Skip to main content

2009 (old posts, page 7)

Premium data subscriptions in Windows Azure

Microsoft Codename "Dallas" looks very interesting:

Microsoft® Codename "Dallas" is a new service allowing developers and information workers to easily discover, purchase and manage premium data subscriptions in the Windows Azure platform. Dallas is an information marketplace that brings data, imagery, and real-time web services from leading commercial data providers and authoritative public data sources together into a single location, under a unified provisioning and billing framework. Additionally, Dallas APIs allow developers and information workers to consume this premium content with virtually any platform, application or business workflow.
One of the first challenges in undertaking any data mining task is getting all the (massive amounts of data) into a single place in a usable form.

Amazon already has a similar offering with Public Data Sets on AWS but this is the first time I've seen 'real-time' mentioned.

Disclosure: I'm not on the Windows Azure team and know nothing more about this than the public site mentions. But this does sound exciting.

• −    − •    − • •    − • − −     − − −    • −    • − •    • − • •    •    − • − −

Burslem Normals

I enjoyed this for some reason, mostly for the entirely deadpan acting and the eerie background music.

[youtube=http://www.youtube.com/watch?v=NqD171d0VBc]

Where are you going?

America.

Whereabouts?

All over.

Why do you want to go there?

I know it's had some bad press recently but it's the place that's taken my fancy.

• −    − •    − • •    − • − −     − − −    • −    • − •    • − • •    •    − • − −

Visualizing the interstate system

The best visualizations are information dense while being deceptively simple. This example takes the US interstate system and distills it into its key components.

From the post:

Drawing my cues from the original and best metro diagram, H.C. Beck's wonderful London Underground diagram, I have rendered the Interstate system in a much simpler form. I have made the "major" highways (those divisible by 5) the framework of the map, with the "minor" highways reduced in importance and rendered as thinner grey lines. Even with these highways, a difference in the greys indicates whether they are even-numbered (west-east) or odd-numbered (north-south). Dots on the highways indicate interchanges: large dots where major highways meet other major highways, smaller dots where major meets minor and tiny dots where minor highways begin or end. A full key at the bottom indicates clearly where each highway begins and ends.

Eisenhower Interstate System

• −    − •    − • •    − • − −     − − −    • −    • − •    • − • •    •    − • − −

Academia vs business

Working in a heavily research backed environment is one of the things I really enjoy about my job. Experimentation and measurement are central components to making progress in any 'hard problem' type of area and with that comes both delight and disappointment. It demands a certain persistence to continue to try new ideas when seemingly promising ones fall through but the pay-off follows eventually. For me, that reward is actually making something better, getting it production ready, and see its impact take hold in the wild.

I love shipping things.

(this post was inspired by the the most delightful xkcd)

• −    − •    − • •    − • − −     − − −    • −    • − •    • − • •    •    − • − −

Tipping point of cost

Georgia Pacific released the enMotion touch-free paper towel dispenser in 2003 and I remember first seeing one in the wild in about mid 2004. In the last couple of years, based on my anecdotal observation, that type of device is becoming a standard fixture in a majoenMotion[1]rity of public restrooms and by April 2009, they'd sold 1,000,000 boxes of towels. (Standard caveat - I am not everyone and don't visit everywhere.)

Lacking good data on a) penetration rate and b) price trends in that industry, the following statement is entirely conjecture but should be testable.

The list price for a standard dispenser is now about $70 which is actually much lower than I would have guessed. I suspect that one or two things has likely happened: the cost has become neutral or value positive, and/or GP has struck distribution deals with servicing companies like ABM to push the new technology into the mainstream.

I haven't dug into the company reports to see whether the break out of information is in there although some of it may be. Nor have I yet found a good source of historic market price trends for specific items (e.g. 'what was the typical market cost of SKU X on 4/14/2004?') although that would be an awesome resource to have in exploring patterns like this.

