2015 in review

As in previous years, I would like to make a review of the year 2015.
Here is my github contributions for this year, as my current company works on github, this pumps up the numbers :p.

tinygraphs:

The main project that I finished in 2015 might have been tinygraphs. A web app that generates consistent avatars with different themes. The app is available on tinygraphs.com. The code is available at github.com/taironas/tinygraphs. It is under the taironas organization which I created to centralize apps that I code with friends. The app is done in Go and the front end is done in angularjs. While doing this app I learned many things about developing web applications and wrote the following articles:

• Building tinygraphs an avatar web service in Go
• Testing HTTP caching in Go
• Learning HTTP caching in Go
• Building a chessboard image in Go
• Playing with images in HTTP response in golang

route:

I also built a minimal Go package with Remy Jourde called route. This package handles http routing in Go web applications. This package is currently used in gonawin and tinygraphs apps.

idm:

I started coding idm (it doesn't matter) an interpreted programming language based on APL) and ivy.
This idea came after watching this two videos:

• APL demonstration 1975
• Implementing a bignum calculator - Rob Pike

It is not finished and the code needs a lot of clean up but it does work and it was very fun to code :) .

machine learning things:

One of the things I did during the past year was continue learning machine learning, a field that I am very passionate about. I followed the course Information Theory, Inference, and Learning Algorithms, the lectures available online.
I started to compete on simple machine learning problems in kaggle. I decided to have a repository to centralize the different competitions santiaago/kaggle. For this I set some conditions/challenges. First, all code had to be done in Go. Second, I had to code all algorithms and not use any framework. This forced me to start my own machine learning framework github.com/santiaago/ml right now it has linear regression, logistic regressions, support vector machines and linear transformations. As I was coding a different method the results of the competition improved. I would like to code neural networks models, random forest and PCA, that should be fun.

Project wish list for 2016:

As in previous years, this is what I wish to accomplish during this year:

• pomotime: is a web app to record your pomodoro progress This app is on hold as I have too much on my plate right now. This is done in Go, Dart and mongodb.
• xprize: I am participating with a friend on a competition called xprize, we are building some apps for tablets to teach kids to read, write and count. If you want more information go to the web site listed above. This is done in C++ and cocos2d-x. The deadline is around November 2016.
• I am working with a friend on a very interesting machine learning project that I cannot talk a lot about right now, but we should have a prototype ready for late 2016. This is currently done in python as most of the machine learning frameworks are in python.
• finish a deep learning course: I started a deep learning course by Udacity and google, the repository is available here
• ml framework: keep working on the machine learning framework, add neural networks, PCA and random forest.
• modify the ml framework to be in sync with other ml frameworks, mainly use the same terminology as other frameworks.
• keep doing some interesting Kaggle competitions.

Let's see what 2016 brings :-)

Follow me at @santiago_arias to be notified about more posts like this.

Santiaago