When do we do a project? When not?
It is often difficult to say whether we do a project with an organisation. In this guide, we want to present the criteria that should guide that decision making process.
You can use this page to learn more about the criteria that we have used mostly implicitly for the past years when deciding on whether we want to do a project with a potential partner organisation. We have tried to put into words the informal thought and decision processes that have evolved over the past few years. If something is unclear, please let the ethics committee (#ask-the-ethics-committee on Slack) know.
Relevant documents:
Template for Ideation Phase (Questions that help guide talks with NPOs)
Relevant other pages:
How to use this guide
Go through the criteria and answer the questions: they'll either be recognisable because they're followed by one or more of the emojis ✅, ❓, ❌or because they are printed in bold.
Note down whenever you have to answer with a ❓
If you have❓, the project idea is apparently not a clear case. In this case, you should involve the ethics committee that will review the case.
Overview of criteria
| Category | Criterion | Main Question |
|---|---|---|
Partner | legal form | is the legal form one we collaborate with? |
Partner | financial situation | does the partner have money to pay someone for this? |
Partner | purpose | is the partner aligned with CorrelAids values? |
Project | educational factor | can our volunteers learn something in the project? |
Project | for-good factor and ethical considerations | will this project have a positive impact for the partner and society as whole? will it cause no harm? |
Project | sustainability | is the project sustainable for the partner? |
Project | role of project in partner org | do we develop someone else's core product/project? |
1. Characteristics of the partner
1.1. Legal form
What is the legal form of the potential partner?
Non-profit organisation / e.V. ✅
Typically, we collaborate with Non-Profit organisations (NPO). In Germany, they are typically registered as a "Verein" (e.V.) and have a certain charitable purpose ("Gemeinnützigkeit"). Collaborating with NPOs is typically no problem - but still check all the other criteria!
Foundations❓
Foundations ("Stiftungen") are also a potential partner that you might encounter. Depending on their size and supposed financial resources (see 1.2.), you should double-check with the core team.
Social entrepreneurs❓
Social entrepreneurship has been on the rise in the last few years, resulting in quite a lot of "charitable companies" (gGmbH, etc.). If your potential partner organisation is one of those, please check back with the core team call.
Individuals❓
From time to time, we also encounter individuals who are pursuing a for-good project on their own, most often in addition to their job and open-source. One example is the MyAiGuide project.
Non-official initiatives ❓
Especially in political contexts, we might find initiatives that are not (yet) organised in a legal form and are rather loose.
Political parties ❓
CorrelAid is a non-partisan network. That means that typically we do not collaborate with political parties and/or organisations affiliated to political parties (e.g. their youth organisations). However, feel free to come to the core team call to discuss your idea and/or use the Slack community to find like-minded folks to pursue your idea outside of the official CorrelAid context.
Academic institutions / individuals❓
Sometimes, university professors, PhD students or postdocs might approach you with a project to collaborate on. In those cases, it comes down to the content of the project (is it the main project of the person? does it have a for-good impact beyond research?) and other circumstances (cf. 1.2. could they pay someone?).
For-profit companies❌
Unless in very specific circumstances, we won't do projects for for-profit companies. If you think your project idea is one of the exceptions, please join the core team call and pitch it :)
1.2. Their financial situation
Typically, CorrelAid Data4Good projects do not come at a cost for non-profits (although they can donate to us) except for the costs of in-person kickoff workshops (as of April 2021, this has obviously not been "a thing" for a while). This raises the question whether we should do projects with organisations who have large budgets or are for-profit in nature.
Some rules of thumb / questions to ask:
How many employees / how much budget do they have?
NPOs (e.V.) and foundations up to ~10-15 full-time employees: ✅
other circumstances: please check with the core team ❓
could the organisation (realistically) pay someone for the work? e.g.
a working student / student assistant?
an external consultant?
an employee
If this is clearly the case and you get the impression that they are just looking for free labor, we should not do the project: we do not want to replace paid work! ❌
Another way is to ask: "would/could this project happen without CorrelAid?" --> most NPOs will deny this so it's important to stay critical! If in doubt, ask the core team call!
Getting a feeling for this aspect can be a quite tricky and there have been quite a few discussions concerning this. So don't worry if you feel unsure about it - join the core team call and we'll figure it out together!
1.3. Purpose
All potential partner organisations / individuals should be aligned with our values. Due to the nature of our network and the connections we make, this is typically not a problem, but please do some research to find out what they do and who they are - for example, check out their website and social media accounts. If you have any doubts, please ask the core team.
Main question: Is the organisation aligned with our values?
2. Characteristics of the project
What's a "good" CorrelAid project? When shouldn't we do a project?
