2021 - Remote

November 12

Keynote: Data and Impact Measurement

Speaker: Johannes Müller (Chairman of CorrelAid e.V.)
In the social sector we often ask (or get asked) “Does this program work and does it create impact?”. To answer it, we need to answer a different question first: “What if my program or intervention wouldn’t exist?”. In this talk we will explore what impact evaluation is and how data can help to answer the question “what if?”. Johannes shares his experience in evidence-based impact evaluation and puts it into the context of the non-profit sector.
  • Slides

Project Showcase

In the project showcase, volunteers from our Data4Good project teams show how they have been supporting non-profit organizations with their data challenges. Topics include: report automation, creating an interactive report based on survey data, analysis of community dynamics, predictive modelling and using image data to improve cargo bike routing. The projects are:
  • Hacklab Developer Census: Hacklab Ghana Developer Census 2020 is the first and most comprehensive survey of people who code in Ghana. We developed an interactive HTML Rmarkdown report that showcases the results covering everything from developers’ favorite technologies and languages to their job preferences.
  • CargoRocket: Classification of Road Quality Using Crowd-Sourced Images: A cargo bike has its own unique requirements in terms of road conditions, which must be considered when planning a route. Therefore, CorrelAid volunteers developed a Python program that uses machine learning to classify road surface and pavement quality based on crowd-sourced street images.
  • Citizen Science Dynamics: We’ve analyzed data of the Citizen Science Platform iNaturalist, where users can upload observations of plants and animals, and other users can help identify them. We found out a lot of cool things on how the user base changed over time, what types of users there are, and much more - we’re writing a paper about it! I’ll also share some lessons learned. Spoiler: Exploratory data analysis can eat a lot of time.
  • Citizens for Europe Report Automation: Together with Citizens for Europe, we developed a workflow that allows for flexible generation of reports on discrimination and diversity within organizations. We will present how to go one step further than parameterized reports, building R-Markdown templates in a modular, more adaptable way.
  • Science on Stage: As many other NGOs, Science on Stage collects anonymized data from different sources (online-seminars, website, downloads) to measure the effectiveness of their work. We helped Science on Stage, an NGO supporting STEM-teachers from different backgrounds, to make sense of their data sources by automizing the data collection and cleaning process. In a second step, we visualized the results to create valuable reports in order to facilitate inhouse decision-making as well as keeping the Science on Stage shareholders up-to-date.
  • The impact of Voting Advice Applications on local politics: The CorrelAid team supported “Team Tomorrow” with the analysis of data from the “KOMUNAT”, an entirely candidate-based voting advice application (most famously known in Germany: the “Wahl-o-mat”) for the local town elections in Stuttgart in May 2019, developed to address the traditionally low voter turnout amongst the young first-time-voters. Overall, two major topics had to be investigated: the user statistics for product improvement and the political analysis to assess both the political preferences of the app users and the potential impact of the VAA on local politics.
  • Mare Liberum: Data visualizations to bring attention to illegal pushbacks: Push-backs are a set of state measures by which refugees and migrants are forced back over a border without considering individual circumstances and allowing them to apply for asylum. Mare Liberum gathers data on the illegal practice at the Mediterranean sea and publishes documentation on observed occurrences. Supporting their mission through data, CorrelAid has helped to conduct data analysis and provide visualisations for their annual report and consulted them on data collection and storage.
  • Addressing deforestation in Kosovo: Correlaid Netherlands, the Institute of GIS and Geospatial Technology of Kosovo (IGGTK), and Sustainability Leadership Kosova (SLK) worked on a joint project to address the deforestation issue in Kosovo. The main goal of the web platform (https://www.slkosova.org/mosemerrmalin-welcome)? Displaying maps of the deforestation events that occurred in the last 20 years to educate and raise awareness of this massive issue affecting Kosovo. #MoseMerrMalin
  • Gmoji: G-Moji is developing a mental health support app for adolescents and wanted CorrelAid’s help in assessing the feasibility of mood prediction from its pilot phone sensor and survey data.
  • Kulturen im Kiez: Description will be added soon.
Speakers: François Delavy (CorrelAid volunteer and data analyst), Sofia Thai (CorrelAid volunteer and data analyst), Alina Dallmann (CorrelAid volunteer), Frank Schlosser (CorrelAid volunteer, PhD Student at Humboldt University), Christine Hedde-von Westernhagen (CorrelAid volunteer), Simon Schölzel (CorrelAid volunteer), Tim Fangmeyer (CorrelAid volunteer), Andreas Neumann (CorrelAid volunteer), Aylin Kallmeyer (CorrelAid volunteer), Lukas Gröninger (CorrelAid volunteer), Federico Franciamore (CorrelAidX NL volunteer, GIS & Remote Sensing Specialist and Designer at Space4Good), Wen Qiang Toh (aka Wally, CorrelAidX NL volunteer), Daniela Mondorf (CorrelAid volunteer)

