RNeo4j

RNeo4j is Neo4j's R driver. It allows you to read and write data from / to Neo4j directly from your R environment.

Contents

• Install
• Connect
• Nodes
• Relationships
• Cypher
• Shortest Paths
• Weighted Shortest Paths
• Graph Algorithms
• Visualizations
• Import
• Connection Issues
• Contribute

Install

Neo4j

First and foremost, download Neo4j!

Windows

If you're on Windows, download the .exe and follow the instructions. You'll get a GUI where you simply press "Start" to start Neo4j.

OS X

If you're on OS X, you can download either the .dmg or the .tar.gz. The .dmg will give you a GUI where you simply press "Start" to start Neo4j. Otherwise, download the .tar.gz, unzip, navigate to the directory and execute ./bin/neo4j start.

Linux

If you're on Linux, you have to use the .tar.gz. Download the .tar.gz, unzip, navigate to the directory and execute ./bin/neo4j start.

You may also find neo4j in your distribution's package manager.

Bolt dependencies

These depencies are are only required if you want to use the Bolt interface. They must be present at build time, and libneo4j-client must also be present at runtime.

Windows

• Rust: https://rustup.rs
• Clang: binary releases are available at http://releases.llvm.org/download.html
• libneo4j-client:
  • Make sure you have RTools installed (necessary for building R packages on Windows)
  • Open your MinGW shell (check in start menu, C:\RTools\MinGW\bin, and C:\MinGW\bin)
  • See "Installing libneo4j-from source" section

OS X (with Homebrew installed)

• Rust: brew install rust (or https://rustup.rs but see "Rust Path" section)
• Clang: brew install llvm
• libneo4j-client: brew install cleishm/neo4j/libneo4j-client

Linux

• Rust:
  • Debian based (e.g. Mint, Ubuntu): sudo apt-get install cargo
  • Arch Linux: sudo pacman -S rust
  • Building from source: https://rustup.rs but see "Rust Path" section
• Clang: get it from your package manager
  • Debian based (e.g. Mint, Ubuntu): sudo apt-get install clang libclang-dev
  • Arch Linux: sudo pacman -S clang
  • Other: your package manager almost certainly has clang. It may be called llvm.
• libneo4j-client:
  • Debian based: sudo apt-get install libneo4j-client-dev
  • Other: See "Installing libneo4j-client from source"

Rust Path

By default, on *nix systems (such as Linux and OS X), rustup only sets the PATH in your shell. That means that if you try to build RNeo4j in a GUI application like RStudio, it may fail. To work around this issue, simply build RNeo4j in a terminal.

Installing libneo4j-client from source

Newer versions of GCC require removing the -Werror from GCC_CFLAGS in configure.ac.

Run these commands in your shell:

git clone https://github.com/cleishm/libneo4j-client
cd libneo4j-client
./autogen.sh
./configure --disable-tools
sudo make install

See https://github.com/cleishm/libneo4j-client for more details

RNeo4j

From CRAN

install.packages("RNeo4j")

From GitHub

devtools::install_github("nicolewhite/RNeo4j")

From Source

Go to the latest release and download the source code. You can then install with install.packages.

install.packages("/path/to/file.tar.gz", repos=NULL, type="source")

Load the Package

library(RNeo4j)

Connect

graph = startGraph("http://localhost:7474/db/data/")

If you have authentication enabled, pass your username and password.

graph = startGraph("http://localhost:7474/db/data/", username="neo4j", password="password")

Nodes

nicole = createNode(graph, "Person", name="Nicole", age=24)
greta = createNode(graph, "Person", name="Greta", age=24)
kenny = createNode(graph, "Person", name="Kenny", age=27)
shannon = createNode(graph, "Person", name="Shannon", age=23)

Relationships

r1 = createRel(greta, "LIKES", nicole, weight=7)
r2 = createRel(nicole, "LIKES", kenny, weight=1)
r3 = createRel(kenny, "LIKES", shannon, weight=3)
r4 = createRel(nicole, "LIKES", shannon, weight=5)

Cypher

If you're returning tabular results, use cypher, which will give you a data.frame.

query = "
MATCH (nicole:Person)-[r:LIKES]->(p:Person)
WHERE nicole.name = 'Nicole'
RETURN nicole.name, r.weight, p.name
"

cypher(graph, query)

##   nicole.name r.weight  p.name
## 1      Nicole        5 Shannon
## 2      Nicole        1   Kenny

For anything more complicated, use cypherToList, which will give you a list.

query = "
MATCH (nicole:Person)-[:LIKES]->(p:Person)
WHERE nicole.name = 'Nicole'
RETURN nicole, COLLECT(p.name) AS friends
"

cypherToList(graph, query)

## [[1]]
## [[1]]$nicole
## < Node > 
## Person
## 
## $name
## [1] "Nicole"
## 
## $age
## [1] 24
## 
## 
## [[1]]$friends
## [[1]]$friends[[1]]
## [1] "Shannon"
## 
## [[1]]$friends[[2]]
## [1] "Kenny"

Both cypher and cypherToList accept parameters. These parameters can be passed individually or as a list.

query = "
MATCH (p1:Person)-[r:LIKES]->(p2:Person)
WHERE p1.name = {name1} AND p2.name = {name2}
RETURN p1.name, r.weight, p2.name
"

cypher(graph, query, name1="Nicole", name2="Shannon")

