Facilitating the Spread of Knowledge and Innovation in Professional Software Development

Write for InfoQ


Choose your language

InfoQ Homepage News Making Sense of the Social Web with Microsoft Social Analytics (Vancouver)

Making Sense of the Social Web with Microsoft Social Analytics (Vancouver)

This item in japanese

Microsoft is making available a cloud service called Social Analytics for users interested in analyzing Twitter, Facebook, Blogger, YouTube, etc. in order to get insight on the trends on the social web.

Formerly known as Project Vancouver, a Microsoft project code-name Social Analytics is a cloud service now available on SQL Azure Labs for businesses interested in integrating results of the social web analysis into their applications. With Social Analytics, users can retrieve public data from various sources including Twitter, Facebook, Wordpress, Stack Overflow, Blogger, You Tube, and have it analyzed in order to get insight in what is currently going on various social networks. This helps companies getting feedback on users’ reaction to some of their business decisions or some products, being able to better react to what is going on on the social web, the service being recommended to “connect with customers and respond quickly to customer concerns or comments.” It is not clear if the analyzed sources are limited to those chosen by Microsoft or customers can select their own sources including forums where their users exchange ideas about certain products.

Microsoft provides two ways to consume Social Analytics services: a proof of concept tool called the Engagement Client and a RESTful OData-based API. The Engagement Client is a Silverlight application that collects tweets on two topics: Windows 8 and Bill Gates, users having the option to choose their own topic in a future release of the tool. While the tool should also display Facebook comments on the respective topics, for some reason it displayed only tweets in our tests.

The Engagement Client shows how Social Analytics ranks topic-filtered tweets based on their activity, associating a positive/negative/neutral feeling to tweets after performing language analysis, and letting the user filter tweets based on the question/not-a-question character of the message, its importance or its notification status. The tool groups tweets in conversations and lets the user interact with them by re-tweeting or sending a reply. Following is the snapshot of a “negative” tweet on the topic “Windows 8” pointing to a post entitled “How To Disable The Metro UI In Windows 8?”, followed by a positive one linking to another post called “Top 10 Best Features of Windows 8”:


Most posts are considered neutral including the following post which was the top one with 134 retweets at the time we conducted tests and it is somewhat positive actually, showing that the analytics engine needs more tuning : “Microsoft explains how Windows 8's Live Tiles won't kill your battery”:


The tweet analysis includes a graph indicating the activity of a topic over time:


The most important part of the service is its API, enabling users to:

Create clients that have the ability to respond to conversations of interest (reply on Twitter, retweet on Twitter) using OData service operations.

Create clients that programmatically mark content items for later action or analysis using the claim, importance, or annotation OData service operations.

Create clients that access all of the social activity information that an end user of the Microsoft Codename “Social Analytics” application has access to. Access data directly from any of the member resources defined in the OData Collections below.

Client Analytics is currently an Azure Labs project delivered through the DataMarket on the Windows Azure Marketplace. The Engagement Client has some bugs and missing features, and perhaps the API is still under development, but the project is an interesting one worthy to be watched how it goes over time. Access to the Engagement Client needs approval.

Rate this Article