One of my favorite parts about working in data science is being a resource for others in the field. I enjoy connecting with students and those early in their career to share what I have learned from over eight years in data science. These conversations include topics such as whether to pursue a graduate degree, the best approach to tackle a project, promoting your accomplishments in a professional manner and more. I originally presented the contents below this spring at PyCon Lithuania reviewing the top five lessons from my career thus far. If you are interested in the full recording you can check it out here.
1. Connect Everything You Do to Business or Career Priorities
As a data scientist there are many exciting and interesting models to use. It can be easy to get caught up in the latest and greatest technology. However, it is imperative to make sure that you are making visible contributions to high-impact projects that are aligned with your company's goals and driving impact. Make sure that you can summarize the main goals and priorities your company has set for the future. Position yourself for visible opportunities by raising your hand for the projects that really contribute to these priorities. You want to be associated with such projects when it is time for the company to evalute the performance of individuals and teams.
It is also important to do work that moves you towards your own personal career goals. You want to make sure that you are working on something that helps develop the skills and experience to drive your career forward. For example, if I were an individual contributor who wanted to progress into a management role I would join a project where I collaborated with junior colleagues and provided guidance for them throughout the project. That experience demonstrates you have management skills and are taking the initiative to develop further.
2. Communication is an Underrated Skill
I have had to work with non-technical stakeholders on almost every data science project in my career. Some of these teams had experience working with data science previously or a moderate degree of data literacy, but many of them had minimal statistical knowledge. At the end of the day when working in industry a data scientist needs to be able to collaborate with these non-technical teams. This is a skill I learned through internships and early in my career as it wasn't taught in my undergraduate or graduate programs.
In my experience teams trust results and appreciate analysis more when they can understand it at a high level. This helps them understand the recommendations you’re making and how data science projects work. This leads to a stronger and more effective collaboration over time. It is especially important to help stakeholders understand how their expertise fits into the process. If you would like to learn more about helping stakeholders understand data science check out my talk from the NYR conference or slides on tailoring deliverables for stakeholders.
3. The Most Advanced Solution Isn’t Always the Best One
Data science is a field that has always changed rapidly, and its development seems to only have accelerated with the rising popularity of AI models. Companies and teams advertise state-of-the-art models to demonstrate they use cutting edge technology. While there are many use cases for advanced modeling many data science projects might not require them.
Starting a project with a simple model or algorithm establishes a baseline to test against more advanced techniques. Suppose you needed to predict if a customer will convert due to a new email marketing campaign. A logistic regression model is a simple method to build a classifier that can predict the probability of a customer converting. A more complicated neural network also has the ability to make that prediction, but would require more time to build and higher computational costs. You need to consider if the juice is worth the squeeze - does the more complicated model provide enough of a lift in performance to justify its increased cost? I recommend a crawl, walk, run mindset for data science projects. Identify the simplest analysis to deliver what the stakeholder wants for a quicker win, and build upon that as you try more advanced techniques.
4. You are Your Own Best Advocate
When I first started working in industry I believed that if I did my job well and executed on projects I would get a gold star and climb the corporate ladder. I quickly discovered that this was not the case and I had to learn how to promote my contributions and accomplishments. I have worked for a variety of managers and companies of different sizes. Even when I worked for a team lead skilled at personnel development I needed to advocate for myself by sharing goals and accomplishments.
It sounds cliche but sit down and think about where you want to be in the next 1-5 years. Identify what you like to do at work, tasks that excite you and areas you want to explore. Once you define those goals you can then discuss with your senior leadership how to work towards them. Suppose you want to be promoted to senior data scientist in the next two years - make sure you know the expectations and responsibilities of that role. Identify opportunities to demonstrate those skills and track them to bring up during your review. Make sure you know how your company considers promotions - do they happen at a certain time of year, is it possible to get promoted on your current team or do you need to move internally? Do not just sit back and wait for opportunity to come to you - keep track of your contributions and make sure those above you know about them.
5. Dedicate the Time to Build a Network
A strong network can provide resources beyond a job referral - long lasting relationships in your field, learning from industry experts and more. You will do yourself a big favor if you take the time to build a network over time and use it beyond when you need a favor. My favorite networking events combine both online and in-person resources that are a mix of presentations and social events. I am able to learn and also get to know people in my field.
Earlier on in my career I found it challenging to figure out how to have networking conversations. When I am reaching out to someone to learn more about their career and experiences some of my favorite questions to ask include:
- What do you wish you knew when you were at my stage in your career that you know now?
- How have your career goals changed over time and why did they change?
- Who else should I reach out to that could be a good resource?
You can also use resources such as alumni directories and LinkedIn to find people you want to connect with. It can be intimidating to reach out to people but you can’t start building a network without those first messages.
Conclusion
These are the top five lessons I produced when reflecting on my time as a data scientist and use their teachings to this day. It is important to be continuously learning and developing, not only in regards to data science but in many careers. I am looking forward to seeing what the next eight years bring.