Table 1.
Definitions of social network centralities used in this study.
Fig 1.
Number of pig shipments per month between 2011 and 2016 in Argentina.
Fig 2.
Average shipment size per month between 2011 and 2016.
Fig 3.
Distribution of pig movements in Argentina from 2011 to 2016.
Red nodes represent markets and blues nodes farms. Red lines are movements coming from markets and blue lines coming from farms.
Fig 4.
Kernel density maps: of Argentinian (A), farm distribution in 2016, (B) swine distribution in 2016, (C) outgoing shipments from 2011 to 2016, (D) outgoing traded pigs from 2011 to 2016, (E) incoming shipments from 2011 to 2016 and (F) incoming traded pigs from 2011 to 2016.
Table 2.
Farm and pig distribution by province in Argentina in 2016.
Fig 5.
Allocation of shipments involving the provinces of Buenos Aires, Cordoba and Santa Fe.
Edges thickness is proportional to the number of shipments, with the two most important trade relationships highlighted in red.
Table 3.
Provincial distribution of shipments and number of shipped pigs in Argentina from 2011 to 2016.
Fig 6.
Distribution of inter-provincial traded pigs in Argentina between 2011 and 2016.
Table 4.
Network centrality values and characteristics for the yearly networks from 2011 to 2016 and the complete network (whole time period).
Fig 7.
Network graph for December 2014.
Node size represents the log value for node betweeness.
Fig 8.
Time series of the monthly density values from 2011 to 2016 (A) and boxplot of the density values aggregated by month from 2011 to 2016 (B).
Fig 9.
Time series of the monthly mean betweenness value from 2011 to 2016 (A) and boxplot of the betweenness values aggregated by month from 2011 to 2016 (B).
Fig 10.
Map of shipments going to or coming from markets, from 2011 to 2016.
Only market nodes are shown. Red lines represent shipments coming from markets, and blue lines, coming from farms.
Table 5.
Network centrality values and characteristics for the yearly sub-networks from 2011 to 2016 and the full sub-network (whole study period) after removing markets.
Fig 11.
Time series of the monthly density value from 2011 to 2016 in the network without markets (A) and boxplot of the betweenness values aggregated by month from 2011 to 2016 in the network without markets (B). These graph are on the same scale as Fig 8.
Fig 12.
Time series of the monthly mean betweenness value from 2011 to 2016 in the network without markets (A) and boxplot of the betweenness values aggregated by month from 2011 to 2016 in the network without markets (B).
Fig 13.
Correlation circle for continuous variables included in the final factor analysis for mixed data model (FAMD) along dimension 1 and 2.
Variables in green are active and in purple are supplementary. All active variables have positive coordinate values in both dimension 1 and 2. The variable names relate to: indegree, outdegree, betweenness, Pig.in = total number of incoming pigs in a given unit, Pig.out = total number of outgoing pigs, Pig.pop = pig population, Av.pig.out = average size of outgoing shipment Av.pig.in = average size of incoming shipment, Polutry = poultry population, Livestock = livestock population, Distance.out = mean distance of outgoing shipment, Distance.in = mean distance of incoming shipment, Area = area of unit.
Table 6.
Results of factor analysis for mixed data (FAMD).
Coordinates represent the mean location of a variable along the 2 dimensions under study, and contribution represents the discriminatory power of the given variable in dimension 1 or 2. (Supplementary categorical variable of province not shown for clarity, due to the large number of provinces).
Fig 14.
Location of nodes in dimensions 1 and 2 following coordinates obtained from factor analysis of mixed data (FAMD) with color coding from hierarchical clustering.
Cluster 1 represents small and backyard productive units (low degree, betweenness, pig population and shipment size), cluster 2 represents large and industrial farms (high degree, pig population and shipment size but low betweenness), Cluster 3 represents markets (high betweenness but low degree and shipment size, and no pig population) and cluster 4 is a single outlying market with extremely high values for betweenness, degree but small shipment size and no pig population.
Table 7.
Characteristics of the four clusters defined by hierarchical clustering based on the active variables selected from factor analysis for mixed data (FAMD).