures the novelty of the potentially-relevant tweet by computing its lexical overlap with the already-pushed tweets using a modified version of Jaccard similarity. A tweet is considered novel if the overlap does not exceed a predefined threshold. This way the system does not overwhelm the user with redundant notifications. Finally, the list of potentially-relevant and novel tweets of each profile is re-ranked periodically based on both relevance and freshness and the top tweet is then pushed to the user; that ensures the user will not be overwhelmed with excessive notifications while getting fresh updates. The system also allows the expansion of the profiles over time (by automatically adding potentially-relevant terms) and the dynamic change of the thresholds to adapt to the change in the topics over time. We conducted extensive experiments over multiple standard test collections that are specifically developed to evaluate RTS systems. Our live experiments on tracking more than 50 topics over a large stream of tweets lasted for 10 days show both effectiveness and scalability of our system. Indeed, our system exhibited the best performance among 19 international research teams from all over the world in a research track organized by NIST institute (in the United States) last year.

Complex adaptive systems exhibit emergent behavior. This type of behavior occurs in volatile environments involving cyber-physical systems, such as those aimed at smart cities operations. Adaptive systems’ maintenance aims at improving their performance by dealing with the treatment of continuous and frequently changing requirements. Adaptive systems’ behavior requires therefore up-to-date requirements traceability. To this end, we need to understand the requirements and to localize the program parts that should be modified according to the description of the new requirements. This process is known as Requirements Traceability Recovery (RTR). The process of generating requirements traceability, when done by a human (e.g., system maintainer), is time consuming and error-prone. Currently, most of the approaches, in the literature, are time consuming and semi-automatic that always need the intervention of the user. In our work, we are specifically interested in following the link between requirements and the code of the software system with the aim of helping the designer to update the system appropriately by automating the traceability process. To do this, we formulated the RTR problem as a complex combinatorial problem that could be tackled using Heuristic Search (HS) techniques. These techniques can intelligently explore a big search space (space of possible solutions) for a problem and find an acceptable approximate solution. Varieties of HS techniques exist in the literature. In our work, we use the Non-dominated Sorting Genetic Algorithm (NSGA-II), which is an improved version of a classic Genetic Algorithm (GA). NSGA-II is a multi-objective technique that aims at finding the best compromise between objectives (Pareto Front). The application of NSGA-II to a specific problem (i.e., the requirements traceability recovery problem in our context) requires the accomplishment of the following five elements: 1. Representation of the individuals (vector, tree, etc.), 2. Definition of the evaluation function (fitness function), 3. Selection of the (best) individuals to transmit from one generation to another, 4. Creation of new individuals using genetic operators (crossover and mutation) to explore the search space, 5. Generation of a new population using the selected individuals and the newly created individuals. The proposed approach takes as input a software system, a set of requirements, and the maintenance history of the software system and produces as output the trace links, i.e., the artifacts (classes, methods, etc.) related to each requirement. Three objectives are defined to support the NSGA-II: (1) semantic similarity and (2) Recency of Change (RC), and (3) Frequency of Change (FC). We used the cosine of the angle between the vector that represents the requirement and the vector that represents the software element to measure the semantic similarity. To calculate the RC measure, we used the information extracted from the history of change accumulated during the maintenance process of the software system. The intuition behind introducing the RC measure is that artifacts (classes, methods, etc.) that changed more recently than others are more likely to change now, i.e., are related to the new requirements at hand. To calculate the FC measure, we used the information extracted from the history of change accumulated during the maintenance process of the software system. The intuition behind introducing the FC measure is that artifacts (classes, methods, etc.) that change more frequently than others are more likely to change now, i.e., are related to the new requirement at hand. Each solution consists of assigning each requirement to one or many artifacts (classes, methods) of the system. The solution should maximize as much as possible the three objectives mentioned before. Experiments have been conducted on three open source systems in order to evaluate the approach. The obtained results confirm the effectiveness of the approach in correctly generating the traces between the requirements and classes in the source code with a precision of 91% on average and a recall of 89% on average. We also compared our results to those obtained by two recent works. We noticed that our approach outperforms the two other approaches and has higher average for precision and recall for all three projects.

