25 > 25

I just read this here (very popular Austrian news source) and thought it was very funny.

German Version:

Autodiebe besonders jung
Eine Besonderheit zeigt sich bei den Verdächtigen im Bereich Autodiebstahl und -einbruch: Während bei anderen Delikten mehr als die Hälfte der mutmaßlichen Täter über 25 Jahre alt sind, sind die im Kfz-Sektor tätigen Kriminellen überdurchschnittlich jung, sprich bis zu 50 Prozent zwischen 14 und 25 Jahre alt.

English Version (best effort from my side):

Car thieves are particularly young

There is something specific about suspects in the area of car theft and burglary: 50% of other criminals are over 25 years old. Car thieves and burglars are in average younger. Up to 50% are between 14 and 25 years.

I'm assuming that there are no significant numbers of car thieves younger than 14. So the median for criminals (excluding car thieves and burglars) is 25 years (half of the criminals are older than 25) and the median for car thieves and burglars is 25 years (half of the criminals are younger than 25). And obviously 25 is much less than 25.

Mind Maps & Todo Lists

Inspired by Mark Levy I've started experimenting with FreeMind (an open source mind mapping tool). The visuals are nice, simple, and flexible. There are lots of keyboard shortcuts, so I don't need to use the mouse that much. I like being able to easily collapse or expand sub-branches, spatially arrange them, use font sizes to visualize priorities. I still need to figure out what FreeMind's calendar features are capable of doing. I'd love to have an alternative view where tasks are sorted by deadlines.

The image shows my current todo list at work.

Fat Pigeon

A super fat pigeon was sleeping on the tree right in front of our kitchen window today. I wouldn't have noticed it if the other birds (some of them can be seen in the picture) hadn't made such a big fuss of it.

Last.fm on the Dachstein

Last.fm in Vienna

This morning snow fall and cold temperatures encouraged me to relocate. Now I'm sitting in front of the only proper heating system in this house.



The view from my desk:

Last.fm in Vienna

Last.fm HQ in Vienna this morning.

Good Morning Vienna

Summer 2009

A lot has happened since my last entry. The biggest change in my life has been that I've decided to follow Gwen and return to Vienna. We had never planned to stay in London forever. But even sticking to the plan wasn't an easy decision as there is a lot I really like in London.

I've put together a tiny selection of some of the pictures from the last months.

In May Gwen and I spent a weekend with friends in Amsterdam. Here's a picture of Gwen, Reinhard, and me. Amsterdam with its canals is one of the prettiest cities I've visited.



One sunny Sunday in June Gwen and I were randomly walking through London when we bumped into a piano. No one was around so Gwen gave it a try. Street pianos are a great idea.



Another picture taken in June is this one. Gwen tried to teach me how to dance at a friends wedding. Tonight Gwen will be taking me to dance lessons.



Here's a picture taken in July. One of my favorite places near London is Dover with it's white cliffs.



In August my little sister visited me in London and we went to Greenwich. Behind the sky scrapers in the picture is where Gwen and I lived in London.



Also in August Gwen and I made a weekend trip to Paris where we saw this cute dragon. One of the nice things about London is that you can take fast trains to many great places (including Paris and Amsterdam). I never realized that - like London- Paris has many museums that don't charge admissions. This picture was taken in one of these visitor friendly museums.



And here's where Gwen and I have moved to. We're spending almost all of our free time trying to bring the house back into shape. The last 1.5 days I spent cleaning up the garden - now it's possible to use the drive way again!



One of the nicest things about the house is the view from the big roof terrace. Here's the evening view looking away from Vienna. It doesn't feel like living in a city.

Rampert Dance Company

Gwen and I saw the Rampert Dance Company tonight at Sadler's Wells.

We particularly liked the piece A Linha Curva. The stage design with the colorful lights, the energy, the choreography, and the samba beats were awesome. The percussionists even used berimbaus at one point.

Epping Forest

We cycled to Epping forest today. It's a beautiful ride along the canals.


Picknicking in the forest.


Huge soccer field: only about one third of the goals are visible in this picture.

Highly recommended: Justitia

Tonight was one of those evenings where Gwen booked the tickets in advance and all I knew was that we'd be seeing something. We were a bit late, and Gwen was running when we arrived at Sadler's Wells to pick up our pre-booked tickets. Except the performance wasn't at Sadler's it was at Peacock. Luckily the two are only a bit more than 5 minutes apart by bike if you move a bit faster than usual and you're really lucky with traffic lights (or pretend to be). We made it just in time.

We saw the dance theatre Justitia. And it was absolutely awesome! The stage design was beautiful yet simple. The story was interesting and kept me thinking. The timeline was non linear to the extreme: time-freeze, rewinding, parallel threads, and non chronological order. The dancing was great, they even had a part which reminded me a bit of this ping pong video. The music ranged from classical to electronic. Gwen and I even discussed if we'd want to see it a second time.

