9
Build a Social Intelligence Platform Jason Xue, Director of Engineering SDL Social Intelligence

Build a Social Intelligence Platform

Embed Size (px)

DESCRIPTION

In this presentation, our Director of Engineering looks at some of the technology that makes our social intelligence software tick. There is a lot of technology that goes into analyzing data and generating predictive analytics. If you have any questions about our software, please let us know.

Citation preview

Page 1: Build a Social Intelligence Platform

Build a Social Intelligence Platform

Jason Xue, Director of Engineering SDL Social Intelligence

Page 2: Build a Social Intelligence Platform

Build a Social Intelligence Platform

Characteristics of Social Media Data

2

Big data Billions of daily posts and tweets are generated from tens of millions of

individuals across social networks and open-source channels

Heterogeneous Sound and movie files, images, documents, geo-location data, web logs,

and text strings Some data are structured but most are not Hundreds of languages and countries

Real time According to Domo, every minute:

• Twitter users average over 100,000 tweets • Facebook users post 684,478 pieces of content• Tumblr Blog owners publish 27,778 new posts

Page 3: Build a Social Intelligence Platform

Build a Platform for Social Media Analytics

Data Process flow for Social Intelligence

3

Charts for Insights

Aggregation and

Modeling

Search and Reports

Data Processing

Data Collection

Social intelligence is to gain insights for a specific brand, product, or subject matter from social media data.

Page 4: Build a Social Intelligence Platform

Build a Social Intelligence Platform

Data Process flow for Social Intelligence

4

• We collect social media data from multiple distributing channels in a real time manner, including Gnip, Sprinn3r and BoardReader.

• We perform language/country detection, duplication elimination, spam detection, blacklist/whitelist filtering and sentiment calculation. We also tag content using category rules of key words.

• We make available reports on result volume, demographics, location, and other metrics, which helps identify the right data sets for generating insights.

• The data sets for products, brands, and customer relevance are aggregated and modeled to provide a comprehensive view of predictive measures that will empower you to make the right decisions on optimizing your customer’s experience.

Page 5: Build a Social Intelligence Platform

Build a Platform for Social Media Analytics

Social Intelligence Platform Components

5

Search framework and data Customers can search by language, country, date range and key words Historic data is available back from Jan 1 2008 Data set can be downloaded and integrated with BI tools Customer data can be uploaded to perform joint analytics

Social analytics and modeling We provide a Customer Commitment Framework with four levels of

customer behavior scores and charts Modeling language for general data transformation and computation

Page 6: Build a Social Intelligence Platform

Build a Platform for Social Media Analytics

Social Intelligence Platform Components

6

Platform APIs APIs will be used for integration with customer applications APIs will include search, modeling and analytics

Application UI framework It includes reports, charts and workflows It provides social widgets that can be integrated with UI applications We have two products now - SM2 for social monitoring and The Customer

Commitment Dashboard for analytics

Page 7: Build a Social Intelligence Platform

Build a Platform for Social Media Analytics

Social Intelligence Platform Users

7

• Social Intelligence Platform can be used by market strategists, product managers and customer service support for analytics, modeling and research.

• Social Intelligence Platform can be integrated with other SDL products to better implement our global vision for customer experience management.

• Social Intelligence Platform can be integrated with third-part applications as part of their business intelligence solution.

Page 8: Build a Social Intelligence Platform

Build a Platform for Social Media Analytics

Social Intelligence Platform – Marketecture

8

`

Application UI Framework Market Strategists Customer Support

Application UI Framework

Reports Charts Social Widgets Workflows

SDL

Pro

duct

s

Third

par

ty A

pplic

ation

s

Platform API

Social Analytics and Modeling

Search Framework and Data

Product Managers

Page 9: Build a Social Intelligence Platform

Build a Platform for Social Media Analytics

Social Intelligence Platform Technologies

9

ElasticSearch – A distributed search engine that indexes all the social media data coming to the platform. The data is composed of content and meta data which are both indexed by date ranges.

Hbase - A non-relational distributed database that stores and accesses the social media data by permalinks.

Hadoop – A programming framework for large data sets that offers super computation power for searching and modeling through MapReduce jobs in a distributed environment.

SQL Server – A relational database that handles structured social media data, such as profiles.

MVC and HTML5 - A software architecture pattern and a markup language for structuring and presenting that are used to build the UI application framework.