I am a mathematician who has similarly been intentional about diversifying my research. My PhD training was in algebraic topology and category theory, very far on the "pure math" side of the spectrum. I can think of a few ways you can diversify your research, listed below. I'm going to ignore all the identity politics stuff, because I don't think it's relevant to the actual question of how to diversity your research and cast a wider net.
- Do some applied math/stats research and choose application areas that stretch you.
I did research related to the opioid epidemic in Ohio (resulting in one publication and one preprint so far), and it connected me to a whole new world of Harm Reduction. As part of my statistical consulting for the organization Harm Reduction Ohio, I met many social workers, sociologists, epidemiologists, journalists, people who use(d) drugs and want to help others beat their addiction, etc. Now I give talks in Ohio and help connect other researchers into this topic.
Later, I did research related to policing and protests and have published two papers from that so far. Through this, I got better at time series models and learned about Hawkes processes from applied math. I also learned about different models of policing, the psychology of protests, collective trauma, etc. I've also done research about Ukraine (one paper in 2015 and one we started in 2021), to understand breakaway movements and civil wars. I learned a lot of game theory, and befriended Ukrainian journalists and sociologists.
Later, I coauthored a paper about wound-healing and diabetes, which taught me a lot about biology, pharmacology, and the lived experience of people with diabetes.
If you are interested in this kind of thing, you might enjoy the research coming out of QSIDE.
- Teach courses that stretch you in the direction you want.
I developed and taught courses in statistical modeling, time series analysis, Bayesian stats, data mining, and big data. I also taught several computer science courses. In every course, I picked real-world datasets for my students to analyze. As a result, I learned about the world bank, racial disparities in median net worth in the USA, how climate change impacts vulnerable populations, racial disparities in criminal sentencing, and what factors predict for high salary, long life span, long-term happiness, etc. I remember learning about the John Henry effect from this. I'm working to develop a course that would teach statistics via social justice applications. It would be called "Statistics for Social Justice." There's already a book "Mathematics for Social Justice" and I look forward to reading it and teaching that material at some point.
- Study the scholarship of teaching and learning.
After observing some trends in test scores in my early years teaching, I read up and learned about "stereotype threat." I then changed how I taught and reduced the impact of stereotype threat. Around the same time, I read up on fixed vs growth mindsets and again changed how I teach. These changes led to several publications in the scholarship of teaching and learning, e.g., a paper in PRIMUS, another in the Annual Review of Statistics, and a couple in the Journal of Humanistic Mathematics.
- Intentionally read papers by authors from under-represented groups, that might be under-appreciated and under-cited.
I tend to believe that papers by mathematicians from developing countries tend to have a harder time getting published, achieve less visibility, and get less citations, than the same papers would get, if written by a mathematician in the USA. So, when I find such papers, I make sure to read them and to cite them if appropriate, because I feel many others in my field will have chosen not to read them.
Let me give you a concrete example. This past fall, I wrote a paper about coalgebras over comonads. As part of this, I wanted to do as full of a literature review as I could. I stumbled upon a paper by a mathematician from Burkina Faso. Despite being written around 2016, this paper had fewer than four citations, and the author had returned to Burkina Faso rather than getting a research professorship in Europe where he'd done his PhD. I might have said to myself "well, this paper is not highly cited, so it's probably not very important or everything in here was well-known." Instead, I read the paper fully and carefully, understood the main result, included a discussion of it in my paper (a bit of an advertisement to the reader that they should read his paper), and cited it.
- Travel to research centers in developing countries, and establish collaborations with researchers you meet during such travels.
I travel every chance I get. Often, I reach out to universities as I go, and they invite me to give talks. For example, I taught a week-long course about model categories to a bunch of Moroccan PhD students in 2015. I gave talks in Argentina and Uruguay in 2018. I gave talks in the Middle East in 2022. My CV has a list. I've not counted but feel confident I've done this in at least 20 countries. It's led to some wonderful collaborations and lots of math I would have never learned about on my own. If you are interested in this kind of thing, you might consider visiting the AIMS centers in four African countries.
- Go to seminars, colloquia, and conferences on topics you are interested in.
I go to the seminar in my main research area religiously, and every week I try to go to at least one other talk, whether in the colloquium, or a different seminar (e.g., topological data analysis applied to neuroscience data), or an interdisciplinary speaker series, of which we have several in central Ohio. Through these activities, I've found lots of new co-authors and expanded my research in lots of new directions. For example, that's how I met the professor who got me interested in the opioid epidemic. Some conferences that helped me a lot include a conference on big data at Wellesley in 2015 (where I met many statisticians and got my start in that direction), an applied math conference at ICERM in 2019 (which led to the policing and protests papers), and a QSIDE social justice math conference in 2021. There's a conference this summer on gerrymandering (specifically, the math you can use to detect unfair maps and to draw fair ones) that seems very interesting. Maybe you can attend.