Technology always seeks to improve the little details in our lives for a faster and a more efficient life-pace. One of these little problems we face in our daily lives is finding relevant events. For example you visit a place like Katara with your kids and you spend your time in vain looking for a fun event, and after you leave the venue a friend of yours tells you about this interesting “Henna and face painting workshop organized in building 25.”. To solve this problem we propose QEvents. QEvents is a platform that provides users with real-time recommendations about events happening around their location and that best match their preferences. QEvents renders the events in a map-centric dashboard to allow easy browsing and user-friendly interactions. QEvents continuously listens to online channels that broadcast information about events taking place in Qatar, including specialized websites (e.g. eventsdoha.com), social media (e.g. Twitter), and news (e.g. dohanews.com). The main challenge QEvents strives to solve is how to extract important features such as title, location, and time from free text describing the events. We will show in this paper how one could leverage existing technologies such as Topic modeling, Named Entity Recognition, and advanced text parsing to transform a plain event listing website into a dynamic and alive service capable of recognizing events’ location, title, category, as well as starting and ending time, and nicely rendering them in a map-centric visualization allowing a more natural exploration.

Tremendous interest in visualizing massive datasets has promoted tiled-display wall systems that offer an immersive and collaborative environment with an extremely high-resolution. To achieve an efficient visualization, the rendering process should be parallelized and distributed among multiple nodes. The Data Observatory at Imperial College London has a unique setup consisting of 64 screens powered by 32 machines providing a resolution of over 130 megapixels. Various applications have been developed to achieve a high-performance visualization by implementing parallel rendering techniques and incorporating distributed rendering frameworks. ParaView is one such application that targets the visualization of scientific datasets by taking computing efficiency into consideration. The main objective of this project is to leverage the potential of the Data Observatory and ParaView regarding visualization by fostering data exploration, analysis, and collaboration through a scalable and high-performance approach. The primary concept is to configure ParaView on a distributed clustered network and associate the appropriate view for each screen by controlling ParaView's virtual camera. The interaction events with the application should be broadcasted to all connected nodes in the cluster to update their views accordingly. The major challenges of such implementations are synchronizing the rendering across all screens, maintaining data coherency, and managing data partitioning. Moreover, the project is aimed at evaluating the effectiveness of large display systems compared to typical desktop screens. This has been achieved by conducting two quantitative studies assessing the individual and collaborative task performance. The first task was designed to investigate the mental rotation of individuals by displaying a pair of a 3D model, as proposed by Shepard-Metzler, on the screen with different orientations. Then, the participant was asked if both models were the same or mirrored. This would lead to evaluate the individual task performance by studying the ability to recognize the orientation change in 3D objects. It consisted of two levels: easy and hard. For the easy level, the second model was rotated for a maximum angle of 30 on two axes. In contrast, the hard level had no limitation on the angle of rotation. The second task was developed specifically for ParaView to assess the collaboration aspect. The participants had to use the basic navigational operations to find hidden needles injected in a 3D brain model in 90 seconds. In both tasks, the time taken to complete the task and the correctness were measured in two environments: 1) the Data Observatory, and 2) a simple desktop screen. The average correct responses of the mental rotation task have been calculated for all participants. It has been shown that the number of correct answers in the Data Observatory is significantly higher than on the desktop screen despite the amount of rotation. The participants could better distinguish mirrored objects from similar ones in the Data Observatory with a percentage of 86.7% and 73.3% in easy and hard levels, respectively. However, on the typical desktop screen, participants correctly answered less than half of the hard level questions. This indicates immersive large display environments provides a better representation and depth perception of 3D objects. Thus, improving the task performance of visualizing 3D scenes in fields that require the ability to detect variations in the position or orientation. Overall, the average completion time of both displays in the easy task is relatively the same. In contrast, the participants required a longer time to complete the hard task in the Data Observatory. This could be because of the large display space, which occupies a wide visual field, thus providing an opportunity to the viewers to ponder and think about the right answer. In the collaborative search task, the participants found all the hidden needles within the time limitation in the Data Observatory. The fastest group completed the task in 36 seconds while the longest recorded time was around one minute and 12 seconds. However, on the desktop screen, all participants consumed the full 90 seconds. In the small screen environment, the mean of the correct responses is estimated as 55%. The maximum number of needles found was 3 out of 4, which was achieved by only one group. To evaluate the overall efficiency of the Data Observatory, the one-way ANOVA test was used to find significant effects regarding the correctness of both tasks. The completion time was discarded from this analysis because of the differences in the tasks’ nature. The ANOVA revealed a significant impact of the display type in the number of correct responses, F1,48 = 10.517, p < 0.002. This indicates participants performed better in the Data Observatory in contrast to the simple desktop screen. Therefore, these results support the hypothesis of the effectiveness of large displays in improving the task performance and collaborative activities in terms of accuracy. The integration of both system solutions provides a novel approach to visualize the enormous amount of data generated from complex scientific computing. This adds great value to researchers and scientists to analyze, discuss, and discover the underlying behavior of certain phenomena.