Topic Sessions

Introduction to Information Security for Non-Profits

Speaker: Ilja Sperling (Freelance Dev & former NGO Tech Lead)
Ransomware, leaks, data loss, and what to do about it (in COVID times): In this talk, Ilja will introduce us to the pillars of Information Security, provide an overview of organisational and individual threats, and outline mitigation strategies - all with a focus on NGOs. The talk will include an actionable plan for your team to get started.

Data Science Education

Speakers: Nina Hauser (Head of Data Literacy at CorrelAid e.V.), Vik Paruchuri (Founder at Dataquest)
Data science is a broad field. Together, we will explore which skill paths a data enthusiast for good may choose and what is needed to excel both professionally and ethically. We will then look at how we can translate those skills to foster data literacy in the civic sector.

Daten im sozialen Sektor: Chancen und Hürden

Speakers: Katarina Peranić (Deutsche Stiftung für Engagement und Ehrenamt), Georg Förster (AWO Bundesverband), Kwesi Aikins (Citizens for Europe), Frie Preu (CorrelAid e.V.)
Während sich in der Wirtschaft in den letzten Jahren ein Verständnis für digitale Technologien entwickelt hat, werden Gesellschaft und zivilgesellschaftliche Organisationen hier bislang größtenteils abgehängt. Deutlich wurde das im Digital-Report 2020: 72 Prozent der 3,400 befragten Organisationen haben eine Organisationsstrategie, aber nur 38 Prozent erheben und/oder nutzen (externe) Daten zur Strategieentwicklung. Eine vollständig durchdachte Wirkungslogik haben nur 30 Prozent der Organisationen und nur insgesamt 28 Prozent messen tatsächlich ihre soziale Wirkung. Es fehlen finanzielle und personelle Ressourcen, um die neuen Möglichkeiten durch Daten und Datenanalyse erfolgreich nutzen zu können. Mitarbeiter:innen schätzen ihre Fähigkeiten und Kenntnisse im Umgang mit Daten als zu gering ein. Dabei können Daten und ihre Analyse ein wichtiger Hebel für gesellschaftliche Wirkung sein. Moderator Johannes Müller diskutiert mit Katarina Peranić Deutschen Stiftung für Engagement und Ehrenamt, Joshua Kwesi Aikins (Citizens for Europe), Georg Förster (AWO Deutschland) und Frie Preu (CorrelAid e.V.) über die Chancen und Hürden der Datennutzung im sozialen Sektor: Was können Daten, Data Science und “KI” für zivile Organisationen leisten, was nicht? Was hindert große und kleine Organisationen daran, Daten zur Verbesserung ihrer Arbeit zu nutzen? Und wie können Akteur:innen der Zivilgesellschaft zu mündigen Datennutzer- und analyst:innen empowered werden?

November 13

CorrelAid in 2021

Speaker: Frie Preu (CorrelAid e.V.)
Frie Preu, Head of Operations, will walk you through the CorrelAid year 2021 with all its milestones, achievements and challenges.

Data Science, Climate Change and the Environment

Tackling Climate Change with AI

Speakers: Marcus Voss (Climate Change AI, Intelligence Architect and AI Expert at Birds on Mars)
Artificial Intelligence (AI) and Machine Learning (ML) provide powerful tools to tackle climate change in various applications: They can support climate change mitigation, for instance, by helping reduce greenhouse gas emissions within various applications. They can help to adapt to a changing climate and even advance climate science itself. However, AI and ML are not silver bullets and can always only be one part of the solution. This talk provides an overview of the strengths and weaknesses of ML, some example applications and recurring themes.