##   p1.name r.weight p2.name
## 1  Nicole        5 Shannon

cypher(graph, query, list(name1="Nicole", name2="Shannon"))

##   p1.name r.weight p2.name
## 1  Nicole        5 Shannon

Shortest Paths

p = shortestPath(greta, "LIKES", shannon, max_depth=4)
n = nodes(p)
sapply(n, "[[", "name")

## [1] "Greta"   "Nicole"  "Shannon"

Weighted Shortest Paths

p = shortestPath(greta, "LIKES", shannon, max_depth=4, cost_property="weight")
n = nodes(p)
sapply(n, "[[", "name")

## [1] "Greta"   "Nicole"  "Kenny"   "Shannon"

p$weight

## [1] 11

Graph Algorithms

library(igraph)

query = "
MATCH (n)-->(m)
RETURN n.name, m.name
"

edgelist = cypher(graph, query)
ig = graph.data.frame(edgelist, directed=F)

betweenness(ig)

##  Nicole   Greta   Kenny Shannon 
##       2       0       0       0

closeness(ig)

##    Nicole     Greta     Kenny   Shannon 
## 0.3333333 0.2000000 0.2500000 0.2500000

Visualizations

igraph

plot(ig)

ggnet

library(network)
library(GGally)

net = network(edgelist)
ggnet(net, label.nodes=TRUE)

visNetwork

Read this blog post and check out this slide deck.

Import

library(hflights)
hflights = hflights[sample(nrow(hflights), 1000), ]
row.names(hflights) = NULL

head(hflights)

##   Year Month DayofMonth DayOfWeek DepTime ArrTime UniqueCarrier FlightNum
## 1 2011     1         15         6     927    1038            XE      2885
## 2 2011    10         10         1    2001    2322            XE      4243
## 3 2011     6         15         3    1853    2108            CO       670
## 4 2011     4         10         7    2100     102            CO       410
## 5 2011     1         25         2     739    1016            XE      3083
## 6 2011     9         13         2    1745    1841            CO      1204
##   TailNum ActualElapsedTime AirTime ArrDelay DepDelay Origin Dest Distance
## 1  N34110               131     113      -10       -3    IAH  COS      809
## 2  N13970               141     127        2       19    IAH  CMH      986
## 3  N36207               255     231       15       -2    IAH  SFO     1635
## 4  N76517               182     162      -18        5    IAH  EWR     1400
## 5  N12922               157     128        0       -6    IAH  MKE      984
## 6  N35271                56      34       -7       -5    IAH  SAT      191
##   TaxiIn TaxiOut Cancelled CancellationCode Diverted
## 1      6      12         0                         0
## 2      4      10         0                         0
## 3      5      19         0                         0
## 4      7      13         0                         0
## 5      4      25         0                         0
## 6      3      19         0                         0

addConstraint(graph, "Carrier", "name")
addConstraint(graph, "Airport", "name")

query = "
CREATE (flight:Flight {number: {FlightNum} })
SET flight.year = TOINT({Year}),
    flight.month = TOINT({DayofMonth}),
    flight.day = TOINT({DayOfWeek})

MERGE (carrier:Carrier {name: {UniqueCarrier} })
CREATE (flight)-[:OPERATED_BY]->(carrier)

MERGE (origin:Airport {name: {Origin} })
MERGE (dest:Airport {name: {Dest} })

CREATE (flight)-[o:ORIGIN]->(origin)
CREATE (flight)-[d:DESTINATION]->(dest)

SET o.delay = TOINT({DepDelay}),
    o.taxi_time = TOINT({TaxiOut})

SET d.delay = TOINT({ArrDelay}),
    d.taxi_time = TOINT({TaxiIn})
"

tx = newTransaction(graph)

for(i in 1:nrow(hflights)) {
  row = hflights[i, ]
  
  appendCypher(tx, query,
               FlightNum=row$FlightNum,
               Year=row$Year,
               DayofMonth=row$DayofMonth,
               DayOfWeek=row$DayOfWeek,
               UniqueCarrier=row$UniqueCarrier,
               Origin=row$Origin,
               Dest=row$Dest,
               DepDelay=row$DepDelay,
               TaxiOut=row$TaxiOut,
               ArrDelay=row$ArrDelay,
               TaxiIn=row$TaxiIn)
}

commit(tx)

summary(graph)

##     This          To    That
## 1 Flight OPERATED_BY Carrier
## 2 Flight      ORIGIN Airport
## 3 Flight DESTINATION Airport

Connection Issues

Couldn't connect to server

Error in curl::curl_fetch_memory(url, handle = handle) : 
  Couldn't connect to server

Neo4j probably isn't running. Make sure Neo4j is running first. It's also possible you have localhost resolution issues; try connecting to http://127.0.0.1:7474/db/data/ instead.

No authorization header supplied

Error: client error: (401) Unauthorized
Neo.ClientError.Security.AuthorizationFailed
No authorization header supplied.

You have auth enabled on Neo4j and either didn't provide your username and password or they were invalid. You can pass a username and password to startGraph.

graph = startGraph("http://localhost:7474/db/data/", username="neo4j", password="password")

You can also disable auth by editing the following line in conf/neo4j-server.properties.

# Require (or disable the requirement of) auth to access Neo4j
dbms.security.auth_enabled=false

Contribute

Check out the contributing doc if you'd like to contribute!