PyCon Sweden 2021

Live Stream: https://youtu.be/fMRBOdHZNds Words from the organizers of PyCon Sweden 2021.

Words from the organizers of PyCon Sweden 2021.

Live Stream: https://youtu.be/clRHB3Jq2Ag "Bridging Productivity, Portability, and Performance with Data-Centric Python" By Tal Ben-Nun, senior researcher with the Scalable Parallel Computing Laboratory, ETH Zurich"

Python is rapidly becoming the language of choice for scientific computing, due to its high productivity and vast software ecosystem. As an interpreted language, however, it is challenging to produce high-performance code from arbitrary programs. In this talk we will present the Data-Centric (DaCe) parallel programming framework (https://www.github.com/spcl/dace), a representation and workflow that enables taking (restricted) Python code and generating high-performance programs that run on multi-core CPUs, accelerators such as GPUs and FPGAs, and clusters thereof. The core concept in DaCe is an intermediate representation and interface that separates the definition of "what" to compute from "how" to compute it efficiently and map it onto hardware. The representation can then be transformed automatically, programmatically, or interactively, without modifying the original Python code. To aid code optimization, DaCe provides programmers with visual transformation and editing tools, integrated in the Visual Studio Code IDE, which allow for manual fine-tuning optimizations. DaCe successfully accelerates several frameworks and applications, including NumPy, deep learning with PyTorch/ONNX, numerical weather prediction systems, and supercomputer-scale quantum transport simulations. The generated data-centric programs are on par and outperform the existing state-of-the-art. The talk will highlight the restrictions and challenges when converting scientific code into a data-centric representation, how performance engineering is facilitated with DaCe, and showcase some of the applications that use the framework.

Live broadcast: https://www.youtube.com/watch?v=OcgLuOs1Hrc Present a solution that integrates various components in its architecture, both computational resources, databases and its own python applications and other open source ones. The idea is to show the problems and challenges posed by traditional scraping and how we have been able to build solutions that reduce them, even more so if what is sought is to do it en masse and in parallel. This also means building an automated flow for the post-processing and transformation of the data using machine learning services such as NLP and classification.

Due to the diversity of content on the web, its formats and technologies, the talk proposes a micro-service architecture solution built in Python, but that integrates a workflow with advanced scraping techniques and that allows the transformation of the data obtained. up to service application for NLP and ML classification. The proposal implies the use of Linux, postgresql, redis, mongodb, clickhouse, airflow, among others, but above all, their own developments and frameworks that consider not only the extraction process but also the consumption of RAM, parallel processing and even the website blocking, as well as the analysis and transformation processes of the data obtained.

Live Stream: https://youtu.be/9ZQxvhdOTlA PySpark is a distributed data processing engine widely used in Data Engineering and Data Science. Another way to think of PySpark is a library that allows processing large amounts of data on a single machine or a cluster of machines. We will go through the basic concepts and operations so you will leave the workshop ready to continue learning on your own.

Workshop steps: - Introduction: Motivation, intro to parallel data processing, Spark's main concepts (transformations versus actions, dataframes versus RDDs), and overall architecture, focusing on Spark SQL - Setup environment: There are two ways of executing the notebook with the exercises. The first one is creating an account on Databricks community and cloning the notebook. The alternative is running the notebook locally as described in the instructions. - Exercises: Going through a series of exercises covering Spark's main transformations (filter, select, groupBy) and ways to visualize them. The idea is to give people some time to complete each exercise and then solve it in an interactive way

Live stream: https://youtu.be/veMSbl2fbXE Flask is favoured for prototyping. It is easy to set up and run. However, choosing Flask as your main 'cheval de bataille', be it a company or individual, requires solid grounding. Flask lets you choose your own ingredients, which lights up the joy of coding but bites if not being careful. This talk covers the standard techniques not to be missed as well as new audacious ones used to help manage BIG codebases.

Flask is great, lean and up to the point. A full blown Flask app presents challenges which time and again causes you to change your codebase. When embarking on Flask projects, people rarely think about having an advanded, simple setup to begin with which fines them along the road. That's why it is imperative to have a great base to start with. For companies choosing Flask as their flagship would want to lay out the groundwork with as much foresight as one can. This talk focuses on twerking the factory pattern to perfection, building painless modules for very organised codebases; going far beyond blueprints, integrating flask-restx with the overall structure, using celery as an integration testcase. It also covers nice-to-haves like the best way to handle migrations, implementing a theming system as well as JWT mechanics. A short 25 mins, feature-packed talk featuring years of usage experience.

Live stream: https://youtu.be/DK9teAs72Do Python web frameworks, like Flask, Quartz, Tornado, and Twisted, are increasingly important for writing high-performance web applications. However, even they posit some bottlenecks either due to their synchronous nature or due to the usage of python runtime. Most of them don’t have the ability to speed themselves due to their dependence on *SGIs. This is where Robyn comes in. Robyn tries to achieve near-native Rust throughput along with the benefit of writing code in Python. In this talk, we will learn more about Robyn. From what is Robyn to the development in Robyn.

Robyn is a web framework written with a runtime written in Rust. Robyn tries to achieve near-native rust performance. This talk will demonstrate the reason why Robyn was created, some of the technical decisions taken behind Robyn, the increased performance by using the rust runtime, and most importantly, how to use Robyn to develop web apps. Finally, I will be sharing the future plans of Robyn and would love to get feedback from the developers to see what they would like to see in it.

