Andrew B. Collier / @datawookie

A project I’m working on has a bunch of different data sources. Some of them have column names in Camel Case. Others are underscore delimited. My OCD rebels at this disarray and demands either one or the other.

If it were just a few columns and I was only going to have to do this once, then I’d probably just quickly do it by hand. But there are many columns and it’s very likely that there will be more data in the future and the process will need to be repeated.

Seems like something that should be easy to automate.

We recently announced the second satRday (Cape Town) conference scheduled to take place on 17 March 2018. Obviously we want this to be bigger and better than this year’s event, so we are paying careful attention to the feedback that we received from the first event.

This is a quick analysis of the feedback. We sold 192 tickets and gave out 11 complimentary tickets to the event. There were 107 responses to the feedback survey, which means that we heard back from more than half of the people who attended, which is hopefully a representative sample.

I was invited to give a talk at Digifest (Durban University of Technology) on 10 November 2017. Looking at the other speakers and talks on the programme I realised that my normal range of topics would not be suitable.

Recipe for installing the NVIDIA binary drivers on Ubuntu.

I’ve now been through the process of setting up OSRM a few times. While it’s not exactly taxing, it seemed like a prime candidate for automation.

Building a presentation with reveal.js is such a pleasure. And the results looks so good. Seriously doubt that I will ever use anything like PowerPoint again. Although it’s possible to export a presentation directly to PDF using a style sheet, this doesn’t always work perfectly (IMHO).

Fortunately there’s another way: decktape. It works with reveal.js and a bunch of other HTML5 presentation frameworks.

I’ve just put together a WordPress site for my older daughter. It’s hosted on DigitalOcean and all of the infrastructure is handled with Docker. This post describes the steps in the (easy) install process.

I’ve got a long running optimisation problem on a EC2 instance. Yesterday it was mysteriously killed. I shrugged it off as an anomaly and restarted the job. However, this morning it was killed again. Definitely not a coincidence! I investigated. This is what I found and how I am resolving the problem.

There are many good reasons to use S3 (Simple Storage Service) storage. This is a quick overview of how to create a S3 bucket.

This procedure works on both my laptop and a fresh EC2 instance.

I’ve been putting together a small proof-of-concept API using R and {plumber}. It works flawlessly on my local machine and I was planning on deploying it on an EC2 instance to demo it for a client. However, I ran into a snag: despite opening the required port in my Security Group I was not able to access the API. This is what I needed to do to get it working.

EC2 Spot Instances can provide very affordable computing on EC2 by allowing access to unused capacity at significant discounts.

The Open Source Routing Machine (OSRM) is a library for calculating routes, distances and travel times between spatial locations. It can be accessed via either an HTTP or C++ API. Since it’s open source you can also install locally, download appropriate map data and start making efficient travel calculations.

These are the instructions for getting OSRM installed on a Ubuntu machine and hooking up the osrm R package.

I know that global variables are from the Devil, but sometimes you just can’t get around them.

I’m building a small package for a client that relies on a data file. For various reasons that file is not part of the package and can reside in different locations on users’ machines. Furthermore there are users on both Windows and Linux machines.

Although it’s very handy (and easy) to set up some cloud resources using the AWS Management Console, once you know what you need it makes a lot of sense to automate the process. Fortunately there’s a handy little command line tools, aws, which makes this eminently possible. The AWS CLI Command Reference is the definitive resource for this tool. There’s a mind boggling array of possibilities. We’ll take a look at a small selection of them.

I have been retrieving some route information using Rodrigo Azuero’s gmapsdistance package and noted that there was some asymmetry in the results: the time and distance for the trip from A to B was not necessarily always the same as the time and distance for the trip from B to A. Although in retrospect this seems self-evident, it merited further investigation.

We’ve been building some models for Kaggle competitions using an EC2 instance for compute. I initially downloaded the data locally and then pushed it onto EC2 using SCP. But there had to be a more efficient way to do this, especially given the blazing fast bandwidth available on AWS.

Enter kaggle-cli.

Update: Apparently that project has been deprecated in favour of kaggle-api. More information below.

Ensuring that your account is configured to run with appropriate time zone information can make your life a lot easier.

Of course, if you administer your own system then you can simply set your system time to local time. However, it’s generally a better idea to set system time to Universal Time (UTC) and then configure time zone information on a per-user basis.

Why does this make sense? Well, suppose that you have remote users logging onto your system. It’s very likely that a remote user will be operating in a different time zone and it’d be handy for them to have system time converted into their local time.

I’m in the process of setting up a Zinnia blog on one of my Django sites. After putting all of the necessary plumbing in place I got the following message on first visiting the blog URL:

Database returned an invalid value in QuerySet.datetimes(). Are time zone definitions for your database and pytz installed?

The solution to this is to copy your system’s time zone information across to the database.

Some quick notes on adding a storage volume to an EC2 instance.

A couple of options for remote access to desktop applications on a EC2 host.

A record of some more or less important events in the history of R.

This is a work in progress. The information is cobbled together from a range of sources. If you have pertinent items to add, please let me know via the comments.

By default an EC2 instance has only a single user other than root. For example, on a Ubuntu instance, that user is ubuntu. If there will be multiple people accessing the instance then it’s generally necessary for each of them to have their own account. Setting this up is pretty simple, it just requires sorting out some authentication details.