The only tiny bit I didn't like was that I didn't understand why her lawyer explained that her husband was a tyrant in court. That seemed counter productive to the story she was trying to sell to the judge.

Python rant: vague built-in data structures

It's surprisingly hard to find out details about built in data structures in Python. For example, the official python data structure tutorial talks about using lists as queues but doesn't mention that list.pop(0) is expensive (O(n)), and doesn't mention that the list is implemented as an array. The documentation of list is very short and omits details related to computational complexity. I like how Java in contrast makes things a lot clearer (LinkedList vs ArrayList etc).

I wonder if the decision was based on the assumption that those using Python wouldn't want to bother with computational complexity? (On the other hand they introduced array in Python 2.6.)

To quantify this rant I Google searched for "<language>" "data structure" "computational complexity". Below are the number of estimated search results.

8990 C++
8780 Java
4090 Matlab
3020 Python

Even Matlab scores higher than Python!

Oxfam Festival Steward Training

Last Thursday Gwen and I attended the Oxfam steward training. Gwen will be volunteering at 4 festivals this year, and together we'll be volunteering again at Summer Sundae Weekender.

Gwen and I should have already attended the training last year, but all training sessions were booked out when we tried to sign up. So this year was our first one.

The training was 3 hours long. The material that was covered would have normally put me to sleep within seconds. (Especially with the lack of sleep I had.) However, the trainer turned out to be a really good entertainer.

We learned lots about what to do if someone comes running out of a festival tent and is on fire. (Depending on how bad the burns are and if the person is still able to follow advice, one option is to kick the persons feet off the ground and role them over. Breaking a leg might be the least to worry about in that situation.)

We were also given lots of fun examples of how to communicate with those under the influence of hallucinogenics. We saw videos on how to not help an ambulance drive through a crowd, and generally learned a lot about various potential dangerous related to crowds.

Btw, Oxfam are still looking for volunteers for this year.

Tower Hamlets Cemetery Park

I pass through this cemetery on my way to work.

Open Windows

Screenshot of my windows task bar.

I particularly like the three icons on the top. Each of those represents a terminal that connects me to one of the many GNU screens that help me keep myself organized. Each of these GNU screens has up to 10 tabs. I also have about 50+ tabs open in Firefox right now. A sign that Firefox hasn't crashed in a while ;-)

I'm also particularly excited about having the 2 Ps and the one R window open. It's been a while since the last time.

Oh, and if TweetDeck wouldn't constantly crash on me I'd have that open too.

My top 20 artists in 2009 so far

Size of the artist images corresponds to number of times I listened to each.

Webservice opening hours

British Gas sent me an email and requested me to read my gas-meter and submit the reading via their web page.

I tried and failed. Their website tells me:

"Access to the application is currently unavailable. If the time is between 6am - 11pm Monday to Friday, 6am - 6pm Saturday or 8am - 11pm Sunday please try again in few minutes. Otherwise, please try again between the times outlined above."

Which reminds me of HSBC which only allows Internet banking money transfers within opening hours.

I wonder if what happens is:
(1) web form is submitted by user (=me)
(2) web form is sent to the printer at British Gas or HSBC
(3) Britsh Gas or HSBC employee gets form from printer and types it into computer

Hanami Season

As seen on my way to work today:



Not comparable to other stuff I've seen but nice nevertheless :-)

Twitter

This video and many discussion with my colleagues convinced me. I need to learn what twitter is about.

Abusing Simple Linear Regression Models

I was very excited about the book I recently bought. However, after a brief first read I'm not sure I'll continue reading it.

Here are two examples of "evidence" presented in the book.



There is a lot to be said about this. For example, that the regression line is obviously a bad fit for the data.

A very quick and rough analysis in Matlab yields the following:



The red line is the regression line the authors use. The blue line is the regression line you get if you drop a single data point (in particular the US). The model the authors use suggests that the homicides go from 10 (low inequality) to about 30 (high inequality) per million (a 200% increase). Dropping the US results in a model that predicts 13 to 22 (an increase of less than 100%). The authors don't bother to point out any of this.

Furthermore, the (95% confidence level) prediction boundaries (dotted lines) indicate that the models need to be treated carefully. The confidence levels for predictions range up to 34 for low inequality and down to 6 for high inequality. The overlap suggests that the authors should gather more data if they want to prove that "Homicides are more common in more unequal countries".

The second example:



The authors briefly mention that there is an issue when using only so few data points. On the other hand they pretend that the data is very clear. Removing the UK and US will completely change the results of their regression model. The data gives their conclusion ("Social mobility is lower in more unequal countries") only very weak support. I also don't like how they scaled it. If I assume that the "low-high" ranges are comparable in both cases I would expect this to be reflected in the plot. Instead the mobility dimension seems visually about 30% less important than the inequality dimension. Anyway, I'm not really sure how useful "low" and "high" without exact definitions are in the first place.