Live broadcast: https://www.youtube.com/watch?v=oBPNk5qN0L4 At H&M Group, we are increasingly adopting machine learning algorithms and rapidly developing successful use cases, one of the applications is a dynamic resources allocation (memory and cpu) using data driven analysis and ML to decrease the cost of infrastructure. The objective of this talk is to show how one of H&M use cases adopted ML workflow using airflow, kubernetes and docker and how to solve the provisioning problem with ML approach.

At H&M we are using Airflow, kubernetes as main components for the machine learning workflow. The increase of the Online shopping during the last two years has impacted the data volume significantly. A lot of companies are struggling with the infrastructure cost when adopting airflow kubernetes/docker as technologies, any person interested can join to have a high level explanation of the solution H&M Group has adapted to encounter this.

Live broadcast: https://www.youtube.com/watch?v=UujU3xOo038 What are the essential software engineering skills a datascientist should have to succesfully bring own work to production? We - Sergei Beilin, Ph.D., software engineering consultant in AI/ML, and his wife Natalia Beylina, Ph.D., datascientist - will go through the most important things a modern datascientist needs to know about software engineering, from both software engineer and datascientist point of views, and using our own experience. We will discuss: * programming language(s): how much of the language should one know? * execution models, orchestration, containerization - kubernetes, kubeflow, airflow, spark/databricks, etc * storage, network protocols/APIs, file formats - from CSVs to delta, from json to avro * modern systems architecture concepts to understand * and how the whole system architecture and infrastructure landscape will dictate the way you deploy and run your work * tools and devops practices * processes: integrating data scientists' workflow into typical agile * bad practices to avoid: a few examples we've seen ourselves

Data science went from universities and research labs to small to big commercial companies in different business areas. From experimentation phase it's going to production and not everyone knows how to build teams around datascience projects, and datascientist need to know more about software engineering, especially when they have to work a lot alone, without proper support from software engineers. No longer is data science just some experimental code, and no, a jupyter notebook is not enough. The industrialization of data science required more, broader skills. We - Sergei Beilin, Ph.D., software engineering consultant in AI/ML, and his wife Natalia Beylina, Ph.D., datascientist - will go through the most important things a modern datascientist needs to know about software engineering, from both software engineer and datascientist point of views, and using our own experience. We both have Ph.Ds in mathematics and worked for quite some time in research and education, so at some point we had this experience of "research to business" mindset shift. In this talk we tried to collect our experience of working in different datascience projects and companies as well as helping others move to data science from different fields.

Live stream: https://youtu.be/kO5Es7KKUIY This talk provides a hands-on deep-dive into the wheel file format and python packaging. First, we will slash the tire, see what's inside, and then build new wheels from scratch. You will learn about the inner workings of a crucial part of the Python packaging ecosystem and understand what your tools do behind the covers.

During this talk, we will explore python packaging. Before we begin our deep dive, I will provide some background and information guiding the audience into where we are today and how we ended up here. First, we will deconstruct a package and deep dive into the file format. Next, we will learn what makes up a python package, how to create a package from zero, and how to create a small tool from scratch to package up a python project and submit it to PyPI. Closing up the talk, I will introduce the reader to modern tooling that makes the process painless. The presentation aims to educate the audience in python packaging and demystify what these tools do behind the scenes. In the end, I'd like to have a Q&A summing up questions. I will publish all the code and content in a GitHub repository. The idea is that said repository can act as a handy reference guide for the audience whenever they need to look something up related to the python packaging format. Rough outline, this might be altered slightly - Introduction - Presentation of me - Outline - Why do we care? - How? - Cheatsheet - What is a Python Package? - Wheels - Distribution types - Crack the egg - What's wrong with sdist? - Deconstruct a package _(practical/demo)_ - Build a package from scratch _(practical/demo)_ - Package it up _(practical/demo)_ - Submit to (test) PyPI _(practical/demo)_ - Modern tooling - More resources - Q&A

Live Stream: https://youtu.be/5EWHYxnOTyQ Not fading away: A tale about a 20-year old Python project By Érico Andrei, Python Software Foundation Fellow, Plone Foundation Vice-President

The Plone CMS is one of the oldest, and most successful open source stories of the Python community. Created by Alex Limi and Alan Runyan in 1999 to be a better UI for Zope, the project grew to be a very stable and secure solution, used by governments, corporations and NGO's to power their public sites and intranets. The community surrounding Plone has been the key to keeping the project alive. This talk will focus on the community itself, and how meeting technical and organizational challenges it has adapted and evolved to avoid fading away.

Live Stream: https://youtu.be/qWvJSIgOcPU With a lot of changes under the hood with Airflow 2.0, the workshop aims to give an overview on major updates in Airflow 2.0 from 1.0, major components and working of Airflow and hands-on demo of implementation and management of an end-to-end Machine Learning pipeline. Without a pipeline in-place, management of multiple Machine Learning stages in production can be difficult. This gives an overview of simplified process and management of Python based ML projects using Airflow.