A good CorrelAid project is a win-win situation: our volunteers can apply and expand their knowledge / skills and the NPO gets help with their data challenge. If you find out during the ideation process that the project would be extremely lop-sided in either direction, you can (and probably should) politely say no. Examples would be:
a project which would only be "cool" for the volunteers, but there would be no real positive gain for the NPO. Every CorrelAid project should be of use to the partner organization. If you want to work with an exciting technology just for the sake of learning about it, you can consider doing an internal project (a project without a partner NPO).
a project where CorrelAid volunteers would be used for free labor which is not rewarding in any way, e.g. a project consisting only of mundane data entry tasks. Every CorrelAid project should entail opportunities to apply and learn useful data science skills.
2.1. Educational factor
As stated above:
A good CorrelAid project is a win-win situation: our volunteers can apply and expand their knowledge / skills and the NPO gets help with their data challenge.
The main question is: could our volunteers learn something in the project? This is most often the case because CorrelAiders come in all shapes and sizes and skill levels. So even if you might find a project with a lot of data cleaning and a bit of data visualization boring, it could still be a very cool educational experience for less experienced data scientists.
That being said, a project should not only consist of mundane tasks like data entry, coding data (e.g. coding tweets into categories), data organisation in spreadsheets, etc.
2.2. For-good factor and ethical considerations
If you have any questions / doubts about the for-good factor or ethical considerations of the potential project (you answered at least once with a ❓to one of the questions below) you should contact the ethics committee.
Does the project have a for-good factor? Will it positively impact and improve the organisation's work? Most often, this will be the case with NPOs - most NPOs just don't do data for the sake of it but come with a clear goal in mind.
Internal projects also should have a for-good factor, i.e. beyond learning about technologies, the project should still have some sort of for-good factor. For example:
open source development
deriving interesting insights from open data to answer social questions
create data visualizations to draw attention to social issues
...
Also very important to ask is: are there any ethical considerations? In particular, you should always question whether the project could cause harm in any way or infringe on the rights of the data subjects. For example:
will the project reinforce existing inequalities, e.g. racial or gender inequalities?
yes. -> the project is not in line with our values, hence we don't do the project ❌
not sure -> ❓
no. -> ✅
Is a technical, data-driven solution in the best interest of the affected people?
no. -> the project is not in line with our values, hence we don't do the project ❌
not sure -> ❓
yes. -> ✅
will the technical solution make moral / ethical decisions?
yes / not sure -> ❓
no. -> ✅
will the project entail data from voice recordings, videos or photos of individuals?
yes -> ❓
no -> ✅
GDPR related questions:
does the organisation have the permission of the data subjects to use and analyze the data? (GDPR)
no -> organisations need the permission of the data subjects ❌
not sure -> ❓
yes -> ✅
is the data content sensitive? The GDPR defines data which are particularly sensitive and are not to be analyzed except in very certain circumstances.
yes -> ❓
no -> ✅
Processing of personal data revealing racial or ethnic origin, political opinions, religious or philosophical beliefs, or trade union membership, and the processing of genetic data, biometric data for the purpose of uniquely identifying a natural person, data concerning health or data concerning a natural person's sex life or sexual orientation shall be prohibited.
(Source: General Data Protection Regulation, Official Journal of the European Union, accessed on April, 16th 2021 here)
You can also refer to the following presentation (in German):
2.3. Sustainability
Will the project be sustainable for the partner organisation? Will the project have an impact beyond the time the CorrelAid team is available? Topics like hosting, maintenance and handovers should definitely be discussed if applicable (for "simple", one-off projects like generating a one-time report or some visualizations this might not be as important).
Example: A NPO approaches you - they need a dashboard to monitor their internal data processes. A shiny dashboard immediately comes to mind. In your talks with them, you should evaluate whether it is possible for them to a) host it themselves, b) host it on shinyapps.io or c) whether it could be used only locally on the laptops of 1-2 people.
2.4. Scope
A central feature of CorrelAid projects is that they are time-boxed, require at most 5 hours of commitment per week and typically don't last more than 6 months. Hence, when scoping a project, this should be an important point: Is the project do-able with volunteers and in a limited amount of time? If the project idea under discussion seems quite "large", it can make sense to break it down into smaller subprojects or propose to do an initial project to explore the idea and then a follow-up project.
2.5. Role of the project for partner organisation
CorrelAid supports partner organisations with their data challenges. And of course, we like to work on challenges that have a real impact either for our partner's work (e.g. improve their internal processes, help them with internal reporting or monitoring) or for the general public (e.g. develop impactful data visualizations, dashboards, ...). It boils down to the question:
Will CorrelAid volunteers "develop the whole core of the partner's business model / project / product"?
An example: a social startup wants to develop a machine learning model to offer a paid service to cities. They ask CorrelAid for assistance. In this case, it would probably be ok to offer some informal exchange. However, CorrelAid volunteers should not develop the ML model - we do not do free labor!
This aspect can be relevant for social entrepreneur partners and academic institutions / individuals (e.g. we should not do the main part of someone's PhD data work), but also could be relevant for NPOs and foundations.
Last updated