The department of Intelligent Computer Systems at the University of Malta, is developing a flat 2D social networking site as part of the deliverables for the Euro South Hub project, which it forms part of. The first stage of this deliverable involves the representation of an online closed community of people with same interests and needs similar to existent social networking site but which will serve as the basis for further research and development in domains of knowledge mining and 3D representation. The main focus on this first basic prototype has been on the flat 2D representation and on the underlying Content Management System (CMS) that will play a crucial role in the future.
b) 2D REPRESENTATION
Presentation is always a challenge when having a significant amount of information to present to the users. Social networks themselves are a large source of information and due to this reason, it was decided that online 2D representation of the hub would be in the form of a dashboard presented in the figure below.
A number of components were chosen to represent the user information vis-à-vis the hub network through this dashboard. Below follows a list of these components and how they contribute to giving context to the said user information.
i. USER INFORMATION
This part displays the basic information about the user. The aim of this component is to allow for quick profiling of the user in question. In an effort to ease the browsing experience a small photograph of the individual is displayed and this is accompanied by the age and location of the user together with the date since the user has been
subscribed to a hub. Subsequently, a 50 word abstract about the user’s background is presented.
ii. KEYWORD CLOUD
The system runs a background process of profiling the user and extracting the user interest. Once the user interests are elicited, the system finds a set of 5 keywords that match the user best. It was decided that instead of presenting these keywords in a traditional list, the format of a tag cloud would be used. An example is shown in the figure below
iii. SKILLS GRAPH
When one browses individual in an effort to establish professional collaboration, it is often the case that the search includes metrics that would give substantial background about the skills available. This was acknowledged and the spider graph below was structured.
This graph is organized in four axes. Each axis represents a particular metric being considered for that particular skill. Each graph may therefore have more than one skill with each skill constituting of different values for the given metrics. The four metrics chosen at a preliminary stage were the following:
• Experience: This reflects the experience the user has in the skill in question.
• Passion: This metric intends to show the individual’s attitude towards that particular skill. This mainly aims to differentiate between something being conducted a hobby, work or somewhere in between.
• Reliability: When measuring individual skills, we also rely on perception. This metric would show how much other people trust the individual doing that particular skill
• Knowledge: The background the individual has about the particular skill in question is reflected in this metric. One may also include the professional qualification in such a metric.
iv. RECOMMENDED MEMBERS
The interests of the user that are concluded in the background process are also used to identify other hub members with similar interests and characteristics. These members are then presented on the same page in order to provide an easy link with other individuals who might be helpful for the person browsing the online 2D representation. This interface was kept as simple as possible by providing the name of the member together with the top interests of the same member. Each element acts as a hyperlink to that member’s dashboard.
v. RELATION NET
This interactive graph shows the relation of the user in question with the other members with similar interests and characteristics. This graph intends to provide the user browsing the dashboard with a visual representation of where the user represented dashboard stands in relation to other members. Each circle in the graph shows a level of similarity. This graph therefore helps people finding other individuals who might be related to the particular person being browsed but at the same time, shares certain similarities that might shed new light on user’s quest. The relations of the member selected in the graph are then
The map aims to present a geographic relation of the interests being generated, related landmarks and facilities and perhaps key individuals. This may be useful since at first sight, the user browsing the profile would relate topics in particular locations with individuals he might know. This establishes a way in which new links and possibilities of how the items highlighted in the map can be reached.
c) CMS JUSTIFICATION
The Euro-South Hub social network is meant to be a social network allowing users of the Social Hub to share ideas and instigate innovation between the users. For this reason it is essential for the network to be able to grow and develop as the users and their requirements change throughout time. In order to be able to provide good levels of performance and updatability it is always suggested that any new system is built up on an existing content management system; which is preferably open source and well supported. With power to further expand and change things in the future.
For this reason several common Open source systems were analyzed. Joomla & WordPress whilst very well supported and containing a large amount of extensions/plugins are very specific in themselves; that is they are built to cater for News websites and blogs respectively. The templates are handled in php, and whilst they try to use‘MCV’ architecture the data and the view are very much intertwined.
