Social media networks are widespread on the internet. However, this experience is good for some people but very bad for others. The bad experiences often lead to hightened states of depression. Social media invariably is all about popularity. But, this popularity is also overshadowed by not only influence but also the level of negative sentiments one can receive from people. In most cases, the person that subscribes or follows a person is totaly a stranger. This strangeness and unfamilarity of people leads to a very cold and defensive state of interaction especially among women. Removing someone from your follower/subscriber count may mean nothing for one person while could mean the world to another. In other cases, not getting reply from a person can be quite an issue for others. Other cases might involve blocking which the person might take quite personally. In general, people of celebrity status have had it quite easy as they already can gain plenty of followers from just influence or even hiring a separate marketing agency to manage their social accounts. However, other people likely would have to work towards it. People often may only reply to you based on your popularity as that would increase their follower count. Social media also seems to be a network of hierarchies. In many cases, it reflects the way people climb the social ladder, in associating with people who are more popular or influencial. When people get a very high follower or subscriber count they also have a tendency of becoming quite bigheaded and proud of their achievement. You never really know whether you are talking to a human or a bot on social media. And, whether someone's post is a scheduled post or something they directly replied to. In fairness, this could also be a reason why so many people are less empathetic. Social media as a result tends to be more about playing the game. It also seems to be a very cold place to hang out. Some people obviously don't care for popularity, influence, nor take the whole experience that seriously and this likely also negatively impacts their experience. Often the way the person looks also effects their popularity. Social media networks tend to be a breeding ground for shallow people, like an ego network. Recommendations are also geared towards popularity of content which further compounds the biases. No doubt social media can be addictive. But, it has also become a battle ground for people who want to display their frustrations, their hypocrisy, and be who they want to be outside of the confines of the real world. This often leads to some people being very unempathetic towards the people they interact and communicate, often with an unconscious bias. In many respects, social media reflects the real world, only worse. It also can be an opportunity to take a glimpse through the looking glass as to the reality of character and ideological mindset of people, especially as so many feel they can say whatever they like and treat people however they like without much regard for consequence. They are also great for mining data and analyzing human behavior. Take a step away from social media and see how the world suddenly feels simpler, more productive, likely less stressful, and frustrating.
Showing posts with label social media analytics. Show all posts
Showing posts with label social media analytics. Show all posts
23 January 2023
Social Media Culture
7 May 2019
25 December 2018
19 May 2017
Social Media Evolution
From Moments to Memories
From Images & Videos to Dynamic Stories
From Static Content to Reactions & Live In-Moment Feeds
From Messages to Communities
From Social Networks to Knowledge & Sentiment Search
...
From Images & Videos to Dynamic Stories
From Static Content to Reactions & Live In-Moment Feeds
From Messages to Communities
From Social Networks to Knowledge & Sentiment Search
...
11 May 2017
8 May 2017
Social Network Metrics
1. Engagement Metrics
3. Reach
4. Impression
- Amplification Metrics
- Applause Metrics
- Conversation Rate
3. Reach
4. Impression
- Total Audience
- Number of Unique Users
- Number of Active/Passive Users
6 May 2017
Sentiment Analysis Services
- Subjectivity Classification
- Polarity Classification
- Opinion Summarization
- Opinion Visualization
- Sarcasm Detection
- Irony Detection
- Entity, Opinion Holder, Time Extraction
- Coreference Resolution
- Word Sense Disambiguation
- Sentiment Lexicon Generation
- Opinion Search & Retrieval
- Opinion Spam Detection
- Cross-Language Sentiment Classification
- Cross-Domain Sentiment Classification
- Sentiment Rating Prediction
- Contextual Polarity Disambiguation
- Aspect-Based Sentiment Classification
- Sentiment Risk/Threat Classification
- Emotion Detection
- Mood Detection
- In-Image Sentiment Detection
- In-Video Sentiment Detection
- In-Text Sentiment Detection
- In-RichText Sentiment Detection
- In-Speech Sentiment Detection
- In-Audio Sentiment Detection
- Implied Sentiment Detection
- Symbolic Sentiment Detection
- Profanity Detection
- Intent Detection
- Signal Sentiment Detection
- Fake Sentiment Detection
- Troll Detection
- Tip Detection
- Suggestion Detection
- Advice Detection
- Wish Detection
- Review Detection
16 February 2017
2 February 2017
Text-Driven Forecasting
Text-Driven Forecasting is about building systems that are able to predict on the future by analyzing collection of a body of natural language documents. Often they predict numeric quantities about a certain event based on various textual sources/feeds (e.g. news, twitter, facebook, polling data, opinion blogs, financial reports, amazon reviews, economics data, etc) as input and gather information gain from aspects of sentiment analysis and subjectivity. Machine Learning algorithms that can be applied to such a domain can range from regression, deep learning, decision trees, and others.
