As a Data Science Content Developer, I contributed to 26 data science courses that have helped more than 50,000 students. I helped develop courses in four different technologies (R, Python, SQL, and Spreadsheets) in topics ranging from query performance in SQL to biomedical image analysis using Python. The content I created had an average rating of 4.6/5—the highest of all Content Developers at the time, a record which I am happy to say that my successors broke! If you’re interested in viewing the content I helped build, please visit one of the links below.
- Analyzing Police Activity with pandas
- Analyzing Survey Data in R
- Analyzing US Census Data in R
- Biomedical Image Analysis in Python
- Building and Optimizing Triggers in SQL Server
- Categorical Data in the Tidyverse
- Error and Uncertainty in Spreadsheets
- Experimental Design in R
- Factor Analysis in R
- Financial Forecasting in Python
- Foundations of Functional Programming with purrr
- Improving Query Performance in PostgreSQL
- Improving Query Performance in SQL Server
- Improving Your Data Visualizations in Python
- Interactive Maps with leaflet in R
- Intermediate Data Visualization with Seaborn
- Intermediate Functional Programming with purrr
- Introduction to SQL (practice pool)
- Model Validation in Python
- Network Analysis in the Tidyverse
- Python for MATLAB Users
- Statistical Simulation in Python
- Structural Equation Modeling with lavaan in R
- Support Vector Machines in R
- Writing Efficient Python Code
- Writing Functions in Python
Online data science courses
Most DataCamp courses are created using a standard instructional design model called ADDIE. Content Developers contribute to the Development, Implementation, and Evaluation phases, which include reviewing content, building content assets, deploying the course on the platform, and evaluating user feedback to prepare the course for hard launch. For all courses, I collaborated alongside one of DataCamp’s Curriculum Managers and an external instructor, who was the subject matter expert (SME) on the course’s topic.
Content development
Working as a Data Science Content Developer requires many hats: project manager, instructional designer, and programmer to name a few. The bulk of the work, however, is diving deep into the content through proofreading & copyediting, providing quality assurance, and writing unit tests to evaluate student-submitted content.
Proofreading & copyediting
While many DataCamp instructors are seasoned teachers, very few have much, if any, experience with online courses. Online courses pose a very challenging problem: there is no student feedback during development. Typically, teachers rely on feedback in the classroom, e.g., students asking questions, which they can quickly respond to by rephrasing a sentence. I helped instructors anticipate such questions by ensuring their courses created a clear narrative flow, used strong analogies and heuristics with global reach, and implemented DataCamp’s voice and tone. As the resident DataCamp expert working on the course, I often made several edits to courses to move the copy from formal to “fun and friendly” to be more in line with the DataCamp brand.
Quality assurance
Something that DataCamp prides itself on is making some of the greatest data science content on the market. While a lot of this greatness is due to an incredibly talented and passionate group of instructors who LOVE teaching, a lot is also due to the dedication of content developers behind the scenes who ensure the following: - Adherence to DataCamp’s content and style guidelines for consistency across the platform. - Accuracy of domain-specific content to ensure clarity and prevent misinformation. - Advocacy for DataCamp learners to ensure that instructors understand and connect with their audience.
Unit testing
In addition to providing affordable access to a wealth of data science content, one of DataCamp’s biggest value propositions is its interactive exercises. DataCamp students write code themselves, submit it, and it is then automatically “graded”. Content Developers are responsible for writing these submission correctness tests, or SCTs. Submission correctness tests are very similar to the concept of unit testing, a type of test used during software development.
Writing SCTs is as much a science as it is an art—and it’s never perfect! Each SCT needs to be crafted to test the “correctness” of the student submission, but should also focus on the learning objective of the exercise, while also being flexible enough to accept different ways of coding. This is a very tall order! These are put to the test when a course is launched and student feedback can be used to either rewrite the SCT or add custom error messages. *Packages used for writing SCTs are open-source and freely available to anyone.
Key takeaways
Developing instructional content, especially technical content requires many skills. One must be an excellent project manager to effectively develop multiple courses at different stages of development at one time. One must be an excellent relationship manager to get the best out of external instructors. And one must be an avid learner to take on new content types, programming languages, design models, etc., at the drop of a hat. It is a challenging but rewarding experience that I would recommend to anyone who loves learning.