On the other hand systems like Symphony-CMS & ProcessWire enable you to define your very own structures without going into details of generating the required SQL, thus making them very flexible if they are required to be extended or improved in the future. Both systems provide a very good MVC structure however Symphony-CMS enables better understanding due to using three different languages for each level. The backend is based upon PHP – which takes care of generating XML data. This is then transformed into HTML using XSLT templates separating the view clearly from the Data.
In our integration we have also opted to use an external API to handle client-sensitive data. This is to be extracted and imported via a JSON API. Symphony CMS provide out-of the box support for data obtained from external sources into the system being XML or JSON formatted. This option could come in very handy when implementing our JSON data.
Following the above CMS analysis – and considering the expertise of the team at our disposal it was clear that using Symphony-CMS would be more beneficial. This open source project also makes use of a lot of popular web standards – such as XML, XSLT and XHTML. Whilst it is based upon Object oriented PHP5 and powered by a MySQL database.
Supporting Extensions – Additionally Symphony-CMS comes with plenty of extensions that can help make our project thicker. Such as multi-lingual support so that we can make the EuroSouth Hub available in languages other than English, say Italian and Maltese. There are also extensions built to support OAuth, the user authentication system used by our API. Finally extending it is really easy so anything which is not available by default can be added in – thus we will always have full-control of the system.
i. UNDERSTANDING THE API
Our social network will be based on the NationalField, a corporate standard for private social networks. This includes all integration of features and usability; they also provide an API for additional extensions/functionality. This API allows us to extract some features from the NationalField network and provide our very own interface for the Euro-South Hub.
First of all – api usage is done using OAuth 2.0 a standard used by Facebook and many other systems. This ensures that anyone using the API at no point in time has access to the user’s password thus increasing security when developing extensions or API. Since all data is managed by the National Field this ensures that nobody can create security flaws by obtaining user passwords.
Following authentication all requests require that an access-token is passed along each request so that the user on whose behalf the request is being made can be verified – allowing for customized data to appear on each profile. This could potentially allow us to retrieve feed data which is more related to the user in question – reflecting his subscriptions and groups. Information from the API is not only useful to display information to the user but also to potentially parse and expand upon. By for example doing data-mining and thus understanding topics the user is knowledgeable about enabling us to further customize his experience on the EuroSouth Virtual Hub.
The API also enables us to update user data, as well as to post status updates and comments. Whilst not all these features are likely to be implemented since they are also covered by the NationalField system they can be useful – especially when trying to connect users between each other.
In other words the API gives us scope to increase the functionality behind our virtual hub.
ii. DATA MINING
In this project we consider Data Mining and Artificial Intelligence to be a very important component. One of our targets is to allow users to connect to others, in order to increase innovation. The hub in itself is a place where people share ideas and expand them, unfortunately it’s not always so easy to find the right person(s) to share your ideas with. By using Data Mining and other AI techniques we plan to understand each user and individual, knowing their skills and strengths.
Once we have successfully built user profiles we can then try to match people and groups in order to allow for increased innovative value from within the group. Various studies show for example that people with knowledge in multiple areas are more likely to be innovative by merging both. Whilst not exactly the same a group with different knowledge bases can give increased overall knowledge and a better perspective on various things. This tends to give rise to better and more complete innovations.
By analysing user profiles we have to identify which combination(s) of skills work best to induct innovation in certain projects. A case in point would be suggesting a physicist/biologist to someone who wants to improve his farm watering system for increased productivity rather than a web-developer with little or no knowledge of physics and biology. This comes by understanding the domain of each user and where they are trying to apply innovation, and that step comes from data mining.
Current social networks usually suggest friends or people to network with depending on our connections, so it always suggests friends of friends. Our innovative system can suggest anything from people of the same likes, or people with complementary skills to provide help and ideas on particular projects. Thus helping expand your network far beyond people you know, and more into the realm of people that can benefit from each other.
The initial prototype together with the functionality and the CMS potential at the background, is very promising, encouraging and exciting. The development team at the department of Intelligent Computer Systems is eager to proceed and investigate new ways and novel strategies in this ground-breaking area which is gaining popularity and interest with the academic and social fora alike.