The 5th edition of the GOTO Berlin conference took place on the 14th-17th November. We didn't want to miss this opportunity to learn from the some of the best in the software industry so we went along to attend some of the workshops and talks at the event.
The GOTO Conferences are created for developers, by developers with events in Chicago, Amsterdam, Copenhagen and Berlin. Having an office in Berlin we didn’t want to miss the opportunity of attending this event which took place in the Berlin Congress Center.
The Conference took place a block away from Alexanderplatz
On the first day of GOTO Berlin there were only workshops running, there was a wide variety of topics and all of them sounded really interesting. Here’s a summary of the workshops we attended.
Working Effectively with Legacy Code by Michael Feathers
We were excited about the workshop since we had greatly enjoyed the book by Michael Feathers with the same name.
The content was half presentation style, half coding exercises. The most interesting part was hearing about Michael’s experience and learnings from decades of software engineering and consulting. The coding examples had just the right size to focus on a certain aspect of refactoring and testing.
All in all we really enjoyed that day, thanks Michael.
Volker Leck and Tobias Heine
Data Science & Analytics for Developers by Phil Winder
I really enjoyed this workshop. It was a great introduction to practical data science and machine learning problems. The instructor gave a lot of good tips that would help an aspiring data scientist. During the workshop we followed the following modules:
- Introduction. A short introduction to the python programming language.
- Basics. Fundamental terms, such as mean, variance, entropy, etc. and how to calculate them.
- Data. We learned how to use the most common data science python libraries, which included pandas, matplotlib and numpy among others, to normalise and plot data in different ways so that we can get as most insights from it as possible and find out which features are important and which ones aren't.
- Regression. Here's where the Machine Learning part began and it got quite fun as we started using scikit-learn to play with linear classifiers, such as linear regression and support vector machines to solve linear problems.
- Generalisation. Where we learned about learning curves, overfitting, underfitting and how to visualise them to make sure we can understand when we're making mistakes.
- Nearest Neighbours. In this module we got into the unsupervised learning algorithm known as k-nearest neighbours, which can be used to crete simple recommendation algorithms, and more generally, to find 'similar' stuff.
- Clustering. This was the last module and we learned how to use agglomerative clustering and dendograms to figure out the structure of our data and find what features represent a cluster.
Phil using the whiteboard to explain the ins and outs of data science
These modules were distributed as a list of Jupyter notebooks and they were a great way to get started in solving practical problems using data science techniques and some machine learning algorithms.
On the second and third days there were plenty of talks split in 5 tracks each day (10 in total) and each track had a specific topic (i.e. Connected World, Agile, Programming Languages, etc.). Here are a the talks that stood out for us.
Drones on Broadway by Raffaello D'Andrea
Professor D’Andrea gave a talk on how they created a breed of interactive and autonomous flying machines that have successfully been operated within Broadway shows and concerts.
It was interesting to learn about the technical challenges his teams had to solve in order to make drones fault-tolerant and safe to operate above a crowd of people.
An amazing talk about the state of robotics, with lots of compelling little demo videos and first-hand experience and learnings.
Number Crush by Hannah Fry
Hannah Fry presented how we can use mathematics to identify and explain a lot of behaviour and social patterns in our societies.
She explained how we can model real life problems with mathematics, so that we can create formulas that estimate how likely it is for a couple to break up, based on the number of arguments in the relationship.
She also showed really cool projects she worked on, such as one that showed how the public transport in London moved during a 24-hour period.
Hannah presenting the science behind long-lasting relationships
This was a really inspiring talk, would totally recommend watching some of Hannah's videos on YouTube.
To sum up, we definitely enjoyed attending the GOTO Conference, and we’re looking forward to attending the next edition.