Data Science for Social Good Berlin (DSSG Berlin) is an initiative of Data Scientists from Berlin, which supports non-profit organizations pro bono to use their data effectively and to gain insights suitable for projects. This is because they often lack the necessary expertise. Katharina Rasch, Data Scientist and part of the core team at DSSG Berlin, explains in detail what the preparatory work and know-how goes into the project and what concrete results can look like.
Hi Katharina. What exactly does DSSG do for social and charitable organisations?
We, DSSG Berlin, offer non-profit organizations (NPO) advice on data and data analysis. This means: We advise and support them in recognizing the value of their data and using it strategically to achieve their goals.
That sounds exciting – but for the inexperienced it may still be a little incomprehensible. Do you have an example?
A first typical example is the analysis of needs. So to find out where the target groups are that the non-profit organization wants to reach.
For example, we have worked with an organisation that places students in an internship where they can try out if the job they are aiming for is the right one. We carried out an analysis that highlighted the differences between the internships students are looking for and those offered by companies or government agencies. Afterwards, the organisation knew in which region there was a shortage of providers offering really suitable internships.
This opens up new application possibilities. What else can your work look like?
Actually, we very often have projects that involve making existing data applicable. For example, one organisation has been collecting very detailed questionnaires from its participants for years. In paper form. But for a very long time they lay unnoticed in boxes, with an incredible amount of knowledge and feedback on the work of the organisation. Here the project consisted of digitizing and analyzing all this data. There was too much full text in the questionnaires to read and analyse it all manually. Part of the analysis was therefore to extract relevant information automatically.
Could you carry out such analyses for any kind of NPO without restrictions? For example, could you do it as well for the local animal shelter?
The analyses we do are of course based on the fact that certain data is already available. This is what we look at first when we conduct an initial interview with an NPO. Often their employees simply need feedback on what kind of database they should use or what data should be collected at all, that can result in exciting analyses.
Especially with smaller organizations, it often happens that we have a conversation and then give them precise instructions on what they need to do first. After one or two years we talk again and carry out concrete analyses – once we have the data.
So your support is usually a longer process?
Exactly. We all volunteer alongside our jobs. So the hours we can devote to a project each week is limited. Therefore, we have initial meetings with some non-profit organizations and give advice on what the next steps should be, and only see each other again after a longer period of time.
With other organisations we also carry out lighthouse projects. After all, everything to do with data analysis is still very vague for people from outside the field. For the employees of NPOs it is therefore very informative to witness a complete analysis. To do this, we work closely with the organisations and identify a typical problem, such as the analysis of the questionnaires, to inspire and motivate them to think along with the data analysis from the very beginning and thus understand: What data do I need to collect so that I can later do the analysis that will actually bring me to my goal?
How exactly does an entire process like this work?
In a first phone call, we identify what the project is about. Then we start a call in our network of about 150 data analysts. For example: We have an organization that needs someone who knows about text analysis. Do you want to work on this project? We try to put two or three people on each project who work closely with the organisation. They then conduct a two to three hour workshop with the organisation, where they look at the project in detail: What are your goals? Where does data accumulate? How do you store it and what other data should you store? And, to a certain extent: What personal data should not be stored in the context of the GDPR? However, we cannot offer any legal advice on this point.
Basically, the workshop is about the definition of realistic questions which should and can be solved. Of course you can have a lot of wishes, but since we are working on a voluntary basis, you have to be realistic about the time needed for the project. After the workshop we will have a look: What would be a suitable lighthouse project? The whole thing culminates in our so-called Datathon.
What exactly is a Datathon?
It is like a hackathon – only for data. For us, it’s a weekend we hold once a year, with two to three organizations and about 30 data analysts to work on the issues. During such a weekend you can only do a certain amount of work, so it is very important to specify the questions beforehand.
And we want to make sure that something really comes out of it, like the analysis of the questionnaires, for example. Here we check in advance: How must the data be digitalized and cleaned, in which fields might there be unexpected information and where do we have to keep it anonymous? So that by the time of the Datathon there were both clear questions and data on which to work immediately.
How does the cooperation during the Datathon look like? Do the participants from the organizations just sit on the sidelines and watch?
No, all participants get together in small groups and work on the previously defined questions. The employees of the organization are there all the time. That is very important. They are right in the middle of it, and there is a wonderful exchange with the Data Scientists. They are also there to evaluate the data and its quality. This provides a lot of practical experience for the next project.
The data analysts also learn and see how their work helps the nonprofit organizations. Many are more technical in their work and rarely have the opportunity to provide strategic advice. Hopefully, this will enable our analysts to advise, not just implement.
You said that most organizations lack understanding of the importance of data. So how does cooperation come about?
It’s not that easy. In many cases it has to do with the fact that employees are very busy and chronically understaffed, so there is simply not the time to think about IT issues. Or that the issue of data tends to be rather deterrent for nonprofit organizations.
So we actually go out and try to reach the organisations. There are several conferences in Berlin where they meet. There we are present whenever possible and give lectures about what we do. Some of the NPOs then feel addressed in their pain points.
Do you also have cooperations with companies like Unbelievable Machine, with whom you implement joint actions?
Apart from volunteering, most of us are working in companies full-time. These support our projects by giving us the time to have a consultation during working hours. We also receive a relatively large amount of support for premises. We ourselves do not have any offices. For our meetings and sometimes workshops with NPOs that don’t have an office in Berlin, we simply rely on the support of companies like Unbelievable Machine. This includes sponsoring the catering at the Datathon.
Can you identify certain topics particularly concerning to many non-profit organizations?
The topics are not so different from those of the private sector. For example, we have marketing issues like: Which Google Adwords should we use to help people find us? Non-profit organizations often receive budget from Google for marketing purposes and can advertise there practically for free. Another typical topic is impact analysis. In other words, to make measurable what the non-profit organisation is actually trying to achieve by recording and comparing its work results before and after a project.
NPOs therefore benefit from the cooperation. What are the benefits of working at DSSG in a personal and professional context?
For me personally, the already mentioned strategic part. Professionally, I’m more involved in the technical implementation, and especially the initial discussions with organizations are very different and exciting for me compared to my normal work. When we are looking for data ambassadors for a project, we also try to bring in two more experienced people and one person who is just starting out in professional life so that they can learn from each other.
Also at the Datathons, we always try to formulate questions for each project that are more suitable for newcomers, and really very complex questions where the more experienced people contribute their knowledge, so that they can also learn technically from each other.
So your involvement with DSSG means that you approach issues in your work differently?
Absolutely. I do think more about the big picture of what I’m implementing and what we want to achieve with it.
What is your personal motivation?
My personal motivation is: As data analysts, we have a great deal of technical knowledge that we use for the company’s purposes – and I think it’s great that with the help of DSSG we enable ourselves and other data analysts in Berlin to use this knowledge for the common good.
What’s next for you? Are there any exciting new projects?
We are currently in initial talks with several non-profit organizations. But I don’t want to go into detail about that yet. It has turned out that a large part of these talks will not become a Datathon project, simply because the data is not yet available or there is not enough time at the NPOs. But I think that by fall we will have exciting projects for a Datathon again.
In any case, we are always looking for non-profit organizations. We are always happy to have initial talks and then see if it suits both sides and what the next steps could be.
Thanks for the interview, Katharina. And all the best with future projects!
This might interest you, too:
Data science for Social Good: Walking in a Data Wonderland
Data Science for Social Good: *um is supporting the 2nd Data Drive in Berlin
Data Science for Social Good: *um is supporting the initiative and the first event in Berlin
This post is also available in: German