Beating the 4-Year Slump: Mid-Career Growth in Data Science

This article provides a list of resources for data scientists who are transitioning from early-career/entry-level positions to more established roles. Surveys have shown a sharp decrease in satisfaction starting around 4 years into the profession, and resources are less obvious and readily available for professionals who have a good handle on the basics of data science than they are for beginners.



There are a lot of resources for early career and aspiring data scientists today, and many bootcamps, books, and MOOCs cater to those hoping to enter the field. However, after a few years, one learns the industry, is a veteran of random forest and boosting algorithms, and can set up a solid A/B test design in a short period of time. It’s an awkward period in which one is no longer a novice but not yet an expert in the field, and for many data scientists, career development is not straight-forward. Recent polls show a satisfaction slump somewhere around 4 years on the job (https://www.kdnuggets.com/2018/04/poll-data-scientist-job-satisfaction.html).

Avg. Job Satisfaction vs Years on the Job.

However, there are many excellent resources for data scientists who are a few years into their careers, a few of which are described in detail below.Certain geographic regions may have more opportunities and events than other regions, but most metropolitan areas include at least a few opportunities a year.

Attending conferences can be a great way to learn new skills, network with other data scientists in the area, and explore new technologies/methods in data science and machine learning. Many academic conferences offer workshops focused on a particular method, as well as opportunities for publication. For those in the US, the American Statistical Association, Institute of Electrical and Electronics Engineers, and American Mathematical Society all provide conference lists for upcoming years, and many of these conferences welcome data scientists and industrial mathematicians, as well as academics. Industry conferences abound, and most US cities/international data hubs now host one or more data science/big data/machine learning conferences. Some conferences even allow for remote participation, providing access for data scientists in less-populated areas or working abroad.

Speaking at conferences can be a great learning experience that hones presentation and public speaking skills; many conferences make an effort to provide a few presentations by local data scientists or analytics managers.

My own experience speaking at Miami WiDS 2018 helped me focus on delivering high-level overviews without skipping the relevant mathematics and helped me familiarize myself with the process of speaking at conferences (meetings with conference organizers, practice sessions, online submissions…). The experience has helped me navigate the submission process for Miami Data Science SALON 2018 and create a system for external content review within my department.

Academic journals and conference publications are another good source of continued learning. Sites like arXiv, open-source journals like Journal of Machine Learning Research or Data Science (Methods, Infrastructure, and Applications), and material published by professional societies (such as Casualty Actuarial Society) often overview state-of-the-art methods covering a broad range of problems. Methods like XGBoost, persistent homology, or HodgeRank often appear in one of these publications before trickling into software packages and blogs. ArXiv and ResearchGate in particular have been a constant in my career since switching from academia to industry. Some employers even allow data scientists doing research to publish in academic journals or though professional societies.

Blogs, discussion forums, and LinkedIn data science groups can also aid in mid-career development and provide data scientists with meaningful connections within the field, particularly for those in rural areas or cities in which data science positions are scarce. The LinkedIn Data Scientists group is quite active (and has >50,000 members). Posts range from career guidance to technical papers to business topics/skills, and many articles/papers are shared every day in the group.KDnuggets (of course), Quora, Data Science Central, Kaggle, and GitHub all provide blogs and/or discussion topics ranging from early career topics to advanced algorithms/managerial topics. Many of these online communities also post local and online conference notices.

Finally, many metropolitan areas have local meetups for data professionals featuring local speakers or math/tech workshops. I’ve only recently been able to attend the meetups in my general area, but it’s been a good experience. I learned a lot putting together my first workshop, and one of the booster sessions provided some good insight on data wrangling from another industry that seems to work well in my industry, as well. It’s also a great way to stay in touch with other data scientists in the area who have varied expertise and span many industries.

While not as many resources exist for continued career development after working in data science for a few years as exist for new data scientists, many good options exist for mid-career development, both online and in-person. Conferences, professional societies/groups, publications, and local events offer a wide range of options for a variety of career trajectories. Happy learning!

Bio: Colleen M. Farrelly is a data scientist whose industry experience includes positions related to healthcare, education, biotech, marketing, and finance. Her areas of research include topology/topological data analysis, ensemble learning, nonparametric statistics, manifold learning, and explaining mathematics to lay audiences (https://www.quora.com/profile/Colleen-Farrelly-1). When she isn’t doing data science, she is a poet and author (https://www.amazon.com/Colleen-Farrelly/e/B07832WQG7).

Related: