PyConPk 2019 at Habib University Karachi: Huge Event in a Maze, Diverse Participants, & Observations
I have recently started learning Python from Udemy to get familiar with it and then switch my career to Data Science. One of my Facebook friends invited me to this event and I get myself registered instantly on sastaticket.pk. After a day or two, I got my ticket at my doorstep without any hassle. Being an online learner, a beginner in academic research and a computer programmer I had a lot of thoughts and queries in my mind for the event as it was going to be my very first PyCon experience. I am thankful to all the organizers, sponsors, and everyone who worked behind the scenes and made this gigantic event possible for us and provide the platform to learn and interact with industry professionals working on real-world problems. One of the key successes of the event from my point of view is that it attracted participation from all walks of like: Students, Professionals, Entrepreneurs, and academicians. As I had a vital IEEE Karachi Section meeting to attend, on the same day, I had to rush there so I missed many important talks. Without wasting any of the readers’ precious time lets head to my humble thoughts on various sessions that I attended myself.
1 — Formal Opening of The Python Track by Mr. Masood Rastgar
Mr. Rastgar is Chief Technology Officer at Sastaticket.pk, Google Developer Expert, and an active partaker in tech events. Thanks to his organization I got my ticket for the PyConPk at my doorstep without going out in Karachi’s traffic mess. The best session among all was the formal opening of Python track by Mr. Rastgar as he reviled the long-awaited WIFI SSID and its password which saved a lot of mobile data of all the participants. He briefly talked about the partners, sponsors, and the whole team of volunteers. He also encourages all of us to join the Python Karachi group and take part in the quarterly meetups.
2 — Open Source and Free Software by Mr. Van Lindberg (Key Note Speech)
Mr. Lindenberg is the member of the Board of Director of the Python Software Foundation, author of Intellectual Property and Open Source, and he was one among the “America’s Top 12 Techiest Attorneys” by the American Bar Association Journal in 2012, he is a software developer and regular speaker at various tech talks.
He gave a very informative presentation on the topic of Open Source and Free Software. Before attending his session, I myself use both of these terms interchangeably but there is a subtle difference between them. In simple words, free software is one that allows users to freely share and modify it and open source is the same thing with an objective and philosophy to build a development community for the common good of society. You can read more about the Open Source Definition here. Now producing software has a huge cost associated with it and providing it for free to the whole society is surely helpful for the betterment of the community but the producer gets nothing out of it. The idea of the patent comes to bridge the gap between personal and community gain from the information produced by an individual or an organization.
Information is a form of intangible goods that we can exchange with other goods, for instance, we purchase different software (basically a product that is the result of a lot of information) in exchange for money. If an individual is selling is information then he is inclined towards his personal gain more then his community betterment. This does not mean that the idea of Open Source is problem-free. There is an economic problem called Free Rider Problem which is simply the same as what a leecher does to the torrent if the ratio of leechers is higher than seeders. In other words, it is a problem of community members not contributing their fair share to the cost of a shared resource.
Almost all of the things on which Mr. Lindenberg talked about treated software as a piece of information and considered information as a good as economics treats apples and cars as goods. If I have to summarize his session in two words then I will opt “Information Economics”. He talked about all the things from an economics point of view like excludability of information is the possibility to prevent consumers who have not paid for it from having access to it. Now if we do not pay for the subscription of any expensive software then we can’t use it. Similarly, there is a term called rival goods which reflects the economic idea of a good that can only be possessed or consumed by a single user. A common example of rival goods includes clothes, foods, smartphones, and other electronics.
3 — Kaggle Workshop by Mr. Adnan Zaidi
Mr. Adnan Zaidi is a Ph.D. scholar and Kaggle Ambassador in Pakistan, according to him he is the only Kaggle trainer in Pakistan authorized by Google. Kaggle is basically a Social Media platform for data scientists (DS) and machine learners (ML), owned by Google. It offers short courses to build up skills for beginners, competitions which offer cash rewards, computational and graphical processing resource to fulfill the needs of DS and ML, and last but not the least a ton of data sets available publicly for the whole community to use it for their own development or research problem. The best part of this online community is that it provides all above mention things free of cost in order to promote DS and ML. Just like other online communities such as Stack Overflow, your profile reputation increases with the passage of time through your constant community services by providing solutions to problems and uploading different data sets. The discussion section of this platform is quite obvious with its name and the notebooks sections offer all the ready-made solutions with varying levels of efficiency. One can choose any available notebook solution to tackle the problem in his own research or project he is working on.
I did not know anything about Kaggle before attending the workshop and our instructor encourage us to learn and grow our knowledge by continuously give time to this community. He encourages us to join monthly meetups he uses to organized regarding Kaggle. He also shares his concern that he never saw a Pakistani in different international competitions under the Kaggle. I would like to close with many thanks to and my gratitude for Sir Adnan who not only listen to my basic questions but also patiently answered all of them in a detailed manner.
4 — Speech Recognition in Python by Muhammad Naufil
Mr. Naufil was one of the most engaging speakers thought out all the sessions I attended. He is currently a final year student of BE Electronics at NEDUET, IoT instructor at PIAIC, Freelancer and a self-learner. In his own words, he enjoys the perks of a professional and the fun of being a student at the same time. The curiosity drives him to learn more and more about python and work on speech recognition chatbot on his own. He not only presented how a couple of lines of code in python can do great things in the context of Speech Recognition but also tried to show a demonstration of his own personalized assistant which did not work most probably because of environmental noise. What I liked most about him is that he kept the attention of his audience to him throughout his session with the same joke and you can always watch his whole session here. One of his many recommendations is the following TED Talk on “Why you should make useless things” by Simone Giertz.