## Prerequisites 1. Install [Docker Desktop](https://www.docker.com/get-started) (with minimum 3GB memory allocated) 2. Start Docker engine 3. Clone the workshop repo with `git clone https://github.com/pycon-ml/airflow_workshop.git` 4. Run `docker-compose pull` inside repo folder `airflow_workshop` ## Agenda - 05 min: Introduction - 05 min: Major changes in Airflow 2.0 - 05 min: Pre-requisites setup overview - 10 min: Walkthrough of different backend components - 10 min: Different stages of a DAG file – steps and operators - 10 min: Dynamic DAG creation to improve parallelism - 15 min: How to trigger Airflow DAG runs - 15 min: Debug and clear Airflow task errors - 10 min: Overview of production-level Airflow-based architecture - 05 min: Wrap up questions

Live broadcast: https://www.youtube.com/watch?v=gwLJZVoXWlg Apache Airflow has become one of the most popular data toolings. Due to its high complexity, it could be challenging for all teams and companies. For example, how to effectively construct an orchestrate architecture on diverse cloud platforms, how to productively accelerate your engineering and machine learning workload at scale, and how to smartly decouple your Python codebase for professional testing and easy maintenance.

In this session, we will present five recipes that have helped us scale data jobs and to reinforce as a data-driven AI-leading company. Key take-ways: - Orchestrate Architecture - Auto-Build Airflow DAG - Data Quality - Auto-Cost Evaluation - Auto-Cataloging

Live Stream: https://youtu.be/rka62mJ-vfo A discussion with companies' recruiters from different areas about the expectations on python programmers, the trends, and the difficulties nowadays.

A discussion with companies' recruiters from different areas about the expectations on python programmers, the trends, and the difficulties nowadays. Participants are Björn Hertzberg from H&M, Magnus Perman from Nexer, Carolina J. Säll from 46elks, Lisa Mellor from Klarna and Chris Valle from Funnel.

Live broadcast: https://www.youtube.com/watch?v=EV7SkhRxemA How can you show what a Machine Learning model does once it's trained? In this talk, you're going to learn how to create Machine Learning apps and demos using Streamlit and Gradio, Python libraries for this purpose. Additionally, we'll see how to share them with the rest of the Open Source ecosystem. Learning to create graphic interfaces for models is extremely useful for sharing with other people interesting with them.

How can you show what a Machine Learning model does once it's trained? In this talk, you're going to learn how to create Machine Learning apps and demos using Streamlit and Gradio, Python libraries for this purpose. Additionally, we'll see how to share them with the rest of the Open Source ecosystem. Learning to create graphic interfaces for models is extremely useful for sharing with other people interesting with them. Some demo examples are: - https://huggingface.co/spaces/flax-community/dalle-mini - https://huggingface.co/spaces/flax-community/chef-transformer - https://huggingface.co/spaces/nielsr/LayoutLMv2-FUNSD

Live stream: https://youtu.be/tHlMw9zFgQE Popular programming language index websites (TIOBE index) and developer surveys (Stack Overflow) place Python as one of the fastest-growing programming languages. However, this popularity also puts in the target range of attackers. The attackers perform malicious dependency attacks and use misconfiguration tools to reveal confidential information. In this talk, we will discuss identifying common security issues in Python code and handling malicious dependency attacks using safety.

Popular programming language index websites (TIOBE index) and developer surveys (Stack Overflow) place Python as one of the fastest-growing programming languages. However, this popularity also puts in the target range of attackers. The attackers perform malicious dependency attacks and use misconfiguration tools to reveal confidential information. Jukka Ruohonen, Kalle Hjerppe, and Kalle Rindell in their research paper "A Large-Scale Security-Oriented Static Analysis of Python Packages in PyPI" claimed that they scanned PyPI for security issues in Python packages and found the presence of at least one security issue in about 46% of the Python packages. In addition, security vulnerabilities can be present in the source code of the package. In this talk, we will address the security issues related to python packaging and possible solutions to make python packages secure. The talk begins with the importance of a secure package and vulnerabilities in the Python package index. Then, I will discuss Python packages such as Bandit for identifying common security issues in Python code and “safety” for dependency check. Next, I will discuss verifying and signing Python packages using GPG. Finally, I will discuss general guidelines for secure coding practices in Python. Outline 1. Importance of a secure package and vulnerabilities in python package index. (05 Minutes) 2. Bandit for identifying common security issues in Python code (03 Minutes) 3. safety for dependency check (04 Minutes) 4. Verifying and signing PyPI and conda packages using GPG and Twine(05 Minutes) 5. General guidelines for secure coding practices in Python (05 Minutes) 6. Summary and Questions (03 Minute )

Live stream: https://youtu.be/W1qOPma747k Configuration files used to manage Infrastructure as Code are traditionally implemented as YAML or JSON text files and are missing most of the advantages of modern programming languages. Wouldn't it be better to use the expressive power of Python to define your cloud infrastructure? The AWS Cloud Development Kit (AWS CDK) is an open-source framework from AWS that enables developers to harness the full power of modern programming languages to define reusable cloud components and provision applications built from those components using AWS CloudFormation. In this session, we'll quickly cover the basic concepts of the CDK, and then we'll spend the majority of our time building a serverless application with the CDK. We'll show you how to use the CDK to assemble your AWS infrastructure using the Python CDK quickly.

Configuration files used to manage Infrastructure as Code are traditionally implemented as YAML or JSON text files and are missing most of the advantages of modern programming languages. Wouldn't it be better to use the expressive power of Python to define your cloud infrastructure? The AWS Cloud Development Kit (AWS CDK) is an open-source framework from AWS that enables developers to harness the full power of modern programming languages to define reusable cloud components and provision applications built from those components using AWS CloudFormation. In this session, we'll quickly cover the basic concepts of the CDK, and then we'll spend the majority of our time building a serverless application with the CDK. We'll show you how to use the CDK to assemble your AWS infrastructure using the Python CDK quickly.

