The mission of the Data Analytics team is to maximize the impact of business decisions and strategy with data solutions that are trusted and scalable.
We do this by helping all GitLab teams move up the Data Value Pyramid by means of our GitLab values and our Data Team Principles.
Of the Data Team's Responsibilities the Data Analytics Team is directly responsible for:
Additionally, the Data Analytics Team supports the following responsibilities
At GitLab we are Handbook First and drive this by:
Data Visualization: Tips and Tricks and Data StorytellingAnalytics is the practice of turning data into information, knowledge, and wisdom to:
Our short term goal for FY21-2H is to move GitLab up the Data Capability Model from Level (1) Reactive to Level (2) Advanced.
To do this, the Data Analytics team will be:
Data Analytics, Advanced Data Analytics, and Data Science all start with the Data Analysis Process.
The Data Analysis Process has five steps:
In Data Analytics, context is everything. It guides the way Data Analyst view the problem, the data, and the methodology taken to generate the data insights.
In the Problem Statement step, it is crucial that we define the business problem with the Business Partner. By clearly defining the business problem that the Business Partner is seeking data insights into, Data Analyst can set success criterias to ensure that the problem is technically solvable with the data available and to provide the maximum business insight into the problem.
In the Data Collection step, it's important for Data Analyst to explore and understand the Data Caveats that may limit the impact or generalization of the analysis as well as to evaluate any Data Biases or Data Concerns. Understanding how the data is collected in each business system will ensure that we are reporting on accurate data in an appropriate manner. During this step, it's important to ensure that the Data Types are accurate for reliable transformations.
Therefore, this step is actually a part of a feedback loop with the source system owners. If a Data Team member finds that a data field is missing or inaccurately captured, it is the responsibility of anyone at GitLab to reach out to the source system owners and ask them to update or add the data field in question.
In the Pre-Processing step, Data Analyst prepare and cleanse the data and exam the data quality to design the optimal Data Structures for scalable reporting. It's important in this step to review the model and ensure that the model is business-friendly.
In the Modeling step, we aim to create new data models in the Enterprise Dimensional Model format for all GitLab team members. We start first with an Entity Relationship Diagram (ERD) to ensure that the new data structures reflect accurate business processes. All data models are reviewed by the Data Platform Team.
When GitLab reaches Level (4) Predictive of the Data Capability Model, this step will include Exploratory Data Analysis and Data Feature Engineering, which should include supervised & unsupervised machine learning techniques.
In the Presentation step, Data Analyst radiate in their Data Storytelling skills with strong Data Visualizations. Data Analyst recognizes that it is important to package and display the actionable business insights to effectively communicate with stakeholders. In fact, it is as vital as ensuring that the data solution, such as a Business Intelligence (BI) dashboards with drill-downs to show the granular reasoning behind the insights, has robust data quality checks for our Business Partners (all GitLab team members) to trust the data.
The wonderful thing about Data Analytics is that each insight should propagate a wave of additional business questions to solve, which then allows the Data Analysis Process to cycle again.