Liked: Information Filter Based on the friendfeed API

Zheng @ playlist liked
20090208

Juwang.com has been searching for various methods to show you the best content on social media.

So far, we have providedSocial
Recommendation (Social Recommendation)Social
Dialogue (Social Dialogue)

AndLiked
These three ways to explore the social media world.

Liked is a mashup application based on the friendfeed API.

 

Preface

Zhou weimingProgramThe members are divided into 10 floors, including cainiao, prawns, cattle, experts, scholars, masters, scientists, and scientists.

So if the comparison divides the website into several floors, which floor does friendfeed and Twitter belong? The two are out of the tool type, browser type,CommunityType websites have been upgraded to the platform level, although
S3 and Google App Engine are two or two different platforms, but they cannot be described as friendfeed by means of aggregation or Lifestream, just as you cannot use miniblog to cover
Same as Twitter.

Since friendfeed launched a new search version a few days ago, it is equivalent to converting it from a publisher to a super publisher, and its aggregated data and internal user behavior data are all open, you can easily get it through the friendfeed API.

Among them, my most concern is its user like and comment behavior.

Like indicates that you like this item (links, images, videos, and tweet) by clicking "like/like" after you see an aggregation item ). Now, you can call FF API
Obtain this behavior, such as search:

Service: googlereader, delicious, diigo, Digg, internal
Likes: 2

This means that in the service types supported by these ff, searching for items that are liked at least twice has become social.

Comparison of Douban APIs
Ff api and Twitter API can explore more applications that are different from the original user level. For some examples, the Twitter API can easily make memetracker
(Hotspot tracking and reflection) applications, such as: Hotspot.

I don't know if the fresh fruit API to be released will be better than Douban.

 

Information Filtering Based on friendfeed likes data

Since we can

Service: googlereader, delicious, diigo, Digg, internal
Likes: 2

Service: Flickr, Picasa likes: 1 Comments: 1

Find the favorite buzz items (News, websites,Article, Posts, etc.), similar can also be popular
Videos, images, tweets, and blogs, as long as you control the list of information publishers, you can easily get a batch of Quality Assurance Information. Of course, similar items should be merged to avoid duplication of information.

How can I "Control Information publishers "?

The Big Bang
In theory, Penny and Sheldon discuss whether she and Leonard are suitable for dating. Sheldon uses the "schörör's cat" analogy, that is, observation behaviors will affect the observation results, before observation, the cat is in a dead and active state. The cat's death and the cat's life depend on observation. Penny and Leonard are the stuff in the box, so Sheldon cannot give
Penny's suggestion. Therefore, I will not talk about how to control the information publisher due to the uncertainty theorem. Because the users who make the like action, the users who aggregate information to ff, and those who read and play together Sr and liked are in the same box.

 

Following the e-Example of Flickr

I like this feature very much: caffe, discovering interesting photos of the last seven days on flickrland, and having fun at random. Of course, its mining video
It is also good, but relatively speaking, it is more intuitive to discover interesting photos and whether the content is more open to appetite.

I often want to warn myself that every time I refresh javase, I only refresh it five times to avoid wasting my time on it.

So,

Get favorite items by using the friendfeed API,

Following the E idea of Flickr, mashup came up with this prototype:

Liked

By default, interesting information in the last seven days is returned randomly, and the following categories are implemented:

• Top stories
• Popular videos
• Popular images
• Popular blogs
• Popular tweets

 

Summary

Obviously, liked is just a prototype of information filter.

Its information magnitude and level of excellence are not comparable to those of playmates Sr, because
Friendfeed is just a geeker in China and is not as common (relative) as Twitter and googlereader ).

Not long ago, GFW was also focusing on ff. Although FF subsequently introduced support for HTTPS access, maybe liked just wanted to prove it or could only prove it, A platform-level, fully open data social website is full of imagination and manipulation.

If you have an idea, please do it.

 

Appendix:

Http://like.ju690.com /? V = Rand & n = 16 & D = 7

URL parameter prompt:

D: represents the number of days. The maximum value is 90;

N:
The maximum number of items returned on a page is 100. This number refers to the number of information in each category (story, video, image, blog, and tweet), rather than the total number of information.

Appendix B:

Social
Recommendation (Social Recommendation)

Social
Dialogue (Social Dialogue)

Social
Liked