Structural health Monitoring (SHM) is a very critical component for sustainable civil and mechanical structures in modern urban settings. The sky-scrappers and huge bridges in modern metropolis today are essential aspects of the prosperity and development of a country but at the same time they present a great challenge in terms of maintaining and sustaining the structures in a good health. Due to the complex designs of these structures, it is typically very dangerous to do SHM tasks through human personnel. Deployment of a monitoring team with various forms of equipment and scaffolding accompanied with their hoisting machines becomes extremely exorbitant for the maintenance and planning of the structures causing unnecessary cost-spill on other areas of the available budget. For most of the metallic structures, a fast method of scanning an area more closely is the Magnetic Flux Leakage (MFL) based defect detection. The MFL is considered the most economical approach for inspecting the metallic structures. Traditionally a hand-held device is used for performing the MFL inspection. In this paper, an autonomous MFL inspection robot has been presented which is small, flexible and remotely accessible. The robot is constructed with an Aluminum chassis, driven by two servomotors and holds a stack of very powerful Neodymium magnets to produce the required magnetic circuit. As the robot moves on a metallic surface, the magnetic circuit produces a layered magnetic field just under the scanning probe. The probe is composed of several Hall-effect sensors to detect any leakage in the magnetic circuit, which happens due to abnormality in the surface, thus detecting an anomaly. In this paper, a coordinated robotic inspection system has been proposed that utilizes a set of drones with one positioning robotic crawler platform with additional load hoisting capabilities that are utilized in order to position a specific defect-locating probe on the building under scan. Proposed methodology can play a vital role in SHM since it is capable of scanning a specific area and transmit back the results in a shorter time with a very safe mode of operation. This method is more reliable as compared to fixed sensors that focus a particular area of the structure only. Design for SHM robot involves intelligent integration of navigation system comprising of crucial parts that act as its backbone and assist the robot to work autonomously. These parts include GPS module, compass, range sensor, Infrared (IR) sensor along with MFL probe and winch setup and powerful PMDC Servo Motor controller (MC 160) used to drive two (2) powerful motors. The MC160 brushed Motor Controller proves to be a perfect platform for controlling Brushed DC motors. The controller consists of two power drivers in addition to OSMC connector for a third power driver (winch motor control). All these things add extra degrees of freedom to the robotic system for SHM. Novelty of the methodology is that the robot's program logic is not fixed. It is flexible in terms of path following. It has ability to detect an obstacle while it is on its way to scan the building. It not only detects obstacle but also changes its course and automatically adopts new route to the target destination. Such an autonomous robotic system can play a vital role in Structural Health Monitoring (SHM) in contrast to manual inspection eliminating the need of physical presence of human in severe weather conditions. The presented methodology is condition based in contrast to schedule-based approach. Core scan is easily done and robot is reconfigurable in a sense that it automatically changes its course to adopt to rough terrain and avoids obstacles on its way. Easy deployment makes robot an excellent choice for SHM with minimum cost and enhanced flexibility. Proposed robotic system can perform a coarse level of scan of a tall building using drones and the probe deployment robots (PDR). The drones provide a rough estimate of the location of possible defect or abnormality and PDR inspects the anomaly more closely. In addition, the coarse information about a possible defect can also help in deploying other means of inspection in a much lower cost since the whole structure needs not to be inspected.

