An introduction to the **correlate** Python library.  You tell **correlate** about two datasets that should map to each other, and it determines the best matches for you.  The novel scoring algorithm at the heart of **correlate** means it copes exceedingly well with messy real-world data.  **correlate** supports fuzzy matching, weighted matching, and ordering.

"correlate" finds good matches between two sets of messy real-world data!  How?  Come find out!

Data correlation!  What could be more computer science-y!  Ever needed to find matching items between two sets of data?  Maybe even messy real-world data, with inexact string matches?  Come find out how the novel scoring algorithm and clever heuristics at the heart of **correlate** solve this problem with ease!

Larry Hastings

Correlating messy data with "correlate"

A lightning talk (LT) is a short presentation that must not be longer than five minutes.

To sign up for a lightning talk, you can put your name on the information board during the conference before the second coffee break. For our online participants, we will set up a separate form or Google sheet for you to put your name and topic in - similar to how we run this at the in-person conference. 

We will announce the same every day both online and in person.

Lightning Talk Lightning Talk Lightning Talk Lightning Talk Lightning Talk Lightning Talk Lightning Talk Lightning Talk Lightning Talk

Signing up is on a first-come-first-served basis. The queue is reset every day in the morning.
You may talk about / present (almost) everything with certain boundaries:
- no promotion for products of companies, no call for 'we are hiring' (but you may name your employer).
- conference announcements are limited to one minute only
- one LT per person per conference policy: everyone who has not given an LT at this year's conference, yet, is ahead of you in the queue
- keep it short, the shorter the presentations, the more people can give a talk

See you there and have fun!

Thursday's Lightning Talks

When processing data, validating its structure and its type is critical. Bad record types or changes in structure can often result in processing errors or worst in wrong data output. Yet, solving this problem cleanly and efficiently can be challenging. It often results in complicated code logic and increases complexity; consequently decreasing code readability. Pydantic is an efficient and elegant answer to these challenges

We expect you'll leave this talk with a good understanding of:

- Existing challenges in data validation
- What Pydantic Models, Validators, and Convertors are
- How to leverage Pydantic in your day to day (using real-life examples)
- [Bonnus] How to use Code Generation to create Pydantic Models from any data sources

Leverage the power of Pydantic to validate your data efficiently

Teddy Crepineau

Simple data validation and setting management with Pydantic

How annoying is it to find out that everything went to hell on the pipeline because you forgot to run the formatters? 

Don’t waste precious time and learn how it is possible to automate these little things, but most importantly understand why it is important to have them in your code!

Don't waste your time and learn why you should set up automations to run formatters automatically when you commits. Learn why they are important to improve your code!

Have you ever been in a situation where you check your pipelines to see if he’s finished running the tests and you find that he hasn’t even started because you forgot to run the linters locally?

I contribute to the development of [PyCon Italia](https://pycon.it), [Strawberry GraphQL](https://strawberry.rocks) and Poetry Package Manager and coding with these organizations I discovered fantastic tools such as [pre-commit](https://pre-commit.com/), [Poetry](https://python-poetry.org) and many others. In this talk I would like to share with you my findings, think about the pros and cons and understand a little bit better why you should start using them.

[Slides available here :)](https://www.slideshare.net/EsterBeltrami1/automate-cleaning-code-in-few-easy-steps)

Ester Beltrami

Automate cleaning code in few easy steps!

JSON Web Tokens, or JWTs for short, are all over the web. They can be used to track bits of information about a user in a very compact way and can be used in APIs for authorization purposes. Join me and learn what JWTs are, what problems it solves, how you can use JWTs, and how to be safer when using JWTs on your applications.

Learn what JWTs are, what problems it solves, how you can use JWTs, and how to be safer when using JWTs on your applications

JSON Web tokens dominated the way we give access to APIs and how we carry data from users, but to use JWTs safely we need to understand how they came to life and how JWTs can be useful.
In this talk we will take a closer look at the famous three-part structure that forms a JSON Web Token, and the claims each JWT can carry.
But knowing it’s history and structure is not enough, we need also to understand the algorithms used in creating a token and how you can use JWTs as access tokens or as ID tokens.
After understanding JWTs on a deeper level, we will create and validate a JWT together using the PyJWT library and discuss things you should avoid doing to be safer when using JWTs in your projects.

1. How did JWT come to life? Talk about the JOSE specification;
2. What actually is a JSON Web Token and its structure: header, payload, and signature;
3. What is a claim and its standardization efforts;
4. The different types of algorithms that can be used to create JWTs and what is JWKs;
5. Let's create a token together using PyJWT;
6. What is an access token and an ID token;
7. Things to avoid to be safer with JWTs

Jessica Temporal

Let's talk about JWT

Databases are beautiful beasts built with several layers of abstractions that we rarely have to know or even care about, now it's time to break them open and discuss the different components that maketh a database and how one would go about building one if they wanted. This would be based on my learnings when I went about building my own Toy Database in Python

Lets break them the different components that maketh a database and how one would go about building one in Python !

This talk will help unlock the internal workings of a Database by breaking down the abstractions that make it. We will use Python as our weapon of choice to slowly discuss how you would 
go about building the different components of a database.

1) Talking to your Database: We start by building out an interface and a language that helps us communicate with our database. We will use the Prompt toolkit to build a REPL & use a simple SQL-based language with basic regular expressions that can parse it to instruction to execute.

2) Working with Data: Now that we can communicate with our database using instructions. we start the actual work in building out the Datastore, We initially store all the data in a simple in-memory dictionary and then move to persist this data to disk. We now read the data from the disk to memory every time we query the data and write back the data to the disk but this makes things very slow :(
This problem is our entry into the beautiful world of Indexes so by building a very basic Btree index to store references in memory to quickly access only what we require from the data on disk we can actually speed up our access times for basic row access queries from O(N) to O(1) where N is the number of rows in a table

3) Future: We can now proudly demo our new and polished database that can store data, persist it, and can run queries that are quite fast thanks to our Btree Indexes. We also discuss how this Database can be improved in the future by supporting full ACID Transactions, allowing concurrency, and handling locks

Our Database journey now comes to an End and As they say, The best way to understand something is to build it yourself :)

Sangarshanan

Build-A-Database with Python

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 categorization

Is news media polarization based in fact or is it conditioned due to the consumption of alternative news media?

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.

Correlating messy data with "correlate"

Abstract

Description

The speaker