Data Lab Link Roundup: Worm wars, csvkit, machine learning platforms, Jeep hacking, and a $12 satellite imagery receiver

This post was created automatically via an RSS feed and was originally published at http://blogs.worldbank.org/opendata/data-lab-link-roundup-worm-wars-csvkit-machine-learning-platforms-jeep-hacking-and-12-satellite

Here are some things that caught our attention last week:

  • Matt Gray points out that American NOAA Weather Satellites transmit pictures via FM radio at 137MHz. Which, as Matt demonstrates, means you could use a $12 USB dongle and a laptop to receive and decode images from space.
  • “Worm Wars” may sound like an unlikely Summer blockbuster, but the debate surrounding a recent re-analysis of Edward Miguel and Michael Kremer’s 2004 deworming study has been making headlines in development circles. It provides an excellent argument for publishing data and methods alongside all research (bravo to the authors for doing so) and I’d recommend Ben Goldacre’s excellent overview of the issue, Chris Blattman’s take on it, and World Bank colleague Berk Ozler’s review of the reanalysis as well.
  • Chrysler issued a recall of about 1.4M vehicles after hackers / researchers managed to take control of a Jeep over the internet. It turns out it was possible to remotely hack the “Uconnect” system in these cars just by knowing the vehicle’s IP address. This is *really* alarming and while there are some legislative options on the table, it’s a sobering reminder that security engineering needs to be at the heart of internet connected devices, and that Commander Adama was, as usual, right.
  • Machine learning is a part of modern computer science and statistics and deals with getting systems to make and improve on predictions based on data. Commoditized machine learning platforms are starting to get pretty good, and lett you quickly try out and visualize various models. Check out Microsoft’s Azure Machine Learning, BigML, IBM Watson Analytics and Amazon Machine Learning.
  • If you’re familiar with unix-like command line tools and work with CSV files, I’d highly recommend checking out CSVKit. It’s a set of utilities for displaying, manipulating and analysing CSV files. Three years after discovering it, CSVKit and Open Refine are two of my most used “data-wrangling” tools.. And if you’re not familiar with command-line tools, CSVKit is a great place to start!
  • Finally, if you’ve not already seen it, check out the new worldbank.org homepage – it’s a beautiful example of mobile-friendly responsive design in action. It weighs in at 1.2Mb which these days is not too shabby, but for an overview of how not to design a fast, low-bandwidth responsive site, here’s one blogger’s take on The Verge.

Posted in Benefits of open data, Posts from feeds, Smart communities, Transparency, Visualising data Tagged with: , , , , ,