What makes prospect development a great career?
Apra-IL is asking local and national industry leaders what the field means to them and why and how they have pursued success in prospect development. Through this blog series, we will explore what drives industry leaders to propel their careers and prospect development forward.
For this month's piece, Joan Ogwumike, Apra-IL member and volunteer, interviews Michael Pawlus, Data Scientist at the University of Southern California.
Michael Pawlus is currently a Data Scientist at University of Southern California. Prior to this he was Director of Prospect Development at The Trust for Public Land and before this Assistant Director, Prospect Research and Development at Grand Valley State University.
Michael is chair of the 2018 Data Analytics Symposium planning committee. He has volunteered with Apra in a number of other capacities throughout his time in this profession, most recently as one of the faculty members for the 2017 Apra OverDRIVE/ conference. He also served as programming chair on the Apra-Michigan Board and as a member of the first ARC planning committee. Michael has presented on his work at: Apra-International, Apra-Canada, overDRIVE/, DRIVE/, as well as at regional and online Apra conferences. In addition, Michael assisted with the creation of two online training courses, contributed a section to Prospect Research in Canada: An Essential Guide for Researchers and Fundraisers and was featured in an article in the Chronicle of Philanthropy.
Michael has a Master’s degree in Librarianship from the University of Sheffield and a Bachelor’s degree from Grand Valley State University.
He lives in Greater Los Angeles with his wife, two kids and an orange tabby cat from Korea.
Apra-IL: Describe your journey into your current position.
Pawlus: My journey all started when I learned about R. I wish I could remember how I first heard about R but I do remember going to the library and getting my first R book. I can recollect so clearly sitting on the #4 Bus traveling north up Eastern Ave in Grand Rapids, Michigan and trying to wrap my brain around the syntax and R basics like vectors and data frames.
Even before then I was just extremely fortunate to work at really amazing organizations with really supportive and encouraging supervisors which is luckily still the case. I had the freedom to learn and try out new ideas. This actually started with learning Crystal Reports and then SQL but when I finally started learning R then I felt I suddenly had super powers and I needed to find a place where I could best apply this new skill set which brings us up to the present day.
Apra-IL: What does your role as a Data Scientist mean to the fundraising field?
Pawlus: It is really encouraging that an increasing number of development offices are hiring Data Scientists. As with so many other sectors, all of us working in nonprofits have access to more and more data but this only has value if we can convert this into actionable information which can best be accomplished through the scientific process.
When I used to just do prospect research, the main value that I felt I was adding was verifying data. Knowing that data is accurate is, of course, important however there is a critical next step that was always missing; how much value do these accurate facts have? It always felt like we had these formulas and heuristics that could be applied once we knew some data point however it never felt like these were tested or provable at all. This is when predictive modeling started to become more popular and started to answer these questions.
There is so much more to investigate as Data Scientists though and it is such an exciting and fast-moving profession generally and everything happening in this more general space can be applied to fundraising. In addition to discovering the most valuable pieces of data that we have, we can start thinking about how to then use this data and push it to front-line fundraisers. Integrating external data, extracting value from free form text using natural language processing tools and building recommender systems are all ways that we can provide more information to our colleagues and by extension provide a higher level of service to our donors which means our organizations can provide even more to those we serve.
Apra-IL: Could you tell us one perception people have about professionals in Prospect Development? What's the truth?
Pawlus: I remember being asked why I would want to leave prospect research to start doing data analytics or data science. I was always a little confused by the question because I think I am still essentially accomplishing the same aims as I was when I was in prospect research except now I can I can complete tasks at scale and provide a wider assortment of services but I am still providing fundraisers with actionable insights.
In theory, analytics is one of the three main branches of prospect development along with prospect research and prospect management. However, in practice, it always feels like analytics is off on its own island. It has a separate email list and separate conferences. I am not saying that everyone needs to become a data scientist however I think we should all know a little about what we all do.
Since the World Cup is right around the corner, I feel it is safe to use a soccer metaphor. In soccer, there are ten outfield players and some specialize in defense and some specialize in scoring but those that play in defense can still score and forwards need to be able to get back and help defend at times. The best teams are well-rounded and versatile in this way and I feel this would also be true for prospect development teams.
Apra-IL: What is something interesting you learned within your role this month?
Pawlus: I just learned about graph databases which are a different way of storing data that allows you to travel from node to node along relationships to find objects that are several degrees removed. This is similar to relationship mapping but imagine applying that to everything. You can also use these graphs to do something like this: let's say there are a group of foundations that give to your organization and these foundations also give to other organizations of course but then let's say there are foundations that give to those same common organizations that you discovered but they don't give to you; well these foundations might be good prospects for your organization and using graphs you can find these.
Also, I can't say I completely understand this yet but I have been learning more about this model called GloVe (Global Vectors for Word Representation) which essentially creates this matrix that quantifies the similarity between terms which is really helpful when you are topic modeling on survey data or contact reports. Using the example from the authors of the model, GloVe shows that Litoria and Leptodactylidae are related to frog and it can do this with other word and term groupings which wasn't really possible before. For those working in health, it is possible to find terms that all reference a similar medical condition without having expert subject matter knowledge using pre-trained word embeddings that have been done on the entire Wikipedia corpus for example.
I am sure I will look back at this in a month or two and realize I could have explained both of these concepts much better but then I will be on to learning new concepts and techniques. That is the way that it is right now there is just so much out there to learn and tons being developed constantly. It’s a great time to be a data scientist!