Put your money where your heart is - climate investing for your personal finance

Speaker: Constanze Bayer (2° Investing Initiative (2DII) and MeinFairMögen/ MyFairMoney)
Earning returns while saving the planet. Sounds fascinating, but has many pitfalls. Where do I get reliable information about the climate impact of my investment? How can I make sure that I don’t invest in fossil fuels? And (how) can I even have an impact? Many retail investors want to invest their money according to their values. But information is missing. 2DII has developed the platform MeinFairMögen/ MyFairMoney to help bring a bit of light in the darkness.
  • Google Slides (slide by slide view)
  • Google Slides (full HTML view)

Using AI to help restore the natural world

Speaker: David Dao (GainForest)
Land use and land use change play a critical role in our climate taking up about a quarter of annual anthropogenic emissions of greenhouse gases (GHGs) during 2007-2016. In addition to being a key driver of global warming, careless land use is also destroying valuable ecosystem services and is threatening the livelihood for local populations and a multitude of species. Major conservation and restoration efforts are underway to mitigate and safeguard against these losses, and to highlight the urgency of the issue, 2021-2030 has been declared the “UN Decade on Ecosystem Restoration”. However, we cannot preserve what we cannot measure. Machine learning (ML) plays a significant role in responding to this critical call for action and can accelerate the conservation and sustainable use of our natural world.
  • Slides

Sustainability of AI

Speaker: Friederike Rohde (Wissenschaftliche Mitarbeiterin am Institut für ökologische Wirtschaftsforschung (IÖW))
This talk is about the impacts of AI-systems from a sustainability perspective. Based on current work within the project SustAIn we present first ideas for a comprehensive sustainability assessment of AI-Systemes. With this effort we want to raise discussions amongst industry, developer communities and society about how to reduce social, economic and environmental risks throughout the whole lifecycle of AI-systems and call for action to make AI-systems more sustainable.
  • Slides

Topic Sessions

Pathways to Data Science

Speakers: Andrew Sutjahjo (Data Scientist at datavaluepeople, CorrelAid Netherlands co-founder), Allison Koh (PhD Researcher at the Centre for International Security, Hertie School), Sharon Gieske (Data Scientist at Picnic)
Physics, computer science or political sciene - all these (and many more) backgrounds can be found among the volunteers of our network. As diverse as our network, however, is the field of data science that connects us all. In this talk, three of our volunteers will tell us about their personal journey into data science. Among other things, this should help prospective data scientists to find their way in an exciting and often bizarre job market.
  • Slides

Agile2 - Juggling Innovation and Bureaucracy in IT Project Management

Speaker: Lisa Hehnke (Innovation Project Lead at Tafel Deutschland e.V. | Freelance Consultant & Data Scientist)
Imagine you created a compelling business case, recruited brilliant team members for your cross-functional project team, and even received generous government funding to build your shiny new software product. Being the modern day project manager you are, you plan to adopt agile methodologies by appointing a product owner for your self-organizing team, creating an emergent product backlog, and writing epic user stories. Before you can turn your ideas into reality, though, you have to convince your most influential non-IT stakeholders – and they are not amused. Who is responsible for controlling time, scope, and cost? Where is this important document called „risk register“ that they were told about? And why the unicorn are team members allowed to create their own tasks? That is not how project management works! Now it is up to you, the project manager, to come up with a solution to this dilemma.
  • Google Slides (slide by slide view)
  • Google Slides (full HTML view)


