I attended PyCon India 2015 at NIMHANS Convention Centre, Bengaluru, India between October 2-4, 2015.

Day 1

I took an early train from Chennai to reach Bengaluru, and headed straight to the venue. The workshops were scheduled on the first day. After completing the registration formalities, I went to observe the DevSprint. I met a number of dgplug folks with whom I have been interacting online on #dgplug (irc.freenode.net).

After lunch, I attended Prof. Ronojoy Adhikari’s workshop on “Reasoning under uncertainity with Python”. Dorai Thodla began the workshop with an introduction on “Data Science” and its use and applications in the industry. Prof. Ronojoy then spoke on “Data Science: Theory” and “Probability Theory for Data Science”. We already use Boolean logic in our programming. We also have probability concepts implemented in Python. He then proceeded to demonstrate “Probabilistic Programming in Python” using Lea. He had presented a short preview of this talk at the Chennaipy August 2015 meet-up. Prof. Ronojoy showed very interesting Lea one-liners with IPython Notebook. A sample is provided below:

In [1]: from lea import *

In [2]: faircoin = Lea.fromVals('Head', 'Tail')

In [3]: faircoin.random(10)

Out[3]: ('Tail',
         'Tail',
         'Tail',
         'Head',
         'Head',
         'Tail',
         'Tail',
         'Head',
         'Tail',
         'Head')

In [4]: die = Lea.fromVals(1, 2, 3, 4, 5, 6)

In [5]: die1 = die.clone()
        die2 = die.clone()

In [6]: dice = die1 + die2

In [7]: dice

OUt[10]: 2 : 1/36
         3 : 2/36
         4 : 3/36
         5 : 4/36
         6 : 5/36
         7 : 6/36
         8 : 5/36
         9 : 4/36
        10 : 3/36
        11 : 2/36
        12 : 1/36

I had installed the Anaconda distribution and Lea for the workshop, but, couldn’t get Pomegranate to compile. Later, I worked with Jacob Schreiber to get it built on Ubuntu 14.10.

Prof. Ronojoy then explained Bayesian networks, and how the probabilities can be fed to compute the probability for a causal-effect relationship. He demonstrated numerous examples with live coding. His work is part of the Value from Data initiative. Prof. Ronojoy and Dorai’s “Bayesian Inference and Data Analytics” video lectures are available in YouTube.

Day 2

After collecting my speaker kit, I attended the “KG award” presentation, which was given to Vijay Kumar, followed by the key note by Dr. Ajit Kumar. Vijay Kumar is doing an excellent job in organizing the monthly Chennaipy meet-ups. It was also good to meet Dr. Ajit Kumar after many years, having known his work and involvement with the Phoenix project.

During the tea break, I took the time to test my laptop with the projector in auditorium-1, where I was scheduled to speak. I had few demos in my talk, and I needed to set everything up before the talk. After the tea break, I attended the session by Pratap Vardhan on “Consuming Government Data with Python and D3”. The insights from the data and the analysis by Gramener were quite interesting.

I then moved to the “dgplug” 1130 IST staircase meeting, where I met more #dgplug students, and was able to connect IRC nicknames to people’s faces. My suggestion to the students was to start to identify their interests, and to specialize on them. Around noon, I moved back to auditorium-1 to setup everything that I needed for my talk - “Pretty Printing in Python”. There is one slot which I never wanted - the one before lunch because I know that I will be the one standing between the audience and their lunch, and that is exactly what I got. Nevertheless, the slot had to be taken by someone, and I was prepared for it. I had requested the organizers for a whiteboard and markers for my talk, and they had arranged for the same. I had lot of material to cover in 45 minutes. I did rush through, but, paused at times to explain the concepts. The audience were quite interested in 3D printing, and they asked some really interesting questions at the end of the talk. The video is available in YouTube:

After a quick lunch, I headed for the lightning talks to present “Nursery Rhymes”. We are all familiar with nursery rhymes, and the idea is to teach programming languages and software development using them. The audience needed something to wake them after a heavy lunch, and they thoroughly enjoyed the singing! The video link is provided below:

I then attended the “Machine learning techniques for building a large scale production ready prediction system using Python” by Arthi Venkataraman. After a short tea break, I attended the talk by Pooja Salpekar on “Test Driven Development with Ansible”. She gave an overview of using Serverspec. The subject to test is still determined by the developer, and testing with Vagrant is not the same as testing in production deployments. For the last session of the day, I attended the “Python and Riak DB, a perfect couple for huge scale distributed computing” by Naren Arya, but, it was yet another generic talk.

In the evening, I participated in the discussion on building and running communities organized by Satyaakam Goswami on how Python communities are organizing themselves in India, and the challenges faced by them. I shared my experience with the Chennaipy meet-ups, and insights from my “i want 2 do project. tell me wat 2 do” presentation and book. A video recording of the discussion is available:

Day 3

Nicholas H. Tollervey began the day’s proceedings with his keynote. He shared his experience on Python and education. The Micro Python and BBC micro:bit projects are interesting initiatives. The audience did have questions on why they cannot use Android or any other mobile phone for Python development. But, I think they missed the point that Open Hardware initiatives are important and the openness is not the same with present day mobile devices. Nicholas also covered important points on how to organize Python Dojo meet-ups, and as a teacher, the experiences he shared were very useful.

After the tea break, I attended the session by Sumith Kulal on “SymEngine: The future fast core of computer algebra systems”. It was an introductory session on SymEngine, and he shared the roadmap and plan that they have for the software. I was then networking with people before attending “How to build microservices using ZeroMQ and WSGI” by Srinath G. S. It was yet another generic talk, and I headed for early lunch.

The previous night I prepared for another lightning talk titled “Yet Another Lightning Talk”, to once again, wake the audience after their heavy lunch. The talk went as I had planned, and the audience loved it! The video recording is available in YouTube:

A number of people approached me after the lightning talk, appreciating the creativity, and I did make a lot of new friends. The last talk of the day that I attended was on “Python load balancer; 0 to 1 million requests per second”, and yet again it turned out to be a a very generic talk, without going into specifics. I left early to catch a flight.

It will be helpful if organizers can identify talks where the speaker has contributed to the project. When you hire people for free and open source software (F/OSS) work, you would like to see what their contribution to the project is, and how they have added value to the project. I am interested to hear on what the speaker has done. I can always read the project web site, and find out more information about the project. I wish all conferences add this criteria in their talk selection process.

The workshops have been useful. The content of the talks need to be improved though. I was able to meet fellow F/OSS developers during the conference, and also made new friends. I am pleased that I was able to deliver a talk, and two lightning talks.

I would like to thank Manufacturing System Insights for sponsoring my travel, and allowing me to 3D print during office hours!