Skip to content

Talks

All talks

2026 4 talks

  • Keynote: As Modas e o Futuro

    1 presentation, in 2026 pt

    Keynote on keeping up with technology and its fads. Starting from my own path, from my first computer to ChatGPT, I revisit hypes such as cloud computing, NoSQL, big data and microservices to discuss the problems of always following the latest trend and what to expect from a future with artificial intelligence.

  • Common Voice: o Português precisa da sua voz

    1 presentation, in 2026 pt Lightning talk

    Why speech recognition models such as Whisper still make many more mistakes in Portuguese than in English: there is not enough good, openly licensed data. I present Mozilla Common Voice, where I am the Community Language Expert for Portuguese, and how anyone can contribute their own voice.

  • Criando aplicações de linha de comando amáveis

    2 presentations, from 2025 to 2026 pt Tutorial

    Tutorial on building command-line programs that are a pleasure to use. It starts from the user's point of view (with exercises), goes through what makes a CLI lovable or hateful and ends with the argparse module and a summary of good practices. Requires knowing how to program in Python and how to use the terminal.

  • How electoral data from TSE (Brazil's electoral court) is used in investigations, such as Abraji's CruzaGrafos, and what to do when transparency gets worse: columns that get renamed, files that disappear and candidates' CPFs (taxpayer IDs) that stopped being published in 2024. I show how to cross data from previous elections and scrape what is missing with Python and Scrapy.

2025 1 talk

  • Presentation of mercados, a free software Python library and command-line tool that downloads, extracts and cleans Brazilian financial market data straight from official sources. A person's financial journey (inflation, savings accounts, CDI, Tesouro Direto, real estate funds) guides the talk, which ends with how to contribute to the project.

2024 5 talks

  • With Ana Paula Sales. We present YouTool, a Python library and command-line tool for collecting YouTube data, and journalistic investigations that used it, such as monitoring deleted videos and anti-disinformation projects that led to stories in Brazilian and international outlets.

  • Storing data in QRCodes

    1 presentation, in 2024 en Lightning talk

    Lightning talk showing that QR codes, usually used for short URLs and payments, can store up to about 3 kB of data, with Python examples using the qrcode library, including a contact card (vCard) generated from the command line.

  • PostgreSQL Beyond Django: Strategies to Get Max Performance

    2 presentations, from 2022 to 2024 en pt

    Techniques to get the most out of PostgreSQL in Django applications, with examples from Brasil.IO and the Brazilian federal company registry: importing and exporting large amounts of data, modeling, metrics and full-text search.

  • Workshop, in a Colab notebook, on the techniques and tools of the three investigations supported by ICFJ's Disarming Disinformation, with Instituto Serrapilheira, which resulted in nine stories: collecting YouTube videos and transcripts with youtool, transcribing audio with Whisper, summarizing transcripts with a large language model and full-text search in PostgreSQL. The examples use Abraji's channel, so as not to spread the disinformation videos.

  • Brasil.IO - Dados abertos mais acessíveis

    4 presentations, from 2019 to 2024 pt

    What open data is, what it is for and why so much of it is still hard to use, with examples of news stories built on it, tools for working with data and Brasil.IO, which gathers public datasets and makes them available in an accessible format.

2023 3 talks

  • CruzaGrafos

    1 presentation, in 2023 pt

    Presentation of CruzaGrafos, an Abraji tool born from a 2019 idea and prototype, supported by the Google News Initiative, which uses Brasil.IO data to browse the relationships between people, companies and candidates, search for paths between them and export the data. It also shows the new features and datasets.

  • Introduction to machine learning, natural language processing and large language models such as ChatGPT, with applications in journalism and the problems of using them.

  • ChatGPT e Proteção de Dados

    1 presentation, in 2023 pt

    What large language models such as ChatGPT are, applications in law, such as extracting entities from documents, and the possible problems of using them, including for data protection.

2022 1 talk

  • Common problems when crossing open datasets (which also happen with private data) and methods to reconcile them in PostgreSQL with URLid, with examples from Brasil.IO, such as crossing companies barred from public contracts with campaign donors.

2021 4 talks

  • Brasil.IO COVID-19: dados abertos sobre a pandemia

    16 presentations, from 2020 to 2021 pt es

    How Brasil.IO, with dozens of volunteers, collected and organized the bulletins of the state health departments to build an open dataset of COVID-19 cases and deaths by municipality, in the face of federal data without history, bulletins published as PDFs and images and methodologies changed without notice. It also covers disinformation and how to collaborate.