Speakers: Josef Holnburger ist Political Data Scientist und Co-Geschäftsführer am CeMAS - Center für Monitoring, Analyse and Strategie.
Spätestens seit der Pandemie sind Verschwörungstheorien besonders stark wahrnehmbar. Nachdem die Berichterstattung und auch die Wissenschaft Verschwörungstheorien eher mit Belustigung wahrgenommen haben, vollzieht sich derzeit ein Bewusstseinswandel und viele erkennen die Probleme, die mit der Verbreitung von Verschwörungsideologien verbunden sind: mangelnde Einhaltung von Pandemiebekämpfungsmaßnahmen, geringe Impfbereitschaft, aber auch Menschenfeindlichkeit und Gewaltbereitschaft. CeMAS will mehr Licht ins Dunkel von Verschwörungsideologien, Antisemitismus und Desinformation bringen und gleichzeitig Handlungsstrategien entwickeln. Wie gehen wir vor, was sind die größten Probleme und mit welchen technischen Mitteln nähern wir uns diesem Themenkomplex? Um all das geht es im Talk.
  • Slides

Beginner Sessions

Beginner session R

Speaker: Ann-Kristin Vester (Data Scientist)
In this session you will learn first steps with R and RStudio, like reading data from a csv file, data wrangling and creating simple crosstables and plots. We will mostly use base R and packages from the tidyverse, which are a great starting point for beginners.
  • Slides

Beginner session Python

Speakers: Tilman Kerl (CorrelAid volunteer, MSc student)
We take a look at some simple use cases and how we can tackle them using some prominent libraries. We start with some data cleaning using pandas and continue with forecasting using the prophet library and sentiment analysis.
  • Slides

Advanced Sessions: Working with Geospatial Data

Taking the shortest path - analysis of bicycle infrastructure data in cities

Speaker: Fabian Schaipp (TU München)
This talk will be all about biking in cities: you will see a brief introduction in tools for scraping and visualizing bicycle infrastructure data of cities. Based on a case study, we discuss a simple method for analyzing bicycle networks and explore how this could provide insights for mobility planners.

CargoRocket: How to use OpenStreetMap and Graphhopper to create a custom routing for Cargobikes

Speakers: Alexandra Kapp (cargorocket, CorrelAid)
Navigation apps for cars and bikes fail to properly serve the needs of cargobikes that are allowed to use bike tracks but, e.g., get stuck at too narrow barriers. With the open source project CargoRocket we aim to provide a routing specifically for cargobikes. This talk gives technical details how OpenStreetMap is used to create a “CargoBikeIndex” for every street and how this index is incoorporated into Graphhopper, an open source routing engine.
  • Slides

Introduction to analysing spatial data in R

Speakers: Jasmin Classen (Data Scientist and volunteer at CorrelAid), Aylin Kallmeyer, Andreas Neumann
In our (practically oriented) step-by-step introductory presentation, we will walk you through the process of creating maps in R ranging from reading in and preparing data to handling different shapes of spatial data (e.g. points, lines and polygons). For this we will be using a data set of bike sharing stations and trips in Hamburg.
  • Slides

Advanced Sessions: Modelling Corner

One Million Posts - Training a neural network to detect hate speech in German text

Speaker: Dominik Hohmann (Volunteer at CorrelAid)
Natural language processing project based on the one-million-posts dataset using the neural network architecture BERT. More than 3.000 user comments are written each day on www.derstandard.at. Moderators need to review these comments regarding several aspects like inappropriate language, discriminating content, off-topic comments, questions that need to be answered, and more. We have trained a dense layer on top of BERT to classify arbitrary German text regarding the question if it needs moderation.
  • Slides

Network models for analysis of NGOs

Speakers: Liu Bauer (researcher at CRI, co-leader of CorrelAid Paris, co-founder of City Interaction lab, Montreuil)
In this talk we will discuss different network models and how they can help us to model some processes in nature and especially from data. We will look into some examples as well as talk about packages for their implementation. We will also discuss particular cases of where network analysis could be useful for some NGOs.
  • Slides

mlr3: Object-oriented Machine Learning in R

Speaker: Jakob Richter (co-developer of mlr3)
The mlr3 package and ecosystem provide a generic, object-oriented, and extensible framework for classification, regression, survival analysis. In this talk you will learn how to conduct a benchmark experiment with mlr3, comparing different machine learning methods and preprocessing steps in one go.

