In-class Exercise 5 Mini Challenge 1

Getting Started

In the code chunk below, p_load() of pacman package is used to load the R packages into R environment.

pacman::p_load(jsonlite, tidyverse, SmartEDA, tidygraph, ggraph)

Importing Knowledge Graph Data

In the code chunk below, fromJSON() of jsonlite package is used to import MC1_graph.json file into R and save the output object.

kg <- fromJSON("data/MC1_graph.json")

Inspect structure

str(kg, max.level = 1)

List of 5
 $ directed  : logi TRUE
 $ multigraph: logi TRUE
 $ graph     :List of 2
 $ nodes     :'data.frame': 17412 obs. of  10 variables:
 $ links     :'data.frame': 37857 obs. of  4 variables:

Extract and inspect

nodes_tbl <- as_tibble(kg$nodes)
edges_tbl <- as_tibble(kg$links)

Initial EDA

ggplot(data = edges_tbl,
       aes(y = `Edge Type`)) +
  geom_bar()

ggplot(data = nodes_tbl,
       aes(y = `Node Type`)) +
  geom_bar()

Creating Knowledge Graph

Step 1: Mapping node id to row index

id_map <- tibble(id = nodes_tbl$id,
                 index = seq_len(
                   nrow(nodes_tbl)))

This ensures each id from node list is mapped to the correct number.

Step 2: Map source and target IDs to row indices

edges_tbl <- edges_tbl %>%
  left_join(id_map, by = c("source" = "id")) %>%
  rename(from = index) %>%
  left_join(id_map, by = c("target" = "id")) %>%
  rename(to = index)

The number of observations in edges_tbl should be the same as before running this code chunk.

Before doing leftjoin, there are only 4 variables. AFter doing the leftjoin, there is two additional variables.

Step 3: Filter out any unmatched

edges_tbl <- edges_tbl %>%
  filter(!is.na(from),!is.na(to))

This will get rid of any missing values.

Step 4: Creating the graph

Lastly, tbl_graph() is used to create tidygraph’s graph object by using the code chunk below.

graph <- tbl_graph(nodes = nodes_tbl,
                   edges = edges_tbl,
                   directed = kg$directed)

Directed will be plugged from kg table’s directed column.

Visualising the knowledge graph

set.seed(1234)

This is to ensure reproducibility. ### Visualising the Whole Graph

ggraph(graph, layout = "fr") + 
  geom_edge_link(alpha = 0.3,               # line, alpha is transparency 
                 colour = "gray") + 
  geom_node_point(aes(color = `Node Type`), # point (plot after line so that it doesn't get covered by line)
                  size = 4) +               # size of point  
  geom_node_text(aes(label = name),         # label using name
                 repel = TRUE,              # prevent overlapping names, force words apart
                 size = 2.5) +
  theme_void()

Visualising the sub-graph

In this section, we are interested to create a sub-graph base on MemberOf vaue in Edge Type column of the edges data frame.

Step 1: Filter edges to only “MemberOf”

graph_memberof <- graph %>%
  activate(edges) %>%                     # Focus on edges table
  filter(`Edge Type` == "MemberOf")       # Filter to Memberof

Step 2: Extract only connected nodes (i.e., used in these edges)

used_nodes_indices <- graph_memberof %>%
  activate(edges) %>%
  as_tibble() %>%
  select(from,to) %>%            # Only selected variables
  unlist() %>%                  # beCause it is a graph model, not a list
  unique()

This is to eliminate orphan nodes.

Step 3: Keep only those nodes

graph_memberof <- graph_memberof %>%
  activate(nodes) %>% 
  mutate(row_id = row_number()) %>%
  filter(row_id %in% used_nodes_indices) %>%
  select(-row_id)  # optional clean up

Plot the sub-graph

ggraph(graph_memberof,
       layout = "fr") + 
  geom_edge_link(alpha = 0.5,
                 colour = "gray") +
  geom_node_point(aes(color = `Node Type`),
                  size = 1) +
  geom_node_text(aes(label = name),
                 repel = TRUE,
                 size = 2.5) +
  theme_void()