2,798
edits
Changes
Jump to navigation
Jump to search
Created page with "{{McNair Projects |Project Title=Twitterverse Exploration (Tool) |Topic Area=Resources and Tools |Owner=Gunny Liu |Start Term=Summer 2016 |Status=Active |Deliverable=Tool |Aud..."
{{McNair Projects
|Project Title=Twitterverse Exploration (Tool)
|Topic Area=Resources and Tools
|Owner=Gunny Liu
|Start Term=Summer 2016
|Status=Active
|Deliverable=Tool
|Audience=McNair Staff
|Keywords=Twitter, Visualization, Analysis, Natural Language Processing, Graph, Social Network Analysis
|Primary Billing=AccNBER01
}}
==Exploration Notes==
===NodeXL===
*Summary: Comprehensive in its application of formal graph theory and highly customizable, NodeXL is a MSExcel-based tool that pulls queries tweet data and create graphs populated by nodes (@twitterhandles) and edges (follower/following relationship, mentions, replies, favroites, etc).
*Operates on Twitter's Streaming API, requires user authentication
*Demo: In the following test case done by www.pewinternet.org, where user attempted to graph the community activity regarding the topic "pew internet", he entered a list of search strings all including the keywords "pew internet" over a fixed period of 58 days and some misc hours. His edges are created for each mention and reply that appeared in the time bracket. His edge colors and widths are proportional to the number of mentions/replies that occurred between two nodes (users). The color and transparency of his nodes are related to follower values, i.e. how many followers does each node have..
*Thoughts: In the case of McNair, this can is directly transferrable:
**Identify trending hashtags of technologies and events, and use them as query inputs - this would give us an idea of the parties involved and directly related to an event
**Identify trending mentions of new startups, and use them as query inputs - this would tell us what parties are involved with the founding, funding or IPO/Acq of these startups
**More importantly, NodeXL generates important graph metrics such as:
***Centrality (Eigenvector centrality, Closeness centrality, Betweenness centrality) - with betweenness centrality being the most obviously useful in telling us which nodes connect subgraphs together
***Clustering coefficient
TBC
|Project Title=Twitterverse Exploration (Tool)
|Topic Area=Resources and Tools
|Owner=Gunny Liu
|Start Term=Summer 2016
|Status=Active
|Deliverable=Tool
|Audience=McNair Staff
|Keywords=Twitter, Visualization, Analysis, Natural Language Processing, Graph, Social Network Analysis
|Primary Billing=AccNBER01
}}
==Exploration Notes==
===NodeXL===
*Summary: Comprehensive in its application of formal graph theory and highly customizable, NodeXL is a MSExcel-based tool that pulls queries tweet data and create graphs populated by nodes (@twitterhandles) and edges (follower/following relationship, mentions, replies, favroites, etc).
*Operates on Twitter's Streaming API, requires user authentication
*Demo: In the following test case done by www.pewinternet.org, where user attempted to graph the community activity regarding the topic "pew internet", he entered a list of search strings all including the keywords "pew internet" over a fixed period of 58 days and some misc hours. His edges are created for each mention and reply that appeared in the time bracket. His edge colors and widths are proportional to the number of mentions/replies that occurred between two nodes (users). The color and transparency of his nodes are related to follower values, i.e. how many followers does each node have..
*Thoughts: In the case of McNair, this can is directly transferrable:
**Identify trending hashtags of technologies and events, and use them as query inputs - this would give us an idea of the parties involved and directly related to an event
**Identify trending mentions of new startups, and use them as query inputs - this would tell us what parties are involved with the founding, funding or IPO/Acq of these startups
**More importantly, NodeXL generates important graph metrics such as:
***Centrality (Eigenvector centrality, Closeness centrality, Betweenness centrality) - with betweenness centrality being the most obviously useful in telling us which nodes connect subgraphs together
***Clustering coefficient
TBC