5 — Computer Vision and AI with OpenCV/TensorFlow by Taha Anwar
Mr. Anwar is the Founder and CEO of Bleed-AI, AI and computer vision practitioner, and enthusiastic instructor. I strongly suggest you should subscribe to his YouTube channel. Like many others, he covered a lot of topics in a short time. Being a beginner in this domain I can’t comment much about these tools but what I learned in this session is that TensorFlow is a framework for machine learning, OpenCV is a library for computer vision, and last but not the least you don’t need a sledgehammer to kill a fly! Actually, Mr. Anwar emphasizes in many instances during his speech that a lot of newbies use to solve any given problem with the hammer of deep learning. He suggested solving the problem with Machine Learning first.
The second most interesting information in his speech was the key difference between image processing and computer vision (CV). What I understood is that when the transformation is applied that it is image processing and when Computer Vision deals with (i)Classification, (ii)Object Detection, and (iii) Segmentation of an image. Figure 5 shows the application of CV in which classification and object detection of the various objects is shown whereas figure 6 depicts all operation including the segmentation. However, there are some gray areas in which IP and CV overlap each other.
6 — Natural Language Processing with Genism/NLTK by Dr. Tafseer Ahmed
Dr. Ahmed is currently working as an Associate professor at Mahammad Ali Jinnah University (MAJU) Karachi, he earned his Ph.D. from Universitaet Konstanz, Germany. According to his google scholars profile, there are 40 different research publications to his credit with a total of 256 citations and 9 H-Index. The no of citations and H-Index are considered as key indicators of quality research work all around the globe.
Dr. Ahmed shared his work on Urdu Natural Language Processing and briefed us about the problems and challenges he faced during his research work. On my question of replicating his work for other local languages like Pashto and Punjabi, he suggested that a strong corpus containing millions of words is required to develop an efficient solution and it is not an easy job not impossible at all.
7 — My observation and Humble Suggestion to Organizers
Before I start writing about the room for improvement, allow me to reemphasize the importance of such an event and efforts by all the organizers. Being a volunteer at IEEE Karachi Section I myself know that organizing any sort of national or international event is not a walk in the park. Surely PyConPk 2019 was a huge event and a great success if you consider the fact that PyConPk was being held in Karachi for the very first time. I am cordially thankful to the organizers, volunteers, and all the stockholders who worked day and night to make this event a reality. Mis-communication and lack of coordination happen all the time all around the globe. I believe mistakes are the steps towards success and hope that organizers will consider my observation (O), suggestions (S) and take it as constructive criticism for the next PyConPk in Karachi.
(O) — A lot of participants, including me, did not know that there are two parallel tracks (AI & Python) for sessions and workshops. I received a scheduled email that only shows the AI track’s sessions and workshops. Some even did not receive any schedule at all due to late registration I guess.
(S) — In my humble opinion, a consolidate schedule should be emailed to all the participants multiple times and most importantly after the closing of registrations. Schedules could be placed in the form of standees on the venue to let the participants know where is what.
(O) — Python track was also supposed to start at 10 AM (as per as consolidate schedule on PyCon.pk which I just come across) but it started about 18–20 minutes late, which is common in Pakistan but this reflects a lack of professionalism.
(O) — Before formally starting the session Mr. Masood Rastgar (Sastaticket.pk) did tell the audience that there are two parallel tracks and that was the very first instance on which many participants, me as well, get to know that there is something else going on as well. One can argue that it was my lack of awareness but I have seen a no of participants in the same confusion.
(O) — I believe Organizers were more focus on topics, speakers, and sessions then engaging potential participants which are depicted from fb/@PyConPakistan page (Apparently looks official). The post (May/21/2019) on the page that announces PyConPk 2019 date and venue only received 11 comments and 31 shares (As of 14/Oct/2019 7:34 Pm). Some of the page’s posts regarding the event only got a few react and not a single comment or share even the page has 4,880 likes and 4,976 followers.
(S) — Social Media platforms like Facebook could be used more effective by involving and asking all the stack holders to consistently share the page’s posts to create curiosity, sensation, and hype among the audience. Social Media has the potential to bring twice the audience then what actually attended the event.
(S) — The event was held in a University and that is why I felt a lack of participates from the academic community. Thanks to Dr. Tafseer Ahmed from MUJU, Mr. Adnan Zaidi, a Ph.D. scholar, I can’t say that the event was totally isolated from academia. More academicians and researchers should be there in the event to let the partakers see both pictures of academia and industry in the future.
(S) — As the registration for workshops was a separate process than the event itself, separate attendance and participation certificates in the particular workshop is a better way to attract participants in my humble opinion. Thanks to the volunteers who provided seats to the couple of participants who asked to take part in Kaggle workshop while standing. If a couple of extra seats have already arranged proactively then it would be great.
(S) — I do witness the active and organized volunteer pleasantly guiding each and everyone to make the event successful which it is for sure. Like many others, visiting HU for the first time, I felt that I am in a maze for a little while and that is the only reason I felt a lack of standees for directions and pointer to where is what.
8 — Acknowledgement
1 — Shireen Fatima did not only inform me about this wonderful event but also shared her experience of sessions she attended. Thank you so much for your help.
2 — I am thankful to Muhammad Uzair for allowing me to use his taken photo which is very first in this article.