Monday, October 26, 2015

Machine learning stack

These are the languages and technologies that will make it easy for you to deploy products based on machine learning to production. Here are the six technologies you will need to build a machine learning product from scratch; AngularJS, HTML5, NodeJS, MySQL, Cassandra and R.

AngularJS

AngularJS is a great tool to have in your data scientist kit. It allows you to extend HTML to work in entirely new ways. It also allows you to tie your rich data structures directly to your HTML templates, using 2-way bindings.

HTML5

Knowing how to work with the HTML5 canvas makes data visualisation on the web easier.

NodeJS

Using javascript on both the front- and backend unifies the stack so that you can work with the same language and data structure, simplifying development. Node makes it possible to create real-time websites with push capability. This is great for serving live data to the web page.

MySQL

While there are many database options available, MySQL is tried and tested and is very efficient for most data sets.

Cassandra

At some point, MySQL stops scaling and you need something else. That is Cassandra.

R

R is probably the best tool for model testing with access to all the latest machine learning techniques. 

Friday, October 16, 2015

Three tips when using a Random Forest in R

1. Make sure to have either factors or numeric variables in the regression. No strings allowed!
2. Make sure that you have a reasonable number of factors. About six should do the trick.
3. Reduce your sample size and the number of trees when testing. You only need a large number of trees to avoid overfitting. If your model is underperforming with a small number of trees, your problem isn't overfitting.

Monday, October 12, 2015

O'Reilly on technology and work

"Free market Economics is like the poetry of Alexander Pope."

"Books were always available only to a small percentage of the population."

"Will books go the way of the vinyl record? - Probably."

Business models today rarely has to face the rigor of asking "will someone pay for this?"

http://files.libertyfund.org/econtalk/y2015/OReillytechnology.mp3

Using TransferWise to save your business money

TransferWise is the cheapest, and probably fastest, way to send money abroad. But did you know you can use it for your business as well? Before you get started, you will need the following documents:
  • A clear scan or photograph of your passport photo page, driver's license or a national identity card.
  • A copy of a utility bill or a bank statement issued within the last 3 months. It should match the personal address entered on TransferWise
  • A copy of the business’s certificate of incorporation (or any other valid business registration document).
Once you have those ready, just follow these steps to get started:

1. Sign up at TransferWise.com


2. Click the close button in the top right corner to exit the transfer screen

3. Click the figure left of the "send money" button and hit "settings".


4. Fill in your personal details and click save

5. Go to the settings page again and click "business" to fill in your business details

6. Follow the link in the orange box to verify yourself and your business





Faster BTYD

The R-library "Buy 'Til You Die" for customer equity forecasting looked promising, but I quickly realised that the implementation relies heavily on loops and is not suited for any data set bigger than a couple of thousand rows.

To speed things up, I switched out a couple of the loops for some data.table magic. Here it is, about 10000X faster than the original, but there is still some room for improvement.

If you're having trouble getting the BTYD package to run, take a look at this post for fixes.

How expensive is it to buy currency at Tallinn's airport?


I was travelling back to London after a visit to our Tallinn office last Friday, and decided to take a snapshot of the currency exchange board at the airport. I always knew I was getting a bad deal at the airport currency exchange counter, but how bad is the exchange rate exactly? 



Buying currency at an airport is probably the most expensive way to get your travel money. I compared the rate at the airport to what I would get from TransferWise. The difference is quite shocking.

Airport rate: 0.55 EUR/GBP
Airport fee: 0%
500 EUR * 0.55 EUR/GBP = 275 GBP

TransferWise rate: 0.74 EUR/GBP
TransferWise fee: 0.5%
500 EUR (100% - 0.5%) * 0.74 EUR/GBP= 368.15 GBP

Difference: 93 GBP.

The cost of changing money at the airport instead of using TransferWise is 93 GBP. That means you get 25% more if you're using TransferWise.

Been at an airport recently? Send me your snapshot of the exchange rate screen to erik.johansson [a] shs.fi.

Thursday, October 08, 2015

“To be a thought leader, you have to have a thought.” (Eric Schmidt, How Google Works)

“If you focus on your competition, you will never deliver anything truly innovative.” 

“Voltaire wrote, “The perfect is the enemy of the good.”188 Steve Jobs told the Macintosh team that “real artists ship.”189 New ideas are never perfect right out of the chute, and you don’t have time to wait until they get there. Create a product, ship it, see how it does, design and implement improvements, and push it back out. Ship and iterate. The companies that are the fastest at this process will win.” 

“A technical insight is a new way of applying technology or design that either drives down the cost or increases the functions and usability of the product by a significant factor.” 

“If we have data, let’s look at data. If all we have are opinions, let’s go with mine.”

Tuesday, October 06, 2015

King Q2 earnings compared to forecast

The consensus forecast for King's Q2 2015 EPS was 0.36. Based on the amount of Google searches for King's main titles by the end of Q2, I predicted an EPS of 0.45. Now, the actuals are out, and the real numbers are somewhere in the middle.

At a realised EPS for Q2 at 0.41, the surprise against the consensus forecast was 13.89%. My estimate had a negative surprise of -9%.

 King Gross bookings Q2 2015


As for revenue, I forecasted Q2 revenue to be $517M. Actuals turned out to be lower by 5%, at $490M.
Entertaining Blogs - BlogCatalog Blog Directory
Bloggtoppen.se