5 Keys to Improving a Remote Data Science Team’s Efficiency and Effectiveness

Due to efforts to lower the spread of COVID-19, the vast majority of data science teams are now working remote.

Unfortunately, many teams are not used to working remote and so, are not aware of the challenges that need to be considered when working in a remote data science team.

With a well defined process and a focus on the 5 keys outlined below, teams can efficiently deliver the desired actionable insight.

1: Focus on Project Coordination

To help ensure a project can be effectively coordinated, one needs to effectively prioritize the tasks, divide up the work, and share the results.

  • Prioritizing the Tasks: Communicating the prioritization of tasks visually (e.g., Kanban board) can help ensure the team stays on the same page with respect to the most important tasks and can facilitate keeping the project sponsors / stakeholders / product owner in the loop.
  • Dividing up the Work: The team should make sure everyone understands how to effectively divide the work across the data science team, and then, how to effectively integrate work across the team.
  • Sharing Results: Sharing the knowledge gained when doing a task often requires more than sharing the code generated or an explanation of the visualizations that were produced, in that it also needs to include the sharing of the higher-level insights generated (such as an identified the dead-end that should not be further explored).

2: Facilitate Discussion and Communication 

 When data science teams work remote, some things, such as being able to walk over and talk to someone, is not possible. Teams can replicate these types of interactions, but these interactions require more thought when working remote. So, teams should:

  • Structure team communication: Synchronous updates are important (e.g., daily stand-ups can be very effective with remote teams). But ongoing discussion are also important and need to be also be facilitated (e.g., asynchronous discussions via slack).
  • Continue to brainstorm: Brainstorming is an important aspect of a data science project, but can easily be deprioritized when everyone is remote, so the team needs to be encouraged to brainstorm when everyone is remote (e.g., potential actionable insight, potential new data sources). One way to take advantage of required remote work is to use, for example, an asynchronous poll prior to a meeting, to get some initial brainstorming ideas that can then be discussed. While this (and other approaches) are also possible for co-located teams, due to the potential for team members to disengage, giving more conscious thought to how to engage the team pays more dividends for remote teams.

3: Account for the Varied Work Environment

As compared to their workplace environment, some people will have a less productive home work environment (e.g., children in the house). With this in mind, to enable synchronous coordination, there needs to be a balance of people working some part their “normal” daytime hours with people having the flexibility to work when is best for them.

4: Encourage Motivation

For some people, it can be hard to be motivated while working from home (due to a myriad list of possible distractions). So, some structure — in terms of expected deliverables and communication, can help provide help to keep team members focused.

5: Don’t Forget the Team Culture

 With everyone being remote, it can be difficult to keep the team culture (attitude, community, etc), in that, with everything going on, it can be easy to have the team drift apart and for the culture of the team evaporate. So, extra effort should be made to help ensure a positive culture and the team should explicitly decide important cultural questions (that might get implicitly decided), such as:

  • How much time is synchronous vs asynchronous should there be during the day?
  • How much overlap in the workday is required?

Finally, to help with team building, I have recently started remote happy hours (everyone brings their favorite beverage to a video conference, where people discuss non-work related items).

Note: To help spread the word, this blog was also posted on towardsdatascience

Monthly Newsletter

The industry is constantly evolving which is why you should stay up to date with the latest in data science project management research, practical tips, and news.