widged/miramar_traffic

Code for traffic Analysis. Developed at the @HackMiramar event on Nov 1.

hack_miramar

Code for traffic Analysis. Developed at the @HackMiramar event on Nov 1.

Getting started

Assuming that you already downloaded or cloned the repository from github:

   git clone https://github.com/widged/miramar_traffic

Move into the miramar_traffic directory and install dependencies.

   cd miramar_traffic
   mkdir data
   npm install

Parsing Wellington Sensor Data

Run

   ./sensor_stream.js 
   mkdir data/cypher
   mkdir data/route 

Requires

• data/wellington_sensor_data_October_2014.csv
• cyper and route subdirectories in the data folder.

These sensor data are not bundled within this archive due to the huge size (20GB). They can be downloaded from the Hack Miramar wiki. https://hack-miramar.wikispaces.com/Data+Sources. Select https://isisgroup.app.box.com/hackmiramar, Hackathon > Araflow Sensors > wellington_sensor_data_October_2014.zip.

Importing the sensor data into neo4j graph database

Run

   
   npm run neo_start
   ./neo_import.js  

Requires

• data/cypher/f_*.txt. Run ./sensor_stream.js to generate these data.
• An install of neo4j-community-2.1.5 in the miramar_traffic directory.

Installing Neo4j Community version

Got Java? You’ll need either Oracle JDK 7 or OpenJDK 7

• MacOS: do not be fooled by Apple’s offer to install Java 6. That won’t work.
• Windows Enterprise: install your preferred JDK 7.
• Windows Community: the installer has the Java you need

Download Neo4j

• head to: http://neo4j.com/download/.

Start Neo4j

• Windows: run the installer. Double-click and enjoy.
• MacOS & Linux: open a terminal, cd to the extracted folder, start with bin/neo4j start
• Open http://localhost:7474 in a browser window.

Explore Neo4j without installing anything

Try out the console:

• http://neo4j-console-20.herokuapp.com/

Fork and write your interactive queries published as Gists

• http://gist.neo4j.org/
• community challenges: First Challenge, Winter Challenge
• Video on how to Model Neo4j Graphs Interactively with a GraphGist

A few queries to try

List all intersections

MATCH (m:intersection) RETURN DISTINCT m.name

Traffic information at a given intersection.

   
MATCH (n {name: '504'})-[r]->(m)  RETURN DISTINCT n.name AS start, m.name as end,  SUM(r.count)

start	end	SUM(r.count)
504	013	1518
504	016	34864
504	004	1320
504	003	16019

Journey time between two intersections

   
MATCH (a:intersection { name:"504" }),(b:intersection { name:"002" }),
  p = shortestPath((a)-[*..3]->(b))
RETURN EXTRACT(x IN nodes(p) | x.name), REDUCE(totalTime=0, n IN relationships(p)| totalTime + n.std_journey_time) as totalTime

MATCH p=(a:intersection { name:"504" })-[*..1]-(b:intersection { name:"003" }) RETURN EXTRACT(x IN nodes(p) | x.name) as intersections, SUM(REDUCE(totalTime=0, n IN relationships(p)| totalTime + n.journey_time)) as total_journey_time LIMIT 10