Data from "An exploration of the Facebook social networks of smokers and non-smokers"

<b> Purpose </b>
<br>
For the purpose of informing tobacco intervention programs, this dataset was created and used to explore how online social networks of smokers differed from those of nonsmokers. The study was a secondary analysis of data collected as part of a randomized control trial conducted within Facebook. (See "Other References" in "Metadata" for parent study information.)

<p>
<br>

<b>Basic description of 4 anonymized data files of study participants.</b>
<br>
<ul>
<li> fbr_friends: Anonymized Facebook friends networks, basic ego demographics, basic ego social media activity </li>
<li> fbr_family: Anonymized Facebook family networks, basic ego demographics, basic ego social media activity </li>
<li> fbr_photos: Anonymized Facebook photo networks, basic ego demographics, basic ego social media activity </li>
<li> fbr_groups: Anonymized Facebook group networks, basic ego demographics, basic ego social media activity </li>
</ul>

<p>
Each network comprises the ego, the ego's first degree connections, and the (second degree) connections between the ego's friends.

<p>
Missing data and users who did not have friend, family, photo, or group networks were cleaned from the data beforehand.

<p>

<b>Each data file contains the following columns of data, taken with participant knowledge and consent </b>
<br>
<ol>
<li> participant_id: Nonidentifying ids assigned to different study participants. </li>
<li> is_smoker: Binary value (0,1) that takes on the value 1 if participant was a smoker and 0 otherwise. </li>
<li> gender: One of three categories: male, female, or blank, which signified Other (different from missing data). </li>
<li> country: One of four categories: Canada (ca), US (us), Mexico (mx), or Other (xx). </li>
<li> likes_count: Numeric data indicating number of Facebook likes the participant had made up to the date the data was collected. </li>
<li> wall_count: Numeric data indicating number of Facebook wall posts the participant had made up to the date the data was collected. </li>
<li> t_count_page_views: Numeric data indicating number of pages participant had visited in the UbiQUITous app up to the date the data was collected. </li>
<li> yearsOld: Numeric data indicating age in years of the participant; right censored at 90 years for data anonymity. </li>
<li> vertices: Number of people in the participant's network. </li>
<li> edges: Number of connections between people in the network. </li>
<li> density: The portion of potential connections in a network that are actual connections; a network-level metric; calculated after removing ego and isolates. </li>
<li> mean_betweenness_centrality: An average of the relative importance of all individuals within their own network; a network-level metric; calculated after removing ego and isolates. </li>
<li> transitivity: The extent to which the relationship between two nodes in a network that are connected by an edge is transitive (calculated as the number of triads divided by all possible connections); a network-level metric; calculated after removing ego and isolates. </li>
<li> mean_closeness: Average of how closely associated members are to one another; a network-level metric; calculated after removing ego and isolates. </li>
<li> isolates2: Number of individuals with no connections other than to the ego; a network-level metric. </li>
<li> diameter3: Maximum degree of separation between any two individuals in the network; a network-level metric; calculated after removing ego and isolates. </li>
<li> clusters3: Number of subnetworks; a network-level metric; calculated after removing ego and isolates. </li>
<li> communities3: Number of groups, sorted to increase dense connections within the group and decrease sparse connections outside it (i.e., to maximize modularity); a network-level metric; calculated after removing ego and isolates. </li>
<li> modularity3: The strength of division of a network into communities (calculated as the fraction of ties between community members in excess of the expected number of ties within communities if ties were random); a network-level metric. </li>
</ol>

<p>
Detailed information on network metrics in the associated manuscript:
"An exploration of the Facebook social networks of smokers and non-smokers" by Fu, L, Jacobs MA, Brookover J, Valente TW, Cobb NK, and Graham AL.