The Project Manager in Data Science

This data science project manager position is an important yet often misunderstood role.

So let’s untangle it and dive into:

• What is a data science project manager?
• 5 tips to be an effective project manager in data science
• The data science project manager salary
• Data science project manager certifications

What is a Data Science Project Manager?

A data science project manager is someone who manages a data science project. I admit that’s not a helpful definition, so let’s break that down…

What is Project Management?

Project management is an established field focused on the planning, delivery, control, and monitoring of a project. Per the PMI, a project is…

“a temporary endeavor undertaken to create a unique product, service or result”

Project Management Institute

Traditional project managers focus on delivering a project on time, on budget, and within the pre-defined scope. They attempt to pre-plan as much of the project upfront to avoid surprises during implementation.

Meanwhile, project management in agile teams focuses more on the creation of value using quick development cycles and continuous feedback. Often this work is divvied up among various titles that don’t explicitly hold “project manager” in the name.

Effective project managers can seamlessly shift between these two camps of thought – generally with a preference toward one camp of thought over the other.

What is Data Science?

Unlike project management, data science is an emerging field without a standard definition. But in general, data science is a broad field that leverages scientific principles to develop actionable insights and strategies from data. It is often depicted as sitting at the intersection of computer science/hacking skills, math/stats, and domain knowledge.

What does a Data Science Project Manager Do?

A project manager in data science is responsible for the delivery of one or more advanced analytics and AI/ML projects. Position descriptions typically read like other IT project management roles but with an emphasis on some specific applications of data science. Common responsibilities include:

• Develop and communicate project roadmaps for data science projects
• Coordinate and monitor day-to-day tasks and workflows of the project team
• Manage stakeholder requests and expectations; provide updates to project sponsors
• Scope and define tasks that fulfill the project vision; manage and document scope using a project management ticketing system such as Jira, Atlassian, or Rally
• Identify and gather data sets necessary for projects
• Proactively identify opportunities and provide recommendations to improve operational efficiencies and implement scalable solutions
• Remove impediments that hinder the team’s productivity
• Ensure the project team and the resulting project output comply with regulatory, ethical, and legal needs
• Manage a cycle of deliverables that meet timeline and resource constraints
• Design, manage, and evangelize effective agile workflows; often using Kanban, Scrum, and Data Driven Scrum
• Manage contracts with vendors and suppliers

Note that some positions also require a technical bent whereby the data science project manager is also expected to be proficient in Python, SQL, and visualization tools like Tableau. Often, such a role is labeled as a “technical project manager”. Regardless, working knowledge of databases, analytics, AI/ML, engineering, cloud systems, and the data science life cycle is key for any project manager in data science. Some roles might also specifically ask for experience within a specific domain such as marketing tech, fintech, or pharmaceuticals.

Related Job Titles

The project management function is not exclusively the job of a project manager. Indeed, some frameworks like Scrum and Data Driven Scrum don’t recognize the specific title of project manager. This doesn’t mean that the project management function doesn’t exist, but rather that various roles implement different project management activities. We’ll explore some specific roles that often hold some project management responsibilities.

• Process Master or Scrum Master – Respectively found in Data Driven Scrum and Scrum, these roles facilitate the overall process framework.
• Product Manager / Owner – This role sets the vision for a data science product and priorities that broader vision into smaller bite-size units of work. They own “scope management” from the broader project management framework.
• Program Manager – Often a senior role to the project manager, the program manager manages umbrella activities that span across multiple projects.
• Data Science Team Lead – Usually this is a senior data scientist who helps define the technical solution. They often support and train less experienced technical team members.
• Data Science Manager – They manage the data science team members. The manager might also be responsible for broader activities such as prioritization and stakeholder management.

5. Tips to be an Effective Project Manager in Data Science

1. Master Project Management Basics

Most principles and many practices of project management apply universally across projects. Regardless of whether you are a construction or a data science project manager, you have resources, constraints, risks, and objectives. And it’s your job to steer a team and its broader resources to hit the objectives as efficiently and effectively as possible. To be a successful project manager in any industry, master fundamental project management practices such as:

• resource management
• schedule management
• scope management
• risk management
• stakeholder management
• budget management
• vendor management
• effective communication
• …and the list goes on…

2. Understand the Nuances of Data Science

As an emerging field, many professionals struggle to grasp the concept of data science. As a project manager in data science, you need to understand these nuances and manage your projects in a way that caters to them.

