Speaker: Ben Sadeghi, Solutions Architect, Databricks

Pandas is the de facto standard (single-node) DataFrame implementation in Python, while Spark is the de facto standard for big data processing. With the recently open-sourced Koalas package, you can be immediately productive with Spark, with no learning curve, if you are already familiar with pandas, and have a single codebase that works both with pandas (tests, smaller datasets) and with Spark (distributed datasets). In this talk, we'll go through the basics of Koalas, along with demos.

About the speaker:

Ben Sadeghi is a Partner Solutions Architect at Databricks, covering Asia Pacific and Japan, focusing on Microsoft and its partner ecosystem. Having spent several years with Microsoft as a Big Data & Advanced Analytics Technology Specialist, he has helped various companies and partners implement cloud-based, data-driven, machine learning solutions on the Azure platform. Prior to Databricks and Microsoft, Ben was engaged as a data scientist with Hadoop/Spark distributor MapR Technologies (APAC), developed internal and external data products at Wego.com, a travel meta-search site, and worked in the Internet of Things domain at Jawbone, where he implemented analytics and predictive applications for the UP Band physical activity monitor. Before moving to the private sector, Ben contributed to several NASA and JAXA space missions. Ben has been a user of Python for over 10 years, and is an active member of the open-source Julia language community. He holds an M.Sc. in computational physics, with an astrophysics emphasis.

Event Page: https://pycon.sg/

Produced by Engineers.SG

Speaker: Liling Tan, Research Scientist

The 'transfer learning' hype has transcended from the computer vision domain to natural language processing. This is made ever popular by the wave of language models, named after 'sesame street', ELMO, BERT, ERNIE, OSCAR, etc. This talk will cover the (i) basic understanding of language modeling, (ii) glossing over state-of-art neural net architectures for transfer learning and (iii) introducing the a transfer learning library based on PyTorch.

About the speaker:

Liling is a data geek who works mostly with text processing and machine translation. He contributes to NLTK and seeks help from StackOverflow whenever he works with Pandas dataframe.

Event Page: https://pycon.sg/

Produced by Engineers.SG

Speaker: Takanori Aoki, Data Scientist, HOOQ

Objective: Main purpose of this session is to help audience be familiar with how to develop stream data processing application by Apache Kafka and Spark Structured Streaming in order to encourage them to start playing with these technologies. Description: In Big Data era, massive amount of data is generated at high speed by various types of devices. Stream processing technology plays an important role so that such data can be consumed by realtime application. In this talk, Takanori will present how to implement stream data pipeline and its application by using Apache Kafka and Spark Structured Streaming with Python. He will be elaborating on how to develop application rather than explaining system architectural design in order to help audience be familiar with stream processing implementation by Python. Takanori will introduce examples of application using Tweet data and pseudo-data of mobile device. In addition, he will also explain how to integrate streaming data into other data store technologies such as Apache Cassandra and Elasticsearch. Note: - Python codes to build these applications will be uploaded on GitHub.

About the speaker:

Takanori Aoki is working as a Data Scientist developing data-driven solution to provide better customer experience in on-demand video streaming service. He has been using Python for 3 years to conduct exploratory data analysis, develop production ETL pipeline, and build machine learning model. He built recommendation functionality for movies and tv shows by using Python as a production system. He is interested in not only machine learning algorithm but also data engineering and software engineering in order to build robust production system. LinkedIn profile https://sg.linkedin.com/in/takanori-aoki-7900a438

Event Page: https://pycon.sg/

Produced by Engineers.SG

Speaker: Christianto, Engineer

This talk will demonstrate how to marry Python and Pandas skills with domain knowledge on value investing. In this talk we will try to find a good and healthy companies to invest in the long term, instead of doing speculation and making decision based on rumors. I will explain basic financial analysis, and then how to automate it with Python and Pandas.