• −    − •    − • •    − • − −     − − −    • −    • − •    • − • •    •    − • − −

Pizza Hut and the iPhone

There were reports last week about the impact that the Pizza Hut iPhone application has had on sales with claims of sales of $1M dollars in the first three months.

It's not obvious whether these are net new customers or sales are simply being transferred from other channels like the regular telephone. I would be curious to see if there's any bias in the time distribution of orders (it's easier to order a pizza via app rather than by phone from the back of a taxi), or order composition (the after-bar feast, for example). I suspect iPhone users themselves have subtly different tastes vs the population so that introduces bias there also.

Regardless, providing easier ways to order your products and services is rarely a bad thing.

• −    − •    − • •    − • − −     − − −    • −    • − •    • − • •    •    − • − −

My daily reading

A couple of my favorite reads right now are the Harvard Business Review Voices and Felix Salmon. I thoroughly enjoyed listening to Stephen Dubner and Steven Levitt in Seattle earlier this week, and their NY Times blog Freakonomics is always entertaining.

On the tech front, TechCrunch, Silicon Alley Insider, Mashable and ReadWriteWeb are good for general news. SearchCap from Search Engine Land has a good daily summary of the goings-on in the search world.

I wish I had more time to spend on fun things like Photoshop Tutorials at psdtuts, The Big Picture and a bunch of other photography and PS blogs, but sadly they barely get read.

News mostly comes from Twitter these days which is both focused (good) and myopic (bad) at the same time. It helps having a job that makes me sensitive to large scale trends around breaking events.

A little fiction would be nice soperhaps that's something for the holidays.

• −    − •    − • •    − • − −     − − −    • −    • − •    • − • •    •    − • − −
• −    − •    − • •    − • − −     − − −    • −    • − •    • − • •    •    − • − −

Post offices and the Bing Maps API

I spent an hour this morning hacking together a quick prototype that brought together Windows Azure hosting and storage, the Bing Maps SDK, and a set of photos that Amy and I have been collecting of US Post Offices. The result is a United States Post Offices mashup which shows a map with pushpins for each post office for which which we have a photo.

postoffices

The practical part of this was really the experience rather than the output. In short, it was really surprisingly easy to pull together the pieces to make this work. It takes a little while to get Visual Studio set up as an Azure development platform but once that's done (once) it's trivial taking existing techniques and having them run cloud-enabled with barely any changes.

A few resources that came in handy:

After all that, two things are clear: it should be a good PDC this year and there are many more post offices yet to be visited.
• −    − •    − • •    − • − −     − − −    • −    • − •    • − • •    •    − • − −

Building an optimal phone tree

I am yet to meet someone that enjoys the prospect of calling an automated customer support number to complete a transaction. Fortunately many such tasks can now be completed on the web and that mode of interaction, while not always perfect, is generally a lot more efficient. There's much more bandwidth in the form of a screen to share information and the interaction model is much richer with keyboard/mouse over touchtone keys.

However, every once in a while a support website dead-ends with 'for more information call 1-866-555-1322' and there's nowhere else to go. What follows puzzles me without fail. Never have I wanted to call to get a reprint of my last statement or change my mailing address. My tasks  seem to end up under the menu 8, submenu 6, choice 9 section, deep down in the tree.

Who is the tree structure being optimized for? I'd like to believe the answer is 'the majority of callers but not me' such that the average caller is faced with the least hassle in navigation. That would be aiming towards optimal which is good. But I wonder if they're really measured and arranged in that way at all. I suspect most trees are built via construction - either by primary tasks, or by departments (these boundaries shouldn't show to customers) or by the time the capability was added to the phone tree (yikes!). Perhaps an ideal phone tree would actually constantly adjust based on tasks and demand so that the 'please listen carefully because our options have recently changed' message actually meant something.

Alas, the best predictor of a good customer support phone call still seems to be an immediate answer by a real person who can relate to service they're representing.