  • Python & PostgreSQL = ❤️

    1 presentation, in 2021 pt

    Python and PostgreSQL used together.

  • Participation in a MundoGEO Connect panel on how geographic data portals can help fight pandemics, talking about Brasil.IO's COVID-19 data.

  • TrateCov NÃO FOI HACKEADO: entenda.

    1 presentation, in 2021 pt

    Technical explanation of the TrateCov case, a Brazilian Ministry of Health website that, according to the former minister at the Senate's COVID inquiry, was "hacked": how websites work (backend and frontend), why anyone had access to the code that runs in the browser and what should really be investigated, such as the lack of transparency and the leak of personal data.

2020 5 talks

2019 13 talks

  • Python para jornalistas

    3 presentations, from 2016 to 2019 pt Tutorial

    Python for journalists.

  • Software Livre e Movimentos Colaborativos

    2 presentations, from 2017 to 2019 pt

    What free software is and why it is not the same as freeware or open source software, what copyleft is and how these ideas relate to other collaborative movements.

  • Ciência de dados no terminal

    1 presentation, in 2019 pt

    Talk at PythonBrasil[15] about the conference's talk selection process and about what I had been working on: the post-truth keynote, data science in the terminal, the PGConf talk and crau, a tool for archiving web pages.

  • Brasil.IO: libertando dados públicos

    7 presentations, in 2019 pt es

    What open data is, what Brazil's Access to Information Law says and how public data enables citizen oversight, with Python tools for working with data and a presentation of Brasil.IO, which frees public datasets.

  • Why transparency depends on good-quality data and how Brazilian public data falls short of it, with stories of badly published open data and what we do at Brasil.IO to solve part of these problems.

  • Keynote on the post-truth era: examples of distorted reality, from the anti-vaccine movement to deep fakes and algorithmic bias, and thoughts on what we can do with data journalism, open data and Python, with live examples using Brasil.IO.

  • Torra caseira de café

    1 presentation, in 2019 pt Tutorial

    Free workshop on roasting coffee at home.

  • Open data, its problems and tools for working with it, with live coding, and how Brasil.IO uses PostgreSQL, plus plans for improvements such as using graphs.

  • Ciência de dados na linha de comando

    3 presentations, from 2018 to 2019 pt Tutorial

    How to analyze data using only the command line: navigating and handling files, filtering and combining commands to deal with files too big for spreadsheets, with examples of Brazilian public data. Some editions were hands-on tutorials.

  • What big data is (and is not): not being able to open a file in Excel does not mean you have big data, and systems such as PostgreSQL handle gigabytes of data with ease, with open data examples.

  • Behind the scenes of data journalism.

  • Journalistic work and applications built on public data, such as those by Gênero e Número and Operação Serenata de Amor, and the relationship between transparency, citizen oversight, copyright and open data.

  • Brasil.IO: Dados Abertos para mais Democracia

    4 presentations, from 2018 to 2019 pt

    Examples of news stories and applications built on open data, such as Operação Serenata de Amor and its robot Rosie, the idea of citizen oversight and Brasil.IO, which frees public data so more people can use it.

2018 10 talks

  • Deu match: limpando dados com expressões regulares

    1 presentation, in 2018 pt Tutorial

    Tutorial on regular expressions for data cleaning: what they are, where to use them (and where not to) and how to represent patterns, testing them on regexr.com.

  • Open data, transparency and Brazil's Access to Information Law in the practice of data journalism, with examples of stories, the data funnel and live coding to solve common problems, such as converting tables from PDF to CSV.

  • Hackeando Dados Públicos

    1 presentation, in 2018 pt Interview

    Guest appearance on PodProgramar #42 about hacking public data.

  • Investigando bases de dados com SQL

    1 presentation, in 2018 pt Tutorial

    Workshop for journalists on investigating public databases with SQL.

  • Dados abertos, Open Source e muito mais

    1 presentation, in 2018 pt Interview

    Guest appearance on episode 10 of the Pizza de Dados podcast about open data, open source and other topics.

  • Brasil.IO: saiba mais sobre dados públicos

    1 presentation, in 2018 pt Interview

    Interview with Abraji about Brasil.IO and public data.

  • The story of Brasil.IO: why software alone is not enough to free data, usability as part of accessibility and the datasets available on the platform at the time, such as elections, company partners, judges' salaries and congressional expense allowances.

