Social Network Analytics of Meetup Group
Analysis using graph theory and visualization of a network
Nashville is the most populous city of Tennessee. It is ranked as the 7th friendliest city in US. Major part of the demographic is of residents between the age group of 24–35, indicating that it would be attractive for young professionals to stay in this city. Nashville is called the Music City and has some of the best music museums. The city also offers great food and ample hiking spots more than just Smoky Mountains. This paper explores the social media platform Meetup for Nashville city.
What is Meetup?
Meetup is a social media platform for finding and building communities, and organizing get-togethers with those who share similar interests. Founded in 2002, it has become quite popular in US with 44 million active users worldwide. It was acquired by WeWork in 2017 and recently sold in 2020 to AlleyCorp, a venture fund and incubator company. About 62% of the users are from US, 18% from UK, Canada, Germany, Australia, and 20% from other countries.
The users of Meetup are organized as groups around sports, hiking, reading, arts, music, knitting, coding, crafting, and social groups and are local to cities. Anyone could be an organizer of events. Organizers need to pay a fee to run a group. Commercial activities are permitted as long as they comply with Meetup policies.
What is Social Network Analysis?
Social Network Analysis (SNA), is a branch of Network Analytics, that investigates social structures. It can be used to analyze what roles individuals play in a social group or structure. Examples of social networks include social media groups (Facebook, Pinterest, Snapchat, etc), Special Interest Groups, Club Network, Business Networks, etc. Visualization of social networks is known as a sociogram. Graph theory is used to study networks.
There are multiple use cases for visualizing social networks. For businesses, analysis could help in understanding how products could be launched and how their adoption could be influenced. We already are aware of the role influencers play in marketing a product. For government bodies, this analysis could help in dissemination of information, adoption of model social behaviors.
About the Dataset
We are using Nashville Meetup Group relationship dataset for this analysis. The dataset has 1.17M members, spread over 456 meetup groups.
Potential Goals for Social Network Analysis
Goals could include understanding the following:
- Who are the top influencers in the group?
- How does information transfer or diffusion take place?
- Who are the people who are best at transferring information in the group?
- Which Special Interest Group should an organization target if they are trying to promote a product or a service?
- Which Special Interest Groups are most common?
- Are the groups personal or random networks?
Key Metrics
- Links: Links are ties in a network.
- Density: Proportion of possible ties/links that could exists vs that actually exist.
- Components: # of unconnected/isolated subgroups within a network.
- Centrality: # of relationships of an individual (aka node)
- Closeness: Distance of path between two nodes
- Degree Centrality of Proximity: Total # of links a node has to other nodes, combined for the entire group. This helps indicate the individuals who have significant power in aiding transfer of information. Individuals with higher degrees have the ability to connect easily with a large group.
- Betweenness Centrality: Individuals/Nodes that are most important in a network based on centrality of a given node in a network. These individuals influence the transfer of information to subgroups.
- Eigenvector Centrality: Relative importance of a node with other important nodes in the network. These individuals are central to a network, have the most number of relations and influence in a network.
- Clustering Coefficient: Degree of nodes clustering or forming groups.
- Network Diameter: Largest value of all the shortest paths between any pairs of nodes, that provides an idea of how long it would take for a message, idea or information to pass through the network.
Visualization of Network
We are using Python for our analysis. Since the dataset is large, we will limit our processing to a smaller subset.
Nashville Meetup Network with 1K Members
Nashville Meetup Network 2K Members
Nashville Meetup Network with 4K Members
Nashville Meetup Network with 20K Members
Insights & Conclusion
Of the 700K population of Nashville (2020), 330K members have meetup membership. It should be noted that these insights are specific to the population of 330K members being analyzed and not representative of the entire Nashville population. From the data analysis within Kaggle, it is observed that some of the members are based in states outside Tennessee and also in cities/suburbs outside Nashville. The 330K meetup members are not constituted entirely from the Nashville population. Some groups may be organized by folks outside of Nashville, also members may be residing in a different city but might consider Nashville their hometown.
The following insights can be derived.
- Being a Music City, “Music” was expected in the top 5 meetup groups, but ranks as the 6th most popular group with 13K members.
- The largest meetup group falls under the “Outdoor and Adventure” category with 60K members. This can be explained by the number of young professionals and demographics.
- Popular outdoor activities include hiking, backpacking, paddleboarding, kayaking, and sailing.
- Tech is the 2nd largest group, followed by “Career and Business”. This explains that the young professionals are tech savvy and serious about growing in their careers and entrepreneurship endeavors.
- Top “Tech” sub-categories include generic coding groups, Python, Javascript, .NET, PHP, Wordpress, User Experience, and Ruby. This can be explained by Python being the most popular machine learning language and increased interest in Web Developer/UX Designer jobs.
- There is also a group of Women programmers with approx. 1000 members.
- The sub-categories observed in “Career and Business” include Mom entrepreneurs, after-work networking events, and women’s business network. This indicates that more and more young women are looking to grow their business and careers by connecting with other women.
- “Hobbies and Craft” and “Support” are the groups with the least number of members.
- In terms of sub-groups, though “Tech” is not the top group based on members, we can see it has 100 sub-groups based on the technology people are interested in. The same theme carries for “Career and Business” which has 96 sub-categories.
- “Socializing” ranks at #3 with the most number of sub-groups. These groups do not specify a special interest but members have signed-up based on the need for socializing and for the opportunity to connect with people in their community. This could also be explained by new people, mostly young, pouring into the city and looking for ways to connect.
It can be concluded that young professionals who have signed up on Meetup are most interested in balancing work and life and growing their careers. This is because “Tech”, “Outdoor and Adventure” and “Career and Business” rank as the top groups, with “Socializing” and “Health and Wellbeing” in the top 5 categories.
Next Steps
The dataset Meta-members.csv includes information for each member, including name, hometown, city, state, latitude, longitude. Member_id is an index column. Since this is an extremely large file, it was not analyzed as part of this project scope. This file would be helpful to:
- identify the top 5 popular,
- most connected members and see which part of the city they reside in,
- Understand which groups they belong to,
- how many groups they have organized,
- which groups are the popular members part of.
Additional analysis can be done by doing joins between each of the datasets and centrality, betweenness for the top members and groups can be explored in detail.
