Direct Social Networks Import

[admin]

This is the thread for asking more details about the Direct Social Networks Import proposal.

[kumarlavr]

Hi,
Can you please specify what type of information do we have to pull out of the apis, so that we make our project proposal more specific to the goal
Thanx
With regards,
Kumar Lav

[admin]

We don't know the user's goals, and the information depends on what social network you're crawling. In Facebook, it is more interesting to map who replies on someone's and with which frequency, than just to get the friends list. And it could evolve according to the cutting-edge related researches, or on how people use these networks.

So it's on your responsibility to provide a configurable import, and to let the user process both simple and complex queries with an easy-to-use GUI.

On each network, you must identify which data could take the role of node, what associated data attach to them, and which one could define an edge.

If you're getting data from Flickr, for example, one could:

• Set people and photos as nodes, and set comments as edges. Associated node data: # of friends for people, license and all tags for photos. Associated edge data: comment's full text.
• Set people as nodes, and identical tags on photos as edges. Associated node data: # of photos for people.
• Set set tags as nodes, and photos as people as edges. Associated node data: nothing. Associated edge data: license for photos...

And important feature is to get the timestamps when it's available: a dynamic network would then be created.

[kumarlavr]

Hi,
Can you please guide me in what all should to include in the project proposal ?
I have studied the Twitter, NYT, and Flickr apis and have created the roadmap of how to proceed with them to create the networks and queries.
Also, I have studied the Import APIs.
Do I have provide some code too for the above ?

With Regards,
Kumar Lav

[mbastian]

Hi Kumar, you don't have to provide code for your application but motivated arguments how much time it would take to get graphs from these APis would be nice. It would be helpful to make proper planning.

[kumarlavr]

Hi,
I have used the following apis from NYT Apis:
1) Movie Review
2) Best Seller
3) Time People
4) Article Search
5) Time NewsWire
as I think these are the ones out of which mesh can be developed.

Do I have to go for other Apis from NYT too ?

[admin]

No, I think these are the main APIs, and would be sufficient for the SoC.

[raj]

Hi,
I was wondering about the main agenda here - is it the API's and what it provides or is it an infrastructure that is scalable when more such social apis can be added to gephi with minimal code and ease. I was thinking that the project would require building a framework where an user would be able to select one or more social data (the source being available by click of a button on the toolbar) and create a graph based on user's requirement and interest.
e.g., There is twitter, NYT and Flickr available on the tool bar
Case 1: The user select the option twitter -> a new option opens up, which says use cross database or default(meaning same database) , the user chooses default -> a new option opens up which says, time variant or default(meaning work with the default time epoch), the user chooses default -> a window opens up that shows the parameters that are available for nodes and edges like nodes={name, follower, following} and edges={tweets, retweets, etc.} -> user selects follower as node and retweets as edges -> (optional login window when required ->) clicks "show" -> user gets the graph on gephi
Case 2: The user selects twitter -> selects cross database -> selects NYT and flickr-> "node-edge" window opens up containing node={movie_name(from NYT), name(from twitter), book(from NYT), photos(from Flickr)} and edges={tweet(twitter),review(NYT),article(NYT),photo_tag(Flickr)} -> user selects movie_name and name as nodes and review and tags as edges -> a window opens up asking for the movie name -> user inputs it say twilight -> selects "show" -> the graph shows the movie name as the node the edges emanating from it shows the tags to the nodes that shows pictures related to the movie and another set of edges contain the tweets that sink to the nodes containing name of the actors(having twitter account) of the movie.
The second case sounds elaborate but I think it is the ultimate goal.
I would like to know if I'm thinking in the right direction. It would help to make my proposal.

best,
raj

[admin]

Hi raj,

You're on the right direction!

Providing a framework for adding further social networks is very important. In fact, the main hurdle of this proposal is a good code design to let the user to make queries in an easy and flexible way, and to be able to extend it in the future.

Cross-networking is a good idea, but is not mandatory. Keep it in mind but if we do a single-network grabbing it will be cool. Another big hurdle is the GUI design, because it has be to generic some way. The student who will work on this proposal will closely draft the interface with me before coding an acceptable solution. The first users of this system will be...the sociologists and scientists involved in humanities, so the GUI must be very very user-friendly.

[cristianregep]

Hi,

From what I've read from the previous discussions in this thread I’m beginning to think that it's more important, and in the same time more difficult, to create a user friendly query environment, along with an easily expandable framework.
My question is if an application which discusses more about the GUI and the generic framework than the actual query mechanism on the API’s would be seen favorably?

Cheers,

Cristian Regep