I’m busy putting together a Docker image for a multi-user Jupyter Notebook installation. I am to have an independent login for each of the users and each of them should also have their own storage space. That space should exist elsewhere from on the container though, so that even if the container stops, the data lives on. This should mitigate user rage.

Amazon’s EC2 Container Services (ECS) is an orchestrated system for deploying Docker containers on AWS. This post is about not using ECS.

I’ll be running a training course in a few weeks which will use RStudio as the main computational tool. Since it’s a short course I don’t want to spend a lot of time sorting out technical issues. And with multiple operating systems (and versions) these issues can be numerous and pervasive. Setting up a RStudio server which everyone can access (and that requires no individual configuration!) makes a lot of sense.

This is what I did to set up Hadoop on my Ubuntu machine.

I’m busy experimenting with Spark. This is what I did to set up a local cluster on my Ubuntu machine. Before you embark on this you should first set up Hadoop.

In the process of uploading a massive CSV file to my Django application my session data are getting pretty big. As a the result I’m getting these errors:

• (1153, "Got a packet bigger than 'max_allowed_packet' bytes") and
• (2006, 'MySQL server has gone away').

The second error is potentially unrelated.

After some research it became apparent that the source of the problem is my max_allowed_packet setting.

I’ve been meaning to set up a VPN and this morning seemed like a good time to tick it off the bucket list. This is a quick outline of my experience, which included one minor hiccup.

A short note on how to set up Jupyter Notebooks with Python 3 on Ubuntu. The instructions are specific to Xenial Xerus (16.04) but are likely to be helpful elsewhere too.

I’m in the process of deploying a scraper on a DigitalOcean instance. The scraper uses RSelenium with the PhantomJS browser. I ran into a problem though. Although it worked flawlessly on my local machine, on the remote instance it broke with an following error.

I have been looking at methods for clustering time domain data and recently read TSclust: An R Package for Time Series Clustering by Pablo Montero and José Vilar. Here are the results of my initial experiments with the TSclust package.

The Bulgaria Web Summit happened on 7 and 8 April 2017 at the Inter Expo Center in Sofia, Bulgaria.

There are a variety of ways to predict running times over the standard marathon distance (42.2 km). You could dust off your copy of The Lore of Running (Tim Noakes). My treasured Third Edition discusses predicting likely marathon times on p. 366, referring to tables published by other authors to actually make predictions. There’s also a variety of online services, for example:

• Runners’ World’s Race Time Predictor (based on Riegel’s Formula),
• Running for Fitness’s Race Predictor and
• Race Result Predictor.

Of these I particularly like the offering from Running for Fitness which produces a neatly tabulated set of predicted times over an extensive range of distances using a selection of techniques including Riegel’s Formula and Cameron’s Model.

Amazon seems to really understand me. Or, at least, my reading preferences. Running, garlic, data. Yup, that pretty much sums me up.

Spent a very diverting few minutes playing with Quick, Draw! this morning, which is one of the cool AI Experiments hosted by Google.

A simple problem sent through to me by one of my running friends:

There are 6 red cards and 1 black card in a box. Busi and Khanya take turns to draw a card at random from the box, with Busi being the first one to draw. The first person who draws the black card will win the game (assume that the game can go on indefinitely). If the cards are drawn with replacement, determine the probability that Khanya will win, showing all working.

Read More →

satRday Cape Town will happen on 18 February 2017 at Workshop 17, Victoria & Alfred Waterfront, Cape Town, South Africa.

I followed up a reference to fast-neural-style from Twitter and spent a glorious hour experimenting with this code. Very cool stuff indeed. It’s documented in Perceptual Losses for Real-Time Style Transfer and Super-Resolution by Justin Johnson, Alexandre Alahi and Fei-Fei Li.

The basic idea is to use feed-forward convolutional neural networks to generate image transformations. The networks are trained using perceptual loss functions and effectively apply style transfer.

What is “style transfer”? You’ll see in a moment.

I’m generally not too interested in fitting analytical distributions to my data. With large enough samples (which I am normally fortunate enough to have!) I can safely assume normality for most statistics of interest.

Recently I had a relatively small chunk of data and finding a decent analytical approximation was important. I had a look at the tools available in R for addressing this problem. The {fitdistrplus} package seemed like a good option. Here’s a sample workflow.

I’m busy working my way through Kyle Banker’s MongoDB in Action. Much of the example code in the book is given in Ruby. Despite the fact that I’d love to learn more about Ruby, for the moment it makes more sense for me to follow along with Python.

Sometimes you’ll want to see how a site behaves on a slower connection. This can be easily emulated using Chrome DevTools. Go to the Network tab and press the “No throttling” dropdown, which will give you a selection of presets and the option to configure custom connections.

When figuring out how to formulate the contents of a POST request it’s often useful to see the “typical” fields submitted directly from a web form.

Seems that I am doing this a lot: deleting my entire graph (all nodes and relationships) and rebuilding from scratch. I guess that this is part of the learning process.

Accessing the Neo4j server running on your local machine is simple: just point your browser to http://localhost:7474/. But with the default configuration the server is not accessible from other machines. This means that other folk can share in the wonder of your nodes edges.

Some instructions for installing Neo4j on Ubuntu 16.04. More for my own benefit than anything else.