Obesity is one of the major health risk factors behind the rise of non-communicable conditions. Understanding the factors influencing obesity is very complex since there are many variables that can affect the health behaviors leading to it. Nowadays, multiple data sources can be used to study health behaviors, such as wearable sensors for physical activity and sleep, social media, mobile and health data. In this paper we describe the design of a dashboard for the visualization of actigraphy and biometric data from a childhood obesity camp in Qatar. This dashboard allows quantitative discoveries that can be used to guide patient behavior and orient qualitative research. CONTEXT Childhood obesity is a growing epidemic, and with technological advancements, new tools can be used to monitor and analyze lifestyle factors leading to obesity, which in turn can help in timely health behavior modifications. In this paper we present a tool for visualization of personal health data, which can assist healthcare professionals in designing personalized interventions for improving health. The data used for the tool was collected as part of a research project called «Adaptive Cognitive Behavioral Approach to Addressing Overweight and Obesity among Qatari Youth» (ICAN). The ICAN project was funded by the Qatar National Research Fund (a member of Qatar Foundation) under project number NPRP X- 036- 3–013. The participants in the study were involved in activities aimed at improving their health behavior and losing weight. All participants and their parents/guardians provided informed consent prior to participation. Data from various sources (social media, mobile, wearables and health records) were collected from subjects and linked using a unique subject identifier. These datasets provided what we have defined as a 360-degree Quantified Self (360QS) view of individuals We have focused on the visualization of the biometrics and physical activity data. We proposed different visualization techniques to analyze the activity patterns of participants in the obesity trial. Our dashboard is designed to compare data across time, and among reference individuals and groups. DATA FROM OBESE CHILDREN Biometric data were measured periodically and included height, weight and the derived body-mass index (BMI), body fat percentage, waist circumference and blood pressure for each individual. Physical activity data was collected via accelerometers. The raw signals have been quantized into four activity levels: sedentary, light, moderate and vigorous, using a human activity recognition algorithm. INTERACTIVE ANALYTIC DASHBOARD The objective of the dashboard is to provide an overview of the actigraphy data and enable primary data exploration by an expert user. In a Control Panel, drop down menus enable selecting two subjects or groups of subjects to be compared based on identifiers and gender. During data collection, some devices were not worn at all times; hence they recorded long periods of «sedentary» activity. The user can use check-boxes to select the biometrics she wants to compare (e.g., BMI and Body Fat Percentage). A Visualization Panel shows both selected (groups of) subjects as bar charts indicating the hours of activity with activity level breakdown per day through time. The color legend of the bars is shown in the control panel: reddish colors for moderate (light red) and vigorous (dark red) activity levels, and bluish colors for light (light blue) and sedentary (dark blue) activity levels. The user can select a time window by brushing horizontally on the time range to zoom in or out. Two line chart show the biometrics selected in the control panel, with line colors corresponding to the selected (groups of) subjects. The average activity breakdown by activity levels is also displayed for weekdays and for weekend days. QUANTITATIVE ANALYSIS AND SUPPORT Thanks to our dashboard we can easily identify trends in biometrics, and compare activity level during week days and week-end days to support lifestyle recommendation. CONCLUSION This interface is a tool to give primary overview of the data likely to orient more detailed analysis. For instance, a more in-depth study of the relation between sleep duration and BMI could be conducted. Another outcome related to the experimental setup would consist of recommending biometrics to be measured more often, or to find incentives for subjects to wear the devices more consistently. A health expert could also provide the subject with a target status (e.g. weight) to compare and converge to, along with recommendations about the activities he/she should improve: e.g., go to bed earlier, wake up earlier during weekend, have more vigorous activities during afternoon, etc. Other available tools, such as the Fitbit dashboard (Fitbit Inc, USA), do not give detailed activity levels across time nor comparison with reference individual. Our next steps include performing qualitative evaluation of our dashboard and improvements based on the end users» feedback.