Live broadcast: https://www.youtube.com/watch?v=9fwOBMWRTiI Facial recognition has been a challenging task for a long time. Nowadays, we can reach and pass the human level accuracy with deep learning based state-of-the-art models. In this talk, you are going to learn how to build highly scalable facial recognition pipelines in python programming language with DeepFace library from its creator. DeepFace is the most lightweight facial recognition and facial attribute analysis (age, gender, emotion / facial expression, race / ethnicity) library for Python. It wraps many state-of-the-art face recognition models: VGG-Face, Google FaceNet, OpenFace, Facebook DeepFace, DeepID, Dlib and ArcFace. Experiments show that human beings have 97.53% score on LFW dataset whereas VGG, FaceNet, Dlib and ArcFace are passed that level already. Besides, OpenFace, DeepID and DeepFace have a close score as well. You can also build and run any one those cutting-edge models with just a few lines of code. The library got almost 2K stars on GitHub and 200K installations on PyPi / Pip.

Facial recognition has been a challenging task for a long time. Nowadays, we can reach and pass the human level accuracy with deep learning based state-of-the-art models. In this talk, you are going to learn how to build highly scalable facial recognition pipelines in python programming language with DeepFace library from its creator. DeepFace is the most lightweight facial recognition and facial attribute analysis (age, gender, emotion / facial expression, race / ethnicity) library for Python. It wraps many state-of-the-art face recognition models: VGG-Face, Google FaceNet, OpenFace, Facebook DeepFace, DeepID, Dlib and ArcFace. Experiments show that human beings have 97.53% score on LFW dataset whereas VGG, FaceNet, Dlib and ArcFace are passed that level already. Besides, OpenFace, DeepID and DeepFace have a close score as well. You can also build and run any one those cutting-edge models with just a few lines of code. The library got almost 2K stars on GitHub and 200K installations on PyPi / Pip.

Live Stream: https://youtu.be/BgzIaEzXEBU Many times we have to write Python extensions, particularly in C. To do various system operations, or doing calculations in a faster manner. But, writing safe C code is always difficult, even for an experienced developer. This is where writing Python extensions in Rust is becoming more popular among developers where people think about speed and security at the same time. In this workshop we will learn about how to create a Python module using Rust. No previous Rust experience is required.

If you are using latest [cryptography](https://cryptography.io/en/latest/) in any of your project (which you most probably already do), you are using one of the most powerful and trusted Python module where a part is written in [Rust](https://www.rust-lang.org/). In this workshop we will go through a given git repository (no prior Rust knowledge is required) and start building a Python extension module step by step. ### Outline - Initial module creation - single function - functions with arguments - Help documentation - Functions to read files - Exception generation - Dictionaries, lists - Creating your own class - A module with some real life work done in Rust We will follow prewritten code for most of the sessions, I will ask you to modify those as exercises during the session.

Live broadcast: https://www.youtube.com/watch?v=cVEkJqcmivQ Marketing attribution is one of the trickiest problems to crack for data scientists working with marketers. To reach potential customers one needs to measure the value of campaigns and channels that the customers interact with. It's easier said than done. One solution to this problem is through the Markov chain. We will see how we can implement the markov chain for channel attribution.

For any organization, measuring the value of the campaigns and channels that are reaching their potential customers is very important but not very straightforward. Data scientists/ Analysts can help in these scenarios . Data-driven attribution models can eliminate the biases associated with traditional attribution mechanisms, and understand how various messages influence potential customers and the variances by geography and revenue type. 1. Introduction to marketing attribution **(3 Minutes)** * Overview * importance * Challenges 2. Introduction to Markov chain model **(8 Minutes)** * Overview * States of the Markov Chain Model * Transition Probabilities 3. Implementation of markov chain model **(12 Minutes)** * Data Preprocessing * Calculate the Transition Probabilities * Calculate removal effects * Interpretation and Prediction * Assumptions and Limitations of Markov Model 4. Conclusion and final remarks **(2 Minutes)** * Summarizing what we discussed and discuss other sources to increase the knowledge 5. Q&A **(5 Minutes)**

Video link: https://youtu.be/lWfJfviWIBU The OWASP Top 10 is a book/referential document outlining the 10 most critical security concerns for web application security. In this talk, we will see how underlying security in Django, protects it against OWASP top 10 vulnerabilities, ranging from SQL injection attacks to authentication and CSRF. It is one of the most complex yet interesting topics in Django that makes it an extremely powerful web framework.

As a web developer, using a framework that guarantees security is always great but it’s even better to measure all the vulnerabilities involved while building your application and to also know how to protect yourself from them. The Open Web Application Security Project (OWASP) is an online community that provides the top 10 vulnerabilities in web application security based on what security experts see while performing penetration testing. These vulnerabilities range from SQL injection attacks to authentication and CSRF and Django was built to minimize those security risks and give developers the ability to avoid and reduce those vulnerabilities by themselves by using better practices. It offers many security-minded functions right out of the box, without sacrificing ease of development and integration with both front-end and back-end components. We will share what we have learn so far and encourage you to try it with your own projects. We'll walk through a simple example, with screenshots and code wherever required.