About the speaker:

Christianto Kurniawan is trying to understand how to enable people to learn technology fast and efficiently. He believes that blended learning is the way to go for learning technology, that's why he founded www.katalis.app as a blended learning tool, and use this tool for teaching technology at www.solacetech.com.sg

Event Page: https://pycon.sg/

Produced by Engineers.SG

Speakers:
- Chinab Chugh, CTO, Jublia
- Fahmi Fauzi, Team Lead, Jublia

Attending an event is fun until you get lost between exhibitor booths and are unable to find your way around. In my work, we have worked with over 1500 B2B Events and we generally see the use of static floorplans (images or PDFs being the most common). The reason behind this is building Interactive Floorplans can be a time-consuming and expensive task. First, you need a high DPI base image. To make it interactive, you then need to use an SVG creator tool to map each booth as a polygon manually. What if you had 1600 booths over a 8-halls venue? And the booths get allocated and change till the very last day. So we decided to SAAS-ify interactive floorplans by automating the creation of the SVG layer and repetitive tasks of booths linkage. We developed a two stages booth detection model with proposal stage using OpenCV and classification stage using a trained TensorFlow network. This covers cases like rotated booths, different-shaped polygons and more. We then used Google’s Tesseract-OCR Engine (through pytesseract) to detect the text inside the polygon, which represents the booth number and/or the company name, and we use that to link the booth to the exhibitor itself. Finally, the processed data is rendered using the Google Map API. This opens up many possibilities as the events we work with, big or small, can effectively use our engine for attendees to find their way in those tight venue spaces. We will also briefly share about routing capabilities across single and multiple halls in a venue.
About the speaker:

Chinab is the co-founder and CTO of Jublia which specialises in Smart Matchmaking at events. Apart from leading the development team, Chinab runs several web projects on the side and is fascinated by new technology research.

Fahmi is a passionate tech junkie. love to research and learn new tech. Currently, the Team Lead at Jublia. PS: He Loves Python.

Event Page: https://pycon.sg/

Produced by Engineers.SG

Speaker: Novia Listiyani, Data Scientist

Difference between selling price and cost price really matters – especially in retail industry. In fact, the ability to exploit that gap has been an integral part of their competitive advantage. Some companies are committed to maintain lower selling price to attract more customers, even if maintaining lower cost price is not easy. We should remember that the challenge is due to the dynamic nature of demand. Having high inventory means higher flexibility to fulfill demand but it comes with higher warehouse cost. The cost will be higher if there is expiry date for the sold goods as its selling price will be zero once such goods pass the expiry date. In the opposite, lower inventory will trigger sunk cost where we lose profit from unfulfilled demand. The simplest approach for this problem is to do a prediction modeling for the demand of the goods. Given that a company needs to decide a final number in daily basis – to trigger its inventory replenishment process accordingly – the next question to answer is whether it is enough to take the prediction number into consideration since every prediction comes with certain level of error. What if taking a slightly lower (or slightly higher) number can result in a lower total cost? Inspired by a real-world problem, we will discuss on the potentials of stochastic programming when implemented in our problem setup. Instead of doing prediction, we will focus on scenario generation: possible demand that may happen in the future and its probabilities. Then, we adopt linear programming formula from the famous newsvendor problem [1]. This formulation allows us to get number of goods to purchase from vendor with minimization of total cost as the main objective. Not only such approach will allow us to contribute directly to the business objective, but it can also provide flexibility to business stakeholders as they can propose different possible scenarios depending on the nature of the business. This approach can be implemented easily in Python with PuLP as the optimization library. [1] Alexander Shapiro, Andy Philpott, 2007, A Tutorial on Stochastic Programming, https://www2.isye.gatech.edu/people/faculty/Alex_Shapiro/TutorialSP.pdf.

About the speaker:

Novia L Wirhaspati met Python when she was working as production planner in a cigarette filter manufacture five years ago. She found it fun to learn Python during weekend. Obviously, back then, she had no idea on how it will affect her life in the future. Moving forward – after completing masters degree in Operations Research, University of Edinburgh – she is working as a data scientist where she uses Python in daily basis. She really enjoys working on her projects, which can be from different domains like product ranking, budget optimization, demand forecasting, topic clustering, and many others. Whether it is related to predictive analytics or prescriptive analytics, she thinks it is always mentally rewarding if she can implement things she has learned in real life. In her spare time, she likes to read and learn new things. She is also interested in mixed integer linear programming, stochastic programming, Bayesian regression, and reinforcement learning.

Event Page: https://pycon.sg/

Produced by Engineers.SG