Advanced Sessions: Data Visualization

Creating your own themes and colour scales for ggplot2

Speaker: Long Nguyen (PhD student at SOEP RegioHub/Bielefeld University)
When producing graphs for others to consume, you may want to have your own distinct visual identity across all graphics, using your own colour palette, your fonts, your logo instead of the built-in theme_minimal(), scale_colour_viridis() etc. Using the CorrelAid theme and colour scales in the {correltools} package as examples, this talk will show how you can easily create reusable functions to apply your own “brand” to your ggplot2 graphics in a time-saving manner.

A picture is worth a thousand words - Visual Data Exploration

Speaker: Nico Kreiling (scieneers)
Data is manifold, which makes it hard to discover its secrets and hidden relationships. Data Scientists can rack their brain on descriptive statistics or choose to explore the data using interactive visualisation tools. While both ways have benefits and dangerous pitfalls, the latter is definitely the more fun. This talk will help you discover your data visually by showing how you can quickly transform simple plots into powerful, interactive dashboards using Panel. Furthermore, it will provide you with a quick overview over Python’s most loved interactive plotting libraries and show how you can combine those in a single, interoperable dashboard.

Revealing Room for Improvement in Accessibility within a Social Media Data Visualization Community

Speakers: Silvia Canelón, Liz Hare
Screening of the talk “Revealing Room for Improvement in Accessibility within a Social Media Data Visualization Community” held by Silvia Canelón and Liz Hare at csvconf. Followed by a Q&A and discussion of tips outlined in this blog post. For a description of the talk, see here. License information: “Revealing Room for Improvement in Accessibility within a Social Media Data Visualization Community” by Silvia Canelón and Liz Hare, available on YouTube, used under CC-BY-SA 4.0

Advanced sessions: Tooling and Software Development

Git internals (or understandig git): The three basic git objects and their relation to git’s way of version control

Speaker: Henry Webel (Researcher at Novo Nordisk Foundation Center for Protein Research)
Git as a version control program can be confusing. This talk aims at helping you to understand git better by knowing some internals of the progam: Who does git version control is build around the three basic git objects? Afterwards you hopefully grasp what git trees are made of, that each node in these trees represent a snapshot of your working directory and how branches work. There will be a brief demo of the core concepts. Audience: You use git, but sometimes get confused why something did not work. Hopefully knowing some basics about the internals will help you with these struggles.

Introduction to Pipelines with Make

Speaker: Malte Kyhos
Managing large (research) projects with multiple data processing and computation steps can be hard. However, there is a wide range of tools that can help in such situations. In this talk, will give an introduction to pipeline toolkits, including a brief tutorial on the classic GNU Make utility.
  • Slides

Big things come in small packages – Tips and tricks for developing your own R package

Speaker: Cosima Meyer (Data Scientist)
R is a great resource for data management, statistics, analysis, and visualization — and it becomes better every day. This is to a large part because of the active community that continuously creates and builds extensions for the R world. During the talk, I will provide you with tips and tricks for contributing to this world by writing packages. If you want to get a quick introduction before the talk, here is a hands-on tutorial on how to write your first R package.

Team presentations of NLP-challenge results

Speaker: tba
Since late September several teams from our different local chapters have taken on this year’s CorrelAid-X Challenge. Here we would like to give them the stage to briefly present their results and tell us about the pitfalls and obstacles they faced along the way and what is more, overcame them as a team. We would like to thank Krystel from BuildUp for providing us with the data for this year’s challenge.
  • Slides

CorrelAidX Pecha Kucha

Speakers: Zoé Wolter, Marco Lax, Valentin Geier, Alexis Gillett, Liam Bailey (Joint coordinator of Berlin local chapter), Freya Moßig (CorrelAid Switzerland and Vice Chair at DataCross), Maike Metz-Peeters, tba, Daniel Manny, Pia Baronetzky, Sebastian Zezulka
Meet our local chapters through this engaging presentation format and see where you could find a chapter that might work for you.
  • Local Chapter Konstanz
  • Local Chapter Netherlands
  • Local Chapter Berlin
  • Local Chapter Switzerland
Pecha Cucha Switzerland.pptx
  • Local Chapter Cologne
  • Local Chapter Bremen feat. Hamburg
  • Local Chapter Munich
  • Local Chapter Stuttgart
Last modified 1mo ago