Live broadcast: https://www.youtube.com/watch?v=j90tdZyK6FA The move to cloud has opened a world of new possibilities in software development. It's so easy to spin up resources in the cloud and together with the adoption of DevOps, software developers are more empowered than ever before. Of course this also puts more demand on the software developers, to take full control and have knowledge of the complete cycle from depolying infrastructure to develop and deploy code. Luckily this process has a lot of benefits and is less reliant on skills of key-persons, if infrasctructure can be deployed as code, this can also be automated with different tools. The end goal is to be able to deploy more code enhancements and at the same time benefit from the rapid pace of hardware and cloud improvements.

For the compute heavy Data Science practice the adoption of Cloud and the flexibility of deploying enviroments has become a vital success factor. You can have a "supercomputer" at your fingertips for a short time and then you can decomission it again when your work is ready. But to be able to use this approach over and over again with the same configuration the actual infrastructure need to be saved in code. Of course this can be done with the help of Python and ideally this should be automated in a true DevOps and CI/CD manner. I will walk through some of my key take aways of working with infrastructure as code for Data Science projects.

Live broadcast: https://www.youtube.com/watch?v=iw9uS8yLax8 Machine learning is not only an interesting technology to use today, but it’s also appreciated by management that will hear that the organisation is using “machine learning” to solve time series challenges, such as demand planning with supply chain management. However, this can result in time spent on complex modelling that in general can be accomplished quicker with much simpler models that are easier to deploy and sustain long-term. Therefore, in this talk we'll show how simple can not only give better results while reducing the complexity in terms of data pre-processing, model development and final deployment. We will look at an example within supply chain management and demand planning for a product and discuss different scenarios based on multiple types of historical demand data. The presentation will show the actual code, but a big focus will be on the strategic decision-making of selection of models and how to deploy these models.

Description We break down the talk into four components: 1. “The problem” - The first 5 minutes are about understanding the problem before diving into the code (2 min context of the time series challenges within demand planning in large organisations today + 3 min on time series forecasting and machine learning vs classical statistical models including the importance of good benchmark models) 2. “The setup” - The next 5 minutes are about getting set up correctly on how to analyse this before we test our models (2 min walkthrough of the structure & plan for the code (python & jupyter notebooks + 3 min in terms of data pre-processing and success metrics (comparison to benchmark) 3. “The Models” - Then the next 10 minutes are to select and test models (3 min model selection and explanation + 2 min running models + 5 min explaining & discussing results) 4. “The Deployment” - The final 5 minutes are about deployment and what the pros and cons are with these, depending on the organisation.

Live Stream: https://youtu.be/4z0ivHhX4h0 Words from the organizers of PyCon Sweden 2021.

Words from the organizers of PyCon Sweden 2021.

Live Stream: https://youtu.be/hv-jKovzeHI Managing cloud infrastructure as code in Python By Alexey Isavnin, senior software developer at Elisa Automate, founder of Rays of Space company

Python is a language of choice for a wide spectrum of software products nowadays. It is applied in APIs, data science models, high-performance computing, and more. But, software also needs hardware/infrastructure to run on. The good news is that you can provision and manage cloud infrastructure of any complexity without leaving the comfort of Python. In this talk, we will learn how to do exactly that while discovering the added benefits of using Python for this task. We will mainly focus on AWS but will touch other major clouds as well.

Live stream: https://www.youtube.com/watch?v=Ew4tKVem6F8 In this talk, we aim to find if polarization is induced in a neural network by feeding it newspaper articles with manufactured sentiments according to the Allsides Media Bias chart for the level of faith people on various aisles of the political spectrum. This project consists of a set of experiments on similar data-sets from news agencies across the various subsets in the ”media-bias” chart. News Media perceived bias is common across consumers that belong to various political affiliations. While anecdotal evidence of this exists and there exist annotated datasets that aim to annotate the ”spin” a news agency puts on certain events and entities, whether this is a widespread problem and whether it can be detected by the neural network topically or temporally is a problem that needs to be explored. The news media bias analysis is modelled as a Natural Language Processing sentiment analysis task and a fake news binary classification task to deduce the level of polarization in a neural network by feeding it headlines embedded using pre-trained sentiment models from news publications across the political spectrum. When it came to fake news vulnerability, news from all kinds of perceived politically affiliated news media holds up well against a fake news dataset with a very good accuracy. None of the accuracies dropped below 95%. This is a significant result that sort of debunks the AllSlides Media categorization - if taken as simplistically as it is presented. These experiments can be extended to include entity based topical studies in the future and to also educate the populace about their perceived biases.

