Mapping Family Connectedness across Space and Time

Understanding the structure and evolution of family networks embedded in space and time is crucial for various fields such as disaster evacuation planning and provision of care to the elderly. Computation and visualization can potentially play a key role in analyzing and understanding such networks. We introduce a measure of family connectedness that summarizes the dynamic relationships in a family network by taking into account the distance (how far individuals live apart), time (the duration of individuals’ co-existence within a neighborhood) and the relationship (kinship or kin proximity) between each pair of individuals.

By mapping the family connectedness over a series of time intervals, the method facilitates the discovery of hot spots (hubs) where family connectedness is strong and the changing patterns of such spots across space and time. To demonstrate our approach, we use family tree data derived from published genealogies of nine families from the US North over a span of three hundred years. 

 

 

A series of demographic events (e.g., births, deaths, migrations) were coded from the genealogy, including the places where events had occurred. From these event locations and dates we can infer the migration paths of each individual in the families. 

Click the image on the right to view the animation of the migration of nine families including the Chaffee (CFE) family in the U.S.

http://129.252.37.169:8400/flowvis/trajectories/index.html

Animation of the Chaffee family migration

 
 

Methodology

Time interval                          

To allow for a temporal analysis of connectedness in a family network, one can employ a data-driven approach such as sliding windows, top-down or bottom-up segmentation algorithms to obtain time intervals. In this paper, we partition time series data into equal intervals and reflect meaningful stages of the family tree data. Because some patterns may fall between time windows and not appear, we use a sliding window approach.

 

Time duration of the co-existences

The horizontal axis illustrates time and the vertical axis represents unique locations (i.e., Loc 1, Loc 2, Loc 3, and Loc 4). An individual at a location is represented with a horizontal line with a beginning and an ending year. For example, AB1 refers to the period that AB lived at location 4 between 1672 and 1685, whereas AB2 refers to the period that AB lived at location 3 between 1685 and 1700.

Time Interval

 

Family connectedness    

Naturally, the potential for spatial interaction between individuals change across time as individuals move, new individuals are added or removed, and relationships develop and change over time. By taking into account the time-varying relationship between geographic and kin proximity between individuals, and the time duration of their coexistence, we introduce a measure of family connectedness as a proxy for potential spatial interaction.

2.1. For each time window, we derive the territory of each individual by using a geographic distance threshold around his location at the time. The figure below illustrates the individual AA’s family connections that are determined by his territory (gray circle).

2.2. We use consanguinity (Leutenegger et al. 2011) to quantify the degree of kinship (relation) between the members of a family, which is widely used in law and genetics. Figure 4 represents a family tree of four generations where A is the ancestor of all members in the family.

 

2.1 Territory of interaction

The potential connections of individual AA from the sample network given in Figure 1. The circular buffer illustrates the neighborhood of individual AA which is used to determine his/her potential connections. Nodes with labels within the neighborhood are potential connections of AA, whereas empty node symbols and labeled nodes outside the neighborhood are individuals that are not connected to AA. A subscript (e.g., AB1, AB2) for an individual indicates his/her existence at each unique location given the time interval.

Bandwidth

2.2 Kinship, consanguinity

A sample family tree with four generations that descend from the ancestor, A. The relation among two people is called lineal consanguinity if one is descendant from the other such as the son and the father (e.g., A-AA), or the grandfather (e.g., A-AAA), and so upwards in a direct ascending line. For people who descend from the same ancestor, but not from each other (e.g., cousins or uncles-nephews), the relation is called collateral consanguinity.

Family Tree

 

 

Results

We produced 29 surfaces of family connectedness each of which corresponds to a 20 year time window. Each surface was produced using a constant divergent classification scheme to enable comparison between each time window. While blue hue illustrates places with low family connectedness (i.e., low potential for spatial interactions), red hue illustrates places where family connectedness is higher.

The animations of family connectedness including the Chaffee family  and all families can be viewed below.

 

 

Future Work

We demonstrated our approach using a family tree data set from a population that was growing and colonizing the Northern part of the US. This measure has demonstrated how important migration, birth, and death of individuals to family connectedness. In this study, we define relationship as kinship and assume that two individuals have a relationship if they are from the same family. Our methodology can readily be extended to develop a measure of social connectedness using other forms of relationships, such as friendship and coworkers.

Depending upon the context of the social network, one can define the connections in any form of interaction and quantify those interactions in a variety of ways such as using the frequency of the shared content, common friends in an online social network; the number of email exchanges or meetings held together in a business network. In this regard, the voluminous data collected from social networking platforms such as Twitter, Flickr and Foursquare and genealogy applications such as Family Search and Ancestry provide an excellent opportunity to study online social networks and family trees using our approach.