Social Media Analysis... according to Net7

Social Media Analysis

… according to @netseven_it

Introducing Net7…

•  Net7 srl is an Italian SME based in Pisa, founded in 2001.

•  Its core business is the development of software applications for the web using Open Source software, with a special focus on Semantic Web and Linked Data oriented projects.

•  Special Expertise in Content Management, E-learning, E-commerce, Social Media, Digital Humanities, Communication & Graphic Design

•  Main customers: Coop Italia, Giunti Editore, NTT Data Italy, Scuola Normale of Pisa, Naples University Federico II, Sassari Province, Unicoop Tirreno, …

•  Participation in many EU funded research projects (5 running, 2 started in 2014)

Introducing Social Proxy …

•  Social Proxy is a SaaS platform for managing Social Network accounts and performing Trend and Sentiment Analysis on Social content. That is to say…

•  Social Accounts Management : services to centrally manage posts on multiple Social Network accounts (and multiple Social Networks as well!) at the same time

•  Monitoring : ”listening” from Social Networks or web sites using parametric rules that perform searches

•  Trend analysis : identifying trends that emerge from conversations on Social Networks

•  Analytics : stats and reports with qualitative and quantitative indicators (eg most active users, number of posts published, etc); integration with Facebook Insights and Google Analytics

•  Campaign Management : grouping social network activities in campaigns to measure their effect

Some new relevant features

•  Influencers Analysis: to identify the most active and relevant users in social network conversations

•  Content editors geotargeting for Facebook: operators are allowed to manage posts on the Social Proxy that will be visible only in specific countries.

•  Demographics Analysis: to identify demographics information of the most active users (location, sex, age, etc)

•  Sentiment Analysis: to identify users’ feeling in social conversations (positive/negative). Supported languages: English, Italian – currently under development

Social Account Management

•  Editorial features of Social Proxy that allow operators to publish new posts, to manage those waiting to be published (Queued) and those already Published (not necessarily through the Social Proxy) in the company’ social places.

•  Posts can be published on multiple social places at the same time, in run-time or at a scheduled time.

•  Posting pictures and link shortening are supported.

•  Feedbacks (e.g. likes, comments, retweets, replies,…) can be managed from Social Proxy.

•  Posts can be manually categorized with a configurable taxonomy system

•  Social Proxy currently supports editorial features for:

–  Facebook (accounts, pages, groups) –  Twitter

–  YouTube (comments)

–  Instagram (like and comments)

–  LinkedIn (updates)

–  Tumblr

Monitoring

•  Monitoring is based on parametric rules that specify how posts can be fetched from social networks and web sites.

•  Social Proxy currently supports monitoring for:

–  Facebook (profiles, pages, accounts, groups, generic search, user’s timeline)

–  Twitter (search, geosearch, accounts, user’s timeline)

–  Instagram (search, geosearch, accounts, user’s timeline)

–  LinkedIn (groups, companies, members, network updates)

–  YouTube (search, channels)

–  Foursquare (tips for venues)

–  Pinterest (users, boards)

–  Feed RSS/Atom (web sites)

•  Content can be indexed on a local search engine (based on Apache Solr), to perform higher precision filtering

•  An extension to crawl generic, predefined, web sites, is also under development

Monitoring: observing the flow of social conversations

•  Retrieved posts are shown in feeds, in a chronological order.

•  We provide a user interface based on “tabs”, that facilitates reading and comparisons of multiple feeds. It is also possible to aggregate multiple posts in a single feed.

•  Posts can be saved and classified with a taxonomy system.

Analytics (Stats)

•  Stats provide quantitative and qualitative statistics for the social accounts and social conversations I’m managing.

•  Some examples:

–  Most relevant users

–  Most commented posts

–  Most used semantic categories in the categorizations of posts

Analytics: integration with Facebook Insights

Trend Analysis

•  Trend Analysis exploits text analytics techniques (Entity Extraction) to identify concepts (and not merely keywords) in conversations. Hashtags are also identified.

•  It is possible to follow the evolution of trends in social conversations in real time, checking weekly, daily or even hourly, the most discussed concepts that emerge.

•  It is possible to visualize the geographic distribution of posts on a map.

Trend Analysis: geographic distribution of posts

Link Analysis

•  Link Analysis shows the most cited web pages in social conversations.

•  It allows to effectively discover “other places” on the web, possibly unknown beforehand, where “they talk about me”.

•  The idea: “if something is relevant it will produce an echo on the social conversations I’m observing”

Traffic Analysis

•  Trend Analysis also shows the amount of posts fetched through rules in time (Volume).

•  The evolution in volume of the most cited concepts and hashtags is also shown.

Key influencers

•  An analysis based on various strategies (*) to identify the most relevant and influential users in conversations.

•  Starting point: observing social conversations to identify interactions and behaviours.

•  It will not identify influencers only through the number of friends/followers that they have.

* Sources: Identifying influential spreaders in complex networks, M. Kitsak et al; Influence Marketing, Danny Brown, Sam Fiorella (Pearson 2013)

Competitors analysis: how-to (1)

•  With monitoring rules it is possible to check in real time the evolution of conversations about two competitors

(in the example, the two largest supermarket chains in Italy: Coop and Esselunga)

Competitors analysis: how-to (2)

•  Trend Analysis can be used to observe emerging trends in conversations about different competitors

(again in the example Coop and Esselunga)

Demographics analysis

•  More insights about users involved in social conversations:

–  Number of friends/followers

–  Number of posts and comments

–  Evolution of posts and comments

–  Gender (for Facebook Friends)

–  Age (for Facebook friends)

–  …

Sentiment

•  It quickly shows the positive, negative or neutral attitude in social posts.

•  English and Italian support

Campaign Management

•  Defining campaigns to group and plan coordinated activities on Social Networks

•  Checking the campaign through a Calendar

interface

•  Scheduling posts for automatic publication

•  Tracking the results through the integration of

Facebook Insights and Google Analytics

Live example #1: Trend Analysis on “Carta di credito”

Volume of posts and main concepts (in a week, in a day)

5 trending concepts

Main concepts “tag cloud”

Sample posts with “American Express”

Top hashtags

Relationship between concepts and hashtags

Link Analysis

Link Analysis

Info: http://www.netseven.it/servizi/social-media-analysis/

Net7 srl Pisa – Italy

http://www.netseven.it - [email protected]

@netseven_it - facebook.com/Netseven

Some of the developments of Social Proxy were co-funded through the Regione Toscana research project SenTaClAus – http://sentaclaus.netseven.it