As social media sites across the world grapple with hate speech, calls for genocides and sexual harassment on their platforms, as policy scientists look up the various biases in our justice system’s usage of language and as most of the people in the world struggle with what is globally called ”media bias”, I believe that as mathematics and statistics became commonplace measures, so will Machine Learning models. This work is an example of an intersection of a non scientific field with computer science and mathematics, trying to quantify, measure and identify non mathematical phenomena in the language of mathematics. It is important because it could be the basis of the scientific approaches that the next generation policy makers, voters, non profit social organizations and governments could use to make life changing decisions for their citizens. 2 The questions that this study tries to answer is whether a neural network can learn biases from the news media based on perceived bias scores obtained from independent agencies. It also seeks to answer whether any of these political leanings of the news media affect the vulnerability of their consumer when it comes to fake news. The results of this experiment aim to show Conclusions 1. SVMs perform better clustering with respect to the categories than neural networks, however the maximum does not cross 67% 2. The most significant conclusion from this work is that though there is a perceived bias when it comes to news agencies, when looked at from a neural networks standpoint, it is negligible. Mainstream news agencies are not able to polarize a neural network with inherent biases in their headlines. 3. There may be topical biases that need to be examined by using an Entity linking and bias calculation approach 4. Most mainstream news agencies do not make the consumer vulnerable to believing fake news. This study needs to be conducted with data from popular social media ”news” groups or popular TV shows that masquerade as news but may technically not even be news channels. 5. It is safe to conclude that the perceived bias that stems from social media polarization is being extended to news media when their contribution to the polarization may be negligible. Significance of Work 1. The significance of this work is to be able to transform a social problem into a technical one and using neural networks and Machine learning techniques to try to gain some insights. 2. Hopefully using these techniques to find deeper trends will become mainstream and help policy makers and the general citizens approach news media bias in a better light. Future Work Some further studies to take up are as follows: 1. Effect of news media on Perceptron Networks 2. Better Annotated Datasets to perform bias analysis 3. Effect on memory models of media bias 4. Experiment on some of the most polarizing news epochs in time 5. Studies on country level news bias

Live stream: https://youtu.be/Sp0tEGqFfN8 Different programming languages have different functionality, different paradigms, and different styles. We have certainly seen other low-level languages like C++ adopting more “pythonic” themes in recent years, like foreach loops. But what about the opposite? Did you ever wonder how we could implement a smart pointer in Python? Whether we can we add Java’s final keyword for real constants? What exactly the inspect module is useful for? How we get private methods in classes? We will take a deep dive into Python's fundamentals to discover how you can make things like C++-style input/output, like cout << "Hello world" << endl; or cin >> my_var;, a reality in Python, using the exact same syntax. And, of course, why you really, really shouldn't. What exactly does pythonic mean? What makes python what it is today? Hint: It’s about more than just the walrus operator.

Brief outline of the talk: - Short comparison of Python with other common languages like C++ or Java, and pythonic features they've received in recent years - Recent updates to Python, and why we don't need to manually do those anymore - like match-case and walrus operator. - Implementing std::cout and std::cin in python using special classes. This is actually quite simple to do. - Implementing final decorator and type hint to achieve "constants". This can already be found in libraries on pip. How to break them. - Marking methods as private - this can be done through decorators. - Why you really shouldn't be doing this – What is considered pythonic? What does it mean? Why are other languages like they are? Will include code examples, using the inspect module, some python hacks, and other ideas on the same topic. Obviously with a comedic touch, but the main idea is to educate on python fundamentals and differences in programming paradigms.

Live Stream: https://youtu.be/Iv9KA2JWwVw Join Carolina & Victoria, developers at 46elks, for a code along workshop 👩🏻‍💻
We will be building an answering machine with Flask. Using Python & 46elks you can setup your very own answering machine. What you need to follow this code along: - A 46elks account, [here's a link](https://46elks.se/register/pycon-2021) with some credits to test your answering machine - A computer and be excited to code some cool stuff 👩🏻‍💻 We will be coding together for about 60 minutes and then we'll answer any questions you might have (literally, ask us anything), or just hang, getting to know new developers friends 🥳

What will we be accomplishing during this code along:
When you receive a phone call, if you're not available the caller hears a recording asking them to leave a message. If the caller leaves a message, the audio file is downloaded to a specific folder. When you are available, the call is forwarded to your real mobile number 📞

Live stream: https://youtu.be/Lsi1ZhmbNDc Stuck in a deeply nested if...else when traversing the pyramid of doom, you pause for a minute to catch your breath. The program’s logic eludes you and it is getting increasingly tiresome to keep track of all the twists and turns of the various conditions and possible return values. You start to dream a dream of a flattering flattening of all this code. A dream of refactoring this bewildering maze into an orderly space, devoid of surprising and unexpected behaviour. A space where things have their obvious place and purpose. You decide that you just just might need to set some rules. Enter the Rules Engine.

This talk introduces the concept of rules and rules engines as well as a way of implementing this in Python. Building on top of the initial, simple implementation, it is then further demonstrated how to expand it to handle additional use cases that goes beyond the basic aim of refactoring nested conditional constructs such as if...else, including parallel execution of your rules. The talk is suitable for anyone using Python as the concepts are pedagogically introduced and at the same time the overarching concept is a powerful tool.

Live stream: https://youtu.be/YnxG8jABqaU A brief overview on how the python ecosystem can be used to build things that would help you to boost your skills and build a next major/minor version of yourself. I'll showcase a few approaches and mathematical model for motivation and how one can build tools to help you lower the resistance of doing those things you think you need to be doing more often

There is a trend of modern time-tracking and self-improving screaming at you from every corner and messing up with your plans to watch the latest season of Rick & Morty. In this talk I want to show how one can use python [and a few other middle-code/low-code tools] so that you can have easier time tracking your habits and personal skills and plans. We will go through building a simple dashboard and how one can approach a side project like that

Live stream: https://www.youtube.com/watch?v=0a0c-aMj1Xs Optimization libraries such as SciPy or Nevergrad are commonly used in different data science workflows, such as choosing optimal hyperparameters for a machine learning model or taking actions based on forecasts. In this presentation, we will discuss how such an optimizer can be used to build reward configurations for games (by rewards configurations here we mean bundles of different in-game items that players may get for completing different tasks/quests in a game) Using rewards in Candy Crush Soda as an example, I will show how the problem can be solved using the Nevergard library from Facebook.