Examples:
Predicting movie reviews using social media
Predicting opinion polls using social media
Predicting stock volatility using financial data
Predicting government elections and referendums
Predicting product sales using social media
Predicting property prices in the future
Predicting risk of a potential course of action or decision
smith whitepaper
Related Courses & Resources:
Priberam Labs
Social Media Analysis & Computational Social Science
Natural Language Processing & Social Interaction
Computational Social Science
Social & Information Network Analysis
Text as Data
NLP for Social Science
Computational Social Science
Computational Linguistics / Computational Social Science
Predicting Economic Indicators from Web Text Using Sentiment Composition
Making Predictions with Textual Contents
12 October 2016
10 June 2016
Brexit Pipeline
Studying Sentiment Analysis in context of Brexit (EU Referendum) is currently an intensive area as the polling stations will very soon be active for voters. Input sources from social media and news feeds can be a focal point for storytelling about the various events. Social media and news feeds can be utilized in form of stream processing which can then be used for machine learning analysis and then indexed for summary into Elasticsearch. A sample workflow example is provided below. Reader will take notice that the sample workflow is also supported as an example for learning Apache Flink. The workflow can be modified as required for example, one could use a Redis cache layer between the machine learning process and Elasticsearch. Also, could extend with an NLP pipeline (Gate/UIMA) or simply OpenNLP/CoreNLP for extracting information. One could even replace Apache Flink with Spark or GraphLab. Alternatively, one could even replace Kafka with Kinesis and simply apply the AWS data pipeline. Also, the data sources can be stored using S3. Furthermore, one could even use DL4J with Spark on ElasticMapReduce to apply Deep Learning approach in form of convolutional neural network model. Although, Python developers may be more inclined to use Theano, TensorFlow and possibly RabbitMQ. For a graph representation one could use Titan, GraphX, Elasticsearch Graph, Cayley, PowerGraph, Gelly, among others. As one can see there are several ways of implementing a solution on a case-by-case basis to translate the requirements of stories. However, prototype in small is always the best way to go before scaling out incrementally i.e fail fast.
Input->Kafka->ApacheFlink->Elasticsearch->Output
Steps:
Steps:
- Collect
- Log
- Analyze
- Serve & Store
List of Input Sources:
- Youtube
- Google+
- Tumblr
- IntenseDebate
- Disqus
- Blogger
- YouGov
- NewsFeeds (Reuters, Digg, HuffingtonPost, BBC, Guardian, Telegraph, Independent, Google News, AP, TheWeek, etc)
- EUReferendum
- EU-referendum
- Fullfact
- InFact
- UK & EU
- Breakingviews
- Pollstation
- UK Cabinet Office
- Economist
- Whatukthinks
- Wikipedia
- NCpolitics
- Survation
- ICM
- ORB
- Opinium
- BMG Research
- Ipsos-mori
- UKpollingreport
- MigrationWatch
- European Central Bank
- EconomicsNetwork
- Global Research
- ...and others
7 September 2014
Big Data Graph Processing
The web with its many hyperlinked documents is a massive graph network for interlinks. Such links provide big data complexities for processing. There are many use cases for where graph processing becomes essential from contextual ads to social network analysis to even linked data. Processing such graphs in the large still remains a challenge even with its many data forms. However, graph processing from standard graph theory and network science has provided many advances for Big Data. The functional programming approaches have also facilitated more robust solutions. In OLTP, it is about the processing low-latency of workloads for accessing small portions of graphs. In OLAP, it is about batch processing workloads for accessing large portions of graphs. A graph can be stored in a specific graph database or even a column store such as Accumulo or Cassandra. They can even be stored on the HDFS. Real-time processing of graphs is also a challenge. In general, standard NoSQL stores will be able to cope with limited lookups and small number of traversals at scale. For complex traversals over the Web of Data, it would require alternative and even combined approaches for scalable batch processing in a distributed way. The below provide some options for frameworks in the big data graph processing.
