Intro to Data Science & Analytics at Nextdoor

Intro to Data Science & Analytics at Nextdoor

Data Science & Analytics at Nextdoor is, at its core, about decision-making … data science is an evolving role, and at Nextdoor, we want to give people the option to grow, explore, and build their skills across a variety of disciplines. And, we’re very much of the opinion that that diversity of skills is beneficial in the data scientist’s work.

Author: Carly Villareal, Lead of Data Science & Analytics

A copy of this article was also shared on the Nextdoor Engineering Blog, which you can see here.

What is Data Science & Analytics at Nextdoor?

One of the first questions I ask any data scientist I meet is, “What type of data scientist are you?” Data science is an extraordinary role; the thing I love most about it is the diversity of backgrounds and skill sets that data scientists bring to the table. But, because the discipline of data science is growing so rapidly, and because “data scientist” can mean so many things, someone will say “I’m a data scientist…” and what they’ll actually mean is:

  • I’m a statistician
  • I build production-level models
  • I’m an ETL whiz
  • I’m an analytics-driven product strategist
  • … or many others!

These are all important, and hypergrowth companies like Nextdoor need them all! But, they’re all very different, and we needed to define a common thread. Data Science & Analytics at Nextdoor is, at its core, about decision-making. If we look across the entire team, across all the disciplines listed above, the thing that unites them is that all of our Data Science & Analytics employees are contributing to decision-making every single day. Our analyses help stakeholders make decisions. Our models make decisions in the product. We make decisions ourselves about how Nextdoor should access and use data.

The most important thing we’ve found in hiring is to make sure that the type of decision-maker we’re bringing in is well-suited to the needs of the team and the product they’ll be working with. A mismatch in expectations between the data scientist and the team is a recipe for unhappy data scientists, unsatisfied stakeholders, and high turnover.

The Problem

But, here’s the problem. People are not just one thing. When I started at Nextdoor in 2015, the data team (including data engineering) was 4 people in total. Since then, the data science & analytics organization has grown significantly, and we’ve had the pleasure of working with and talking to many more along the way. The skill sets and interests in that group are extraordinarily diverse; the only thing they have in common (at least from a technical skills perspective) is that they defy classification.

Data science is an evolving role, and at Nextdoor, we want to give people the option to grow, explore, and build their skills across a variety of disciplines. And, we’re very much of the opinion that that diversity of skills is beneficial in the data scientist’s work; the best modelers are the people who can test and analyze the output of the model, the best ETL writers are people who intimately understand how that data will be used in analysis, and the best analysts are people who can identify how their work might lend itself to simple or complex models to improve the product.

With that in mind, we needed to figure out how to define “what a data scientist is” at Nextdoor. How do you support people who grow across disciplines? How do you support people in roles that require deeper specialization? How do you measure career progression? And, finally, how do you make sure that data scientists are able to see how they themselves can grow and be successful at Nextdoor?

The Role of Data Scientists at Nextdoor

We’ve developed two frameworks to define a Data Science role here at Nextdoor: Big Decisions and Core Technical Skills. The Big Decisions framework helps us evaluate the impact of the employee’s work, while Core Technical Skills helps us evaluate their expertise. To be successful, you need both. Using the two frameworks together has allowed us to show the wide variety of different ways that a data scientist can be successful here, and has allowed us to better define the roles and levels of the people we’re hiring.

Big Decisions

Because the core of Data Science & Analytics at Nextdoor is about decision-making, we can define a concept called a Big Decision, which happens when a person or a product makes a decision that impacts the company.

  • We can build a model, which can make Big Decisions at a small feature level all the way up to a major system level.
  • We can formulate and run an analysis, which can drive Big Decisions at a team level up to a company strategy level.
  • We can make Big Decisions about how a team (or the whole company) should use or interact with our data.
  • We can build a data set or automate data delivery, which can drive Big Decisions at a team, organization, or company level.
  • We can determine how tests and experiments should be run & analyzed, which defines how the company makes Big Decisions about feature changes.

Fundamentally, we measure our team’s impact by the impact of the decisions that we help make. The framework works because it’s inclusive of all of our data science disciplines, and because it ensures that we’re continuously driving value, regardless of the type of data science work the person is focused on.

Core Technical Skills

The other framework we’ve implemented is the concept of Core Technical Skills. One of the strengths of our team is that we have many types of people with many diverse skills. We do not want to dictate or limit learning to particular areas — rather, we want to demonstrate excellence in the areas that matter most to an employee’s particular role, and we want to reward people with deep specialization as well as broader skill sets.

Core technical skills at Nextdoor include (but are not limited to) the following:

  • Mathematics (Linear Algebra, Calculus)
  • Statistics and Probability
  • Machine learning
  • Data manipulation
  • Data table design & architecture
  • Data engineering
  • Data visualization
  • Experiment design
  • Cartography and GIS
  • Product analysis
  • ETL design
  • Computer language fluency
  • Qual/quant crossover analysis
  • Survey design and sampling
  • Software engineering

We have had successful data scientists in all of these areas, including people who go deep into one area and people who can do a little bit of everything. When we combine the Big Decisions framework with the Core Technical Skills list, we’re able to build a comprehensive picture of each data scientist’s role, and determine what growth looks like for them at Nextdoor.

Final Thoughts

Our head of engineering, Antonio Silveira, has been heard to tell new managers that “people are the hardest problem you will ever try to solve.” As data scientists, we’re natural quantifiers, and there’s a tendency to want to map people out, put them on a chart, and set them on a clear trajectory. But — people are more complex than that, and paths change. (We’re growing quickly enough that our needs as a company change all the time, too!) The important things, for us, are that we figure out a way to define what a data science role looks like, focus on decision-making and impact across all of our disciplines, and give our data scientists visibility into what their career development and growth path looks like here at Nextdoor.

On that note, we’re hiring!