Most notably, understand the differences between a data science and software engineering project. There is a strong overlap between these fields but data science is more exploratory and less certain. And, the data science life cycle is distinct from the software development life cycle. As such, don’t shoehorn data science as software but rather accept and accommodate the uncertain challenges in data science.

3. Be Agile

Data science projects typically face a fuzzy problem and an unknown solution. As such, you need to steer your projects to define a clear problem-space and mapping a solution that solves those problems. This is where agility shines. To achieve agile data science:

• Use iterations. In data science, these iterations are often the exploration and testing of a hypothesis. Each iteration should yield an insight that can help prioritize future iterations.
• Keep the iterations as small as possible while still keeping them meaningful. The concept of the data science minimal viable product is a useful construct to help you accomplish this.
• Get feedback on each iteration. Especially early in the project, it’s all about learning. So get feedback as quickly as possible. The project’s strong data element allows you to learn from the data. But be sure to also get feedback from your stakeholders.
• Slice your work vertically. As in try to break up your project into thin slices of value and deliver your project – one value stream at a time.

4. Facilitate Collaboration Frameworks

Great project managers in any field facilitate collaboration both within the project team and among the broader stakeholder and partnering teams. There are many different frameworks you can use. Understand these different frameworks and when to apply which one under which circumstance.

• Kanban – A lightweight framework focused on minimizing work-in-progress and maximizing throughput.
• Scrum – A product-focused framework focused on incremental batches of work delivered in fixed-length bursts of work.
• Data Driven Scrum – A variant of Scrum catering to data science projects.

5. Embrace the Challenge

Data science project management can seem overwhelming – especially at the start of the project. You might be responsible for the delivery of an unknown solution to a vaguely defined problem. You might not even know whether a solution is plausible!

Yet take this challenge with a smile. Your attitude and confidence are often the difference between project failure and success.

Data Science Project Management Certifications

There are several benefits to earning a certification. Namely, a certification can help you:

• Acquire valuable skills
• Broaden your domain knowledge
• Gain confidence
• Network with a broader professional group
• Land a new job opportunity
• Increase your salary
• Drive project outcomes

We’ll review three categories of certification – traditional, agile, and data science project management certifications.

Traditional Project Management Certifications

The world’s two most prominent certifying bodies are the Project Management Institute in the USA and AXELOS in the United Kingdom. They each boast over 1 million certificants across several certifications. The two most notable certifications are:

• Project Management Professional (PMP) – The PMI’s flagship certification tends to focus on more traditional project management approaches.
• Projects IN Controlled Environments PRINCE2) – AXELOS’ primary certification focuses on a process-based project management methodology.

This Simplilearn post explains the differences between the two.

Agile Certifications

Scrum certifications: There are several competing certifying Scrum organizations. They each offer entry-level courses that follow a similar format with minimal or no prereqs, 16 hours of class, and a relatively straight-forward exam. Each also offers advanced agile certifications which require more pre-requisites and effort. Three of the leading organizations and their entry certifications include:

Other common agile certifications include:

• Agile Certified Practitioner (PMI-ACP) is the PMI’s leading agile certification. It is also its fastest-growing certification.
• Certified SAFe Agilist is the base certification for the Scaled Agile Framework which is the most recognized project management scaling framework. There are over 1 million SAFe-trained professionals.   

Data Science Project Manager Salary

Many of the more generalized versions of these roles make Glassdoor’s 50 Best Jobs in America for 2022 list.

Given the demand for executing AI/ML projects, project managers in data science likely command higher salaries. Glassdoor’s salary reports agree.

Project Manager Salary

Data Science Project Manager Salary

Thus, the estimated average total pay for a data science project manager is about $13,000 more than a generalized project manager. Glassdoor last updated these estimates on Dec 13, 2021. See the Project Manager and Data Science Project Manager Salary reports for potentially newer estimates.

Closing Thought

Regardless of the job title, the data science project manager function is a key role for organizations and an exciting and lucrative role for individuals who want to drive success into a nascent yet powerful field.