Giraph
Cassovary
Drill
Impala
JUNG
SNAP
Shark
Hama
GraphX
Titan / Faunas
GraphLab / GraphChi
24 June 2014
Titan
Titan is an interesting distributed graph database which provides for a very useful approach to harnessing connected concepts for analytics. A variety of different database backends can be utilized for reuse to facilitate storage mechanisms. It is also available as an implementation for the Tinkerpop property graphs approach. In real world scenarios, graphs can be fairly huge and managing them in a distributed context is a practical necessity especially within the open source community. Titan provides not only an option for analytics but also a total semantic graph alternative to standard native triplestores with flexibility for scaling and replication. It also provides integration support for many third-party libraries and frameworks. There is also a distributed graph processing integration with Faunus that uses the Gremlin query language as well as connects to Hadoop.
 |
| Titan Stress Poster |
Titan was originally released by Aurelius and was later bought by Datastax. The open source development work later slowed down with no further subsequent releases. JanusGraph continues the work of Titan, as a forked project, with the effort and support of the community.
Janusgraph connects past and future of Titan
Janusgraph connects past and future of Titan
15 May 2014
Hello Beautiful
The Selfridges campaign of 'Hello Beautiful' comes across a bit devilish in the way it mirrors and resembles the symbolic representations of the illuminati. Perhaps, just an oversight in representation. But, if one notices the words and symbols it almost sounds like it is stating 'Hell Beautiful' with the all seeing eye symbol, illuminated to the right, which is really meant to be an 'o'. Or, maybe it is just meant to imply that beauty is in the eye of the beholder.
14 April 2014
Advertising 2014
Online advertising is an ever changing market for consumerism with constant buzz in technological innovations, often disseminated through big data processes for recommendations. Watching the digital trends is important as it provides for the most effective ways of reaching an audience with advertising campaigns. However, advertising online is directly related to the way consumers relay their interests and through the different technology drivers. The shifts in trends are often the big indicators for launching advertising campaigns for mass digital appeal.
The following are a few key trends that are emerging online:
5 April 2014
Emoji
Emojis and emoticons can be extremely useful in capturing subliminal messages and moods within the social context of conversations. An almost perceived sentiment is captured with such simplicity can give added information to a document with actionable intent. However, they may be difficult to define in metadata translation unless each iconic derivation can be interpreted from their pictograph representations. And, as such most have a Unicode representation with a standard pixel grid size. These ideograms add much variety to text speak in reducing verbage but also adding valuable meaning. May be, even a semantic data could be incorporated over such emojis supported with such open sentiment dictionaries like sentiwordnet. They are portable as well as quite varied, available as plugins, website services, on desktops, and mobile phones. Symbols and signs have always held valuable cues in our society from traffic signs, health and safety, to various other domains. The semiotics of such symbols holds much value in understanding language pragmatics over the web to discern, relate, pattern recognize in their universality, and then to even catagorize with their specific distinctions. In general, such pictograph ideograms can be categorized into the level of abstractions they provide for an individual from quality of feeling, to reaction and relation, and for their explicit representations. Such interpretations can often be provided through philosophical logic influenced in part by their psychological and sociological implications for inference.
Subscribe to:
Posts
(
Atom
)