Lots of ideas came up to surface, making the whole project analysis very interesting. Finally we managed to agree upon a more concrete idea that is actually what we have already mentioned in one of our previous posts – but now we have more details at hand.
The main idea is to develop a real-time social aggregator in which the user will be given the opportunity to search for a set of keywords across the social networks that are supported in our application. We separated the application’s features into three categories: core, extended, advanced features.
- Social Networks
Our application will initially support Twitter, Facebook and LinkedIn.
- Login System
In order to use the services that our application provides, the user has to log in on our website. We will be supporting OpenID, as well as creating an entirely new account on our website. In both cases, the user will be given the option to link his/her account with the social networks that we support (the user has to have active accounts on the supported social networks).
- Search System
The application will feature a Google-like appearance regarding its user interface (extremely simple to use and practical), which will be available in two modes: simple or advanced. Although the design of the search input will be simple, the mechanism itself will be very sophisticated, thus resulting in high quality information analysis.
- Tag Cloud
There will be an interactive tag cloud of the most popular search keywords so that a user can click on a keyword and automatically search.
- Results Analysis
The results will be analysed in many possible and interesting ways so that the user can interact with our application and change the way the results are ranked and appeared. We will support results analysis based on: quantity (number of posts/mentions/likes/comments/etc.), date (old/new), social network (apply filters), and area of poster (location-awareness if applicable).
- Advanced Search
Through the advanced search, the user will be able to select a specific date/time interval, geographic location, social network connection type (friend/friend of a friend/anybody), as well as filter results from a specific social network.
For each searched keyword, a graph will be featuring (see the Mockups) the number of posts (tweets, wall posts, etc.) in which the keyword was found along different social networks and across a timespan.
- Search History
The application will keep a search history for statistical purposes – generating tag clouds, graphs, charts, etc.
- Determine Results Quality
We were thinking that it would be really interesting if the application could determine whether the meaning of a result is positive or negative. Our application will support a way of categorising an information as positive or negative, whether it will be a semantic text analysis or a simple keyword search.
- Suggest “expert” or “amateur” users
Another interesting feature would be to further analyse the search results and measure the level of relevance a user has to do with the topic (expert, amateur, etc.). Then we can assume that an expert’s opinion on a subject weights more than an amateur’s opinion.