In this talk, we will go through several aspects of the end-to-end optimization workflow: - What is optimization - The different Python libraries that can be used to perform optimization in practice - How to optimize an in-game economy - How this task differs from other optimization workflows (e.g. tuning hyperparameters etc.) - Case study: using Nevergrad to create optimal reward packs in Candy Crush Soda

Live Stream: https://youtu.be/oIWBW2usic8 The best way to learn Python - for the absolute beginners and improvers by Cheuk Ting Ho, Developer Relations Lead at TerminusDB, Python Software Foundation Fellow, organizer of EuroPython

In the 2021 StackOverflow survey, Python remains the most wanted programming language (5 years in a row). As a Pythonista, I have been asked many times about what is the best way to learn Python. There are many resources, especially online, free or paid, that is available. From working as a data scientist to become an open-source developer, I would like to sum up my Python learning experience to make recommendations to different groups of people. Whether you are an absolute beginner with no programming experience, a beginner with other programming language experience or a beginner in Python who wants to get better and know more about the language, you will get an idea of how to use Python at a whole new level.

Live stream: https://youtu.be/y0PTozH9mZs The purpose of this talk is to share the work as a professor of the Bachelor of Information Systems at the University of Minas Gerais (using the Python language to teach Object Oriented Programming II). We are going to talk about the difficulties encountered by students in learning this subject and how we managed to overcome it with the use of a modern language with a shorter learning curve and how this can contribute to a lower dropout rate from the course. Difficulties encountered, pedagogical approach used, exercise practices performed with students.

In this conversation, we will talk about the seven steps of the methodological proposal to practically implement the use of Python, increase student engagement and make use of real examples and practices: Methodological Proposal Lectures 1. Easy to get started 2. Clear problem definition 3. Growing difficulty 4. Collaborative work 5. Market Practices 6. Multidisciplinary 7. All open and published 8. Online classes

Live stream: https://www.youtube.com/watch?v=Y9OPX75ax0M Controlled experiments such as A/B tests are a gold standard for determining whether changes to a website significantly impacted user behaviour, however they are not always possible. In this talk we walk through a iPython Notebook and describe a non-parametric method for determining whether changes to e-commerce product pages impacted conversion to basket without the use of controlled experiments.

Controlled experiments such as A/B tests are a gold standard for determining whether changes to a website significantly impacted user behaviour, however they are not always possible. In this talk we walk through a iPython Notebook and describe a non-parametric method for determining whether changes to e-commerce product pages impacted conversion to basket without the use of controlled experiments.

Live Stream: https://youtu.be/gnFzZRkQZ2c This workshop will demonstrate a zero-to-hero tutorial on how to solve a classification task using deep learning. The tutorial kicks off demonstrating a simple classification task on synthetic data, first in low and then in high dimension. Then, a harder classification task based on FashinMNIST, a famous dataset containing images of clothes, will be tackled. Apart from solving the classification task itself, we will show how to generate and analyze embedding vectors that can be used to solve other downstream tasks, different from the original classification problem on which the model was trained. Finally, we are going to face a more advanced type of classification problem, namely, predicting links on a graph using Graph Neural Networks. Link prediction will be demonstrated on an open source dataset that contains information about collaborations among authors of scientific papers. The target of this workshop is to show how we can use Python to solve the the aforementioned tasks, taking into account both the data science aspects and the engineering and project lifecycle related ones. In particular, the python packages that we are going to cover in the workshop are PyTorch, PyTorch-Lightning, Deep Graph Library.

A Zero-To-Hero workshop that will demonstrate how to solve classification tasks on datasets and tasks of increasing complexity. The workshop will present both how to solve the tasks and how to structure a codebase according to software development best practices. The first challenge presented in the workshop will be the classification of synthetically generated Gaussian blobs. First, we will be classifying low-dimensional Gaussian blobs and then we will extend the algorithm to higher-dimensional blobs. Moreover, the demo will also showcase Tensorboard as a tool to monitor model learning. The model presented in this initial part of the tutorial is the Fully Connected Multi-Layer Perceptron (MLP), the most well-known type of neural network. The second challenge will be the classification of garments from images. To this end, we will use the well-known FashionMNIST dataset. The model presented in this second part is the Convolutional Neural Network (CNN), the mainstream neural network for image analysis. After solving the classification tasks, we will demonstrate how to obtain numerical embeddings from the CNN that can be used to solve a multitude of downstream tasks. We will demonstrate how these embeddings can be used to find and cluster similar items. In the third and final part of the workshop, we will work with graph data and try to predict whether two authors of scientific papers are co-authors or not. We will demonstrate this task using the collab open source dataset, that is, a graph where the nodes will represent the authors and the edges will connect the co-authors. The problem we will solve is framed as a link prediction task, that is, to a certain extent, analogous to a classification task since we are going to try to predict whether an edge exists or not between two nodes. The model that will solve this task consists of two stages. The model presented in this final part of the tutorial is the Graph Neural Network (GNN), a powerful formalism for analysing graph data. The models are going to be implemented using PyTorch-Lightning, that enforces a modular and maintainable software structure.

Live stream: https://youtu.be/cuNEOtLbB14 The Robot Operating System (ROS) is a set of software libraries and tools that help you build robot applications. In this talk, we will discuss how to create your own custom robot and simulate it in Gazebo along with ROS. We will also learn to add cameras and other sensors which will enable us to move the robot and perform image processing using python.

