Week 08: Apache Spark

Page

Content:

This week will be an introduction to Apache Spark and the Python library pyspark.

Watch my lecture here: https://www.youtube.com/watch?v=KCoLQ7BZosA

Learning objectives:

After this week, you are supposed to know:

  • What Apache Spark is
  • How to write a Spark program
  • How to implement and run a pyspark script locally

Resources:

Exercises:

Exercise 0 (don’t include in assignment):

You should install Spark and pyspark. Follow the following steps:

  1. Download and install Vagrant from here: https://www.vagrantup.com/
  2. Download the following Vagrant-file: https://www.dropbox.com/s/q8yj7qd5hifkhfj/Vagrantfile?dl=0
  3. Go to the folder of the Vagrantfile
  4. Run “vagrant up” to boot the virtual machine (this will probably take some time)
  5. Access the Jupyter web UI for running IPython notebooks by navigating your web browser to “http://localhost:8001” (or “http://127.0.0.1:8001/“)
  6. Download the following iPython notebook with some examples: https://www.dropbox.com/s/hbz98qnhbspc1mf/Spark%20test.ipynb?dl=0
  7. Upload the notebook using the Upload-button in the ipython notebook in the browser. Note: The upload button in iPython is also a good way to copy the data files to the Vagrant virtual machine.
  8. Open the notebook in the browser and run the code!

Exercise 1:

Write a Spark job to count the occurrences of each word in a text file. Document that it works with a small example.

Exercise 2:

Write a Spark job that determines if a graph has an Euler tour (all vertices have even degree) where you can assume that the graph you get is connected. This file https://www.dropbox.com/s/usdi0wpsqm3jb7f/eulerGraphs.txt?dl=0 has 5 graphs – for each graph, the first line tells the number of nodes N and the number of edges E. The next E lines tells which two nodes are connected by an edge. Two nodes can be connected by multiple edges.

It is fine if you split the file into 5 different files. You do not need to keep the node and edge counts in the top of the file.

Document that it works using a small example.

Exercise 3:

You are given a couple of hours of raw WiFi data from my phone: https://www.dropbox.com/s/964gq5o5bkzg7q3/wifi.data?dl=0

Compute the following things using Spark:

1. What are the 10 networks I observed the most, and how many times were they observed? Note: the bssid is unique for every network, the name (ssid) of the network is not necessarily unique.
2. What are the 10 most common wifi names? (ssid)
3. What are the 10 longest wifi names? (again, ssid)

Advertisements

One thought on “Week 08: Apache Spark

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s