  • Roasting Coffee with Python

    1 presentation, in 2018 en Lightning talk

    Lightning talk about automation and coffee: from the fruit to the roasted bean, and how Python can collect sensor data, log everything and control the gas and the doors of a 12 kg roasting machine to reproduce any roast profile.

  • Live de Python #45: rows

    1 presentation, in 2018 pt Tutorial

    Guest appearance on Live de Python #45 about the rows library.

  • Capturando Dados Abertos com Python

    4 presentations, from 2017 to 2018 pt Tutorial

    What open data is and Brazil's Access to Information Law, with examples from Gênero e Número and techniques for collecting data with Python.

2017 7 talks

  • Introdução a Jornalismo de Dados com Python

    2 presentations, in 2017 pt Tutorial

    Introduction to programming for journalism using Python, in three parts: logic and basic concepts of the language, reading CSV datasets and computing statistics, and collecting data from the web.

  • How Python helped Gênero e Número investigate gender issues in Brazil with open data and data journalism techniques, with examples of stories, the open methodology and code snippets.

  • O Fantástico Mundo de Python

    2 presentations, from 2016 to 2017 pt

    Introduction to Python for people who do not program yet or do not know the language: its features, basic concepts with live coding, the ecosystem and the community.

  • Lógica de Programação em Python

    3 presentations, from 2016 to 2017 pt Tutorial

    Introduction to programming logic with Python for people who have never programmed: what programming is for, when spreadsheets are not enough and the basic concepts of the language.

  • Keynote

    1 presentation, in 2017 pt

    Keynote in which I tell the story of starting the PythonQuito community in Ecuador and explain what a software license is, what Brazilian copyright law says and what the freedoms of free software are.

  • Liberando Dados Com Um Comando

    4 presentations, from 2016 to 2017 pt

    Public data that is open but not accessible, and how to automate with free software the boring steps of getting, checking, cleaning and converting it, with the rows library and its command-line tool.

  • Planejamento financeiro

    2 presentations, in 2017 pt Interview

    Two-part conversation on BalCast #27 about financial planning.

2016 8 talks

2015 2 talks

  • Capturando Dados Com Python

    4 presentations, from 2014 to 2015 pt Tutorial

    The problems of accessing open data in Brazil and how to collect it with Python, with live coding using the rows library and the first steps of Brasil.IO.

  • Workshop on how to contribute to free software in Python through GitHub, using the Aleph project as an example: documentation and code style, Git and git flow, eXtreme Programming and TDD.

2014 1 talk

  • Dados Abertos e Meio Ambiente

    1 presentation, in 2014 pt

    The problems of accessing environmental data (water, air), how to integrate systems through APIs, reverse engineering and crawling, and the licenses that define what can be done with the data.

2013 6 talks

  • Python: Introdução

    8 presentations, from 2009 to 2013 pt Tutorial

    Introduction to the Python language: what it is, where it is used, the interpreter, basic syntax, flow control, data structures, functions, classes, modules and the standard library.

  • Robotics with Python and Arduino

    11 presentations, from 2010 to 2013 en pt Tutorial Lightning talk

    Introduction to Python and Arduino and how to integrate them.

  • Oficina de Arduino

    14 presentations, from 2010 to 2013 pt Tutorial

    Free hands-on Arduino workshop, building circuits and programming them.

  • PyPLN: Processamento de linguagem natural com Python

    4 presentations, from 2012 to 2013 pt

    Presentation of PyPLN, a natural language processing platform in Python developed at FGV: NLP problems, its architecture based on pypelinin and a demo.

  • Dicionários persistentes e escaláveis

    1 presentation, in 2013 pt

    Keynote about persistent and scalable dictionaries in Python.

  • Character encoding and Unicode in Python: why accented characters cause so many problems and how to handle them correctly.

2012 3 talks

2011 10 talks

2010 7 talks

2008 3 talks

  • Debian GNU/Linux no desktop

    1 presentation, in 2008 pt

    Free software on desktops, focusing on the Debian GNU/Linux distribution and its derivatives.

  • Software Livre e Telecomunicações

    1 presentation, in 2008 pt

    History of the free software movement, concepts, GNU/Linux and distributions, and free software applications in telecommunications: network servers, VoIP, wireless mesh networks and thin clients.

  • Minicurso de GNU/Linux

    1 presentation, in 2008 pt Tutorial

    GNU/Linux short course.

2007 1 talk

2005 1 talk

  • Software Livre

    1 presentation, in 2005 pt

    My first talk: what free software is, the GNU Project and Linux.