I recently finished the Data Science Specialization offered by Coursera. It is a nine course sequence that covers all aspects of data science, from finding and cleaning data, to creating reproducible research, applying regression models and machine learning, and producing data products. The final class is called Developing Data Products and we were required to develop a project for it. For my project I created an interactive data visualization of data scientist jobs worldwide that I call “Where The Jobs Are”. The app processes data scientist job listings obtained from the Careerjet.com API, showing numbers of jobs and salary ranges in bar charts and box plots by region. Users can drill down one level by selecting the region and see jobs by city. They can also see job listing details and click through to the actual job advertisement.

The app is built with R and Shiny, while the Careerjet API is accessed with a Python script. Here are a couple of screenshots.











A year ago came the announcement of a huge barter deal wherein Thailand would pay China for equipment for a high-speed rail project. The plan was for two links, one that passed through Chiang Mai in the north of Thailand to Bangkok, and another that passed through Sakon Nakhon in the northeast of Thailand and then on to Bangkok.

According to Forbes, the new military government of Thailand announced that the deal is off. It’s not really clear why. Maybe the details turned out to be unattractive for either side. Maybe all that stockpiled rice that was planned for the barter isn’t really in the silos or is already rotten – there’s been plenty of searching and posturing in previous months about it.

Forbes says this would have the been the biggest barter deal of its type in history. The most interesting aspect of the high speed rail project is that if it ever does get completed it will provide a very valuable route from mainland China all the way to Singapore. Goods, people, and troops (as needed) will be able to be rapidly and relatively inexpensively shipped a long way.

I stumbled upon this video clip of a ferry being loaded in Greece. The poster and commenters got a chuckle out of it. Take a look.

It reminded me of what happened to a friend a few years ago while boarding a ferry in Greece. She lost part of her foot. It got caught under that loading ramp when the boat lurched just as she was trying to jump on. Very painful, lots of medical treatment, and long period of rehabilitation to be able to walk again. Not so funny.

Upon reading the news about a proposed barter deal where Thailand would trade rice and other agricultural products for high speed train components for its high speed rail megaproject, it occurred to me that this is another example of bypassing the petro-dollar. There are already a bunch of China currency swap deals aimed at bypassing the US dollar in international trade. Barter deals are another way to achieve that end.

Instead of buying agricultural products in the traditional way, which involves currency transactions that end up passing through the USD, they would barter agicultural goods directly for train components. The amounts involved are fairly significant. They are talking about more than one million tons of rice and 200,000 tons of rubber per year for the next five years. Price per ton for rice varies quite a bit depending on the type and quality (see the Thai Rice Exporters’ pricing page here), ranging from USD400 to nearly USD1200 per ton currently. Let’s just take the average as USD800 for arguments sake. That would be USD800 million per year for rice. Maybe they will hit a billion USD or more per year with the combined agricultural products. Not bad.

A couple years ago China and Iran were talking about a barter deal so Iran could sell oil without being hamstrung by U.S. sanctions. There are also deals or proposed deals between Iran and India and South Korea. These bartering deals, especially involving big ticket items like oil, could add to the erosion of the petro-dollar’s strength.

I’m into the third week of the UI (user interface) design course I am taking via Coursera called Human Computer Interaction. The lectures for this week include a discussion of some prototyping techniques to use to quickly create a mockup of the interface for your system. The objective is to have a prototype to put in front of users to see how they interact with it and identify issues early in the development cycle. One of the techniques is called the “Wizard of Oz” method.

The name is inspired by the movie of the same name in which the wizard is really just a small man behind a curtain who is pulling the levers to make something much bigger appear operational. This technique is really great for early evaluation of user interfaces because you don’t have to spend a lot of time solving complex issues in the backend. You just need to have an interface that looks somewhat realistic. A “wizard”, which is a person behind the scenes not usually visible to the users, operates the controls to make it appear as if an automated system is actually running everything.

Today I was reading a story at Natural News with the headline “Head fake! Is Healthcare.gov only an empty shell MOCKUP of a working Obamacare exchange?” and it struck me right away that maybe healthcare.gov really is just a mockup in a Wizard of Oz prototype fashion. All the evidence would seem to indicate that. Like the fact that the people in charge can’t even say how many people have signed up through the system. That kind of information just falls out of a real system. Maybe they don’t know because it really is just a mockup with a bunch of people behind the scenes pulling the levers and they have to poll all them and collate the information before they know.

Another reason to think it might just be a mockup is how quickly and thoroughly it has been overwhelmed. Seriously, any system designed for this purpose would have to consider this volume of users at initial roll-out. I don’t think there is really any system on the backend.

And finally, anyone who has done contract work for the government knows how ridiculous it is. It is standard procedure to have specifications thrown at you that are bloated, poorly written, and often times scoped by inexperienced people. Then the specs get changed repeatedly, causing rework and delays. Then there is a funding issue at the end of each fiscal year when funding runs out, the project has to shutdown, and then later gets restarted again when the funding finally comes through, but in the mean time the previous development team went off and found other jobs. I know this one well from my previous work at an aerospace firm where I was a project manager.

My guess is that the system was not completed in time due to all the vagaries of government contracting. The government was firmly committed to a roll out date, but the system wasn’t ready so they threw up a mockup. The wizards, of course, were immediately overwhelmed.

Starbucks sues Thai street vendor

Posted: October 21, 2013 in Thailand

In what might seem a case of a large multinational bullying a small local business, Starbucks is suing a Thai street vendor for infringing on its logo. The street vendor’s business is named Starbung and the owner claims his logo is his own original work that is based on how he makes his coffee and the color green for his religion. Here are the two logos side by side:
Uh, yeah. That right there is what you call infringement. It’s so obviously a direct takeoff. That vendor probably thought Starbucks would never bother a little guy like him. He doesn’t realize that they have to, because if they don’t fight such obvious infringements, even small time ones, they eventually lose their copyright. Starbung should have just dropped it when the cease and desist letter arrived. Trading on the Starbucks logo may have been good while it lasted but he has been found out.

On the other hand, enforcement in Thailand is pretty lax so even if Starbucks pursues and wins the lawsuit it is unlikely that anyone will be interested in enforcing the judgement. Although, if local police were to learn of the judgement it would offer them a fresh extortion opportunity – a small regular payment to keep them from shutting down Starbung. That’s probably how it will go.

List of China currency swap deals

Posted: October 19, 2013 in Economy

I saw the news that China and the EU have now signed a currency swap deal, the latest in the big move by China to cut the USD out of their global commerce affairs. I keep reading one story after another about these deals so started searching for a list of which countries have entered into currency swap arrangements with China. I read that as of October 15 there were 24 currency swap arrangements but I could not find a consolidated list so here is my attempt.

Countries with currency swap deals with China:

  • EU
  • Indonesia
  • South Korea
  • UK
  • Australia
  • Iceland
  • UAE
  • Malaysia
  • Nigeria
  • Chile
  • Singapore
  • Vietnam
  • Brazil(?)
  • Taiwan (still in talking stage)

What are the rest of the 24? Some stories I read sounded like the deals were not fully implemented but still in the trial stages. So I’m not sure of the rest. Most of them are in the range of a few tens of billions of US dollar equivalent. Small change taken individually but a few billion here and there adds up after a while. And of course if things go well the swaps will grow in size as time goes on.