The point of using ROS is to create a robotics standard, so that you don’t need to reinvent the wheel anymore when building a new robotic software. In this talk, you will learn to create a robot from scratch using URDF and XACRO files. This will enable us to define the inertial, visual and collision aspects of the robot and also we will be able to integrate sensors into it. The method discussed will be very generic and can be used in other bots as well. Along with building robots, we will also learn to create a new world in a gazebo where we will spawn our robot. This robot will then be controlled using python and we will also capture the video coming from the camera and do image processing over it in python. This will be a major building block towards building smart robots.

Live stream: https://www.youtube.com/watch?v=bZjHWgLnWs8 What can a developer teach a data analyst about data analysis? A few lines of Python code may be enough to solve a tricky data cleaning challenge. Functions can stop you from getting lost in many copies of very similar code. Tips for writing larger programs without tearing your hair out. Start writing code which is still useful in years to come, and which evolves without degrading into a big mess I will share examples of how I've used pure Python in my data analysis and give you simple tips on applying software development best practices to your code.

Pandas, MatPlotLib and scikit-learn are fantastic libraries. Glue them together with a little bit of Python and you can do so many things. Your favourite search engine fills in the gaps when you're stuck. You wield your favourite weapons in the war against meaningless data with ease and style. So what can you learn from someone who started coding in the late '70s? What could an experienced Python trainer/engineer possibly know that you can't find online yourself? Whilst doing data analysis projects I regularly drop back on my Python and computer science knowledge. Sometimes a few lines of code will be enough to solve a tricky data cleaning challenge. Knowing how to write functions stops me from getting lost in many copies of very similar code. Combining the ease of Jupyter cells with the rigours of clean code lets me write large programs without tearing my hair out. Doing a quick analysis is easy. Writing code which is still useful in years to come, which evolves without degrading into a big mess, takes experience. I will share examples of how I've used pure Python in my data analysis and give you simple tips on applying software development best practices to your code. Let's learn from each other. Telling people with much more data analysis experience than myself how to do their job feels a little scary. There may be a better way that I've missed. If so, please tell me after the talk - just be gentle - thanks ;-)

Live stream: https://youtu.be/ji8wYJE0c1I Taking ideas to market faster remains key to any good DevOps strategy. Boilerplate application code, configurations and Infrastructure-as-Code (IaC) are the key components that enable this. Leveraging template engines to build these is an effective strategy to enhance your speed. Aligning with minimalism and keeping things agnostic, the talk shares a simple and easy to use code base to generate all of these. Most organizations effectively manage boilerplate code with Git based services. However, it does not solve the question of "what to customize?" once you have the code cloned. This is where customizable templates are key in identifying the customizable bits and injecting with the right parameters/data with ease. About the speaker: Raza Balbale is currently a Snr. Architect/ Manager at Cognizant Technology Solutions, US - part of the Connected Products BU's Product Engineering Team. He frequently uses Python as part of his DevOps / acceleration toolkit.

Completely revamping the solution accelerator demoed at PyCon Sweden 2020, this year's talk is focused on taking a closer look at a lighter, more slick, easier-to-manage and scale framework, "Zaunic". Zaunic breaks down the code into "templates" and "data" that is injected into templates. Multiple files and project folders can be now easily generated by using YAML based books. The books make the whole process repeatable and easier to manage. Teams working towards onboarding new team members or helping existing ones, should find this intuitive. The demo shows how a Python based framework is being leverage to generate Elixir application code and Infrastructure as Code. For teams that intend to follow maintain high level of consistency and strict patterns in terms of code/resource nomenclature, this is an extremely important useful tool. Feel free to pull down code, experiment and add your own books and templates. The project contains samples to get started: https://github.com/razaibi/zaunic

Live stream: https://www.youtube.com/watch?v=rqy3OZn4y-4 The cutting efficiency of a chainsaw is related to the hardness of the wood, For example, it is affected by the existence of knots (hard structure areas) and cracks (no material areas). The current practice involves clean cuts by avoiding knots and cracks. Therefore estimating the relative wood hardness by identifying the knots and cracks beforehand can significantly automate the process of regulating the chain properties, e.g., consumed power, force, etc., which in turn improves the chain's efficiency. In this talk I will share how I have implemented Mask-RCNN to identify and segment defects in wood cuts and how the result can be used to understand wood hardness to improve cutting efficiency of chainsaw.

Wood cutting properties for the chains of chainsaw is measured in the lab by analysing the force, torque, consumed power and other aspects of the chain as it cuts through the wood log. One of the essential properties of the chains is the cutting efficiency which is the measured cutting surface per the power used for cutting per the time unit. These data are not available beforehand and therefore, cutting efficiency cannot be measured before performing the cut. The efficiency of the chainsaw is closely related to wood hardness and defects like knots and cracks are very important properties when measuring wood hardness. Mask-RCNN is a widely used machine learning model in computer vision that is used to perform instance segmentation. In my work Mask RCNN was used to identify each instance of knots and cracks in a wood cut and then the instance information was used to understand wood hardness. OpenCV was used to perform image processing, open source platform tensorflow and libraries like sklearn, matplotlib, numpy were used to implement the model, perform the tasks and visualise the result. Therefore, I think it can interest other python and machine learning enthusiasts to know about my work.

Live Stream: https://youtu.be/SRqr8OSY0oU Words from the organizers of PyCon Sweden 2021.

Words from the organizers of PyCon Sweden 2021.