Welcome to this, the final lesson in the DataOps methodology course. This lesson outlines the process for reviewing, refining, and changing the DataOps process. DataOps is a fusion of people, process, and technology. The process should be reviewed on completion of every iteration. The review process should examine the people involved, the technology and applications used in the DataOps process, and the processes and workflows used to automate the DataOps process. Following this review, the DataOps process may require refinement and changes in the organization, technology, and or processes and workflows. Let's work through each step of the methodology to determine what we want to question about improving each time. The objective of the established data strategy phase is to define both an immediate and longer-term data strategy. In reviewing the effectiveness of the process, questions should be asked such as; is the strategy documented? How well has it been communicated, understood, and agreed by the organization? Is the technology in place to support the data strategy? Is the data strategy in line with the business priorities? Organization and effective collaboration between different teams is key to achieving a successful data strategy. When assessing the DataOps organization, it is important to focus on ensuring that roles and responsibilities are clearly defined. That the objectives of all the team members are in sync and that the support of key executives for the DataOps strategy is in place. Team communication is also critical and the processes for communicating and dispute resolution should be assessed. Likewise, the process for resource allocation and the skills profile and training requirements of the team members should be reviewed. Automation is a critical success factor in any DataOps strategy, and the established tool-chain phase is where the tool-chain and workflows are defined. Several workflows need to be defined as part of the overall DataOps process, for example, to automate source code control and project change requests. When reviewing these workflows, we need to examine if all source code and project changes were under control. Did any bottlenecks arise? Did the rollback capability work? The process for discovering data, classifying data, defining the full-length cycle of business terms also require automation. When reviewing this, we need to understand the degree of automation achieved, how much manual intervention was required, the speed and accuracy of the process, and what improvements are possible. A tool-chain for communication, project planning, task assignment, and reassignment is also required. Again, when reviewing, we need to understand the automation achieved. We need to understand the speed and the accuracy. During the establish baseline process, the intent is to understand where an organization is in terms of organization maturity, governance, and the inside and external influences. When reviewing the process for establishing the baseline, we need to understand if an inventory exists of systems and applications. Are naming standards in place? Does the organization have governance policies defined and KPIs identified? Is there an understanding of external regulatory requirements and a plan in place to achieve that? The objective of the establish business phase is to define business priorities and identify a prioritized list of supporting data tasks. A subset of these data tasks are then identified for implementation in a given sprint. When analyzing the establish business priorities phase, key issues are; are the business priorities clearly defined with the supporting set of data tasks? Has a data task backlog been defined and are the data tasks fully defined in terms of priority score, data sources, critical data elements? For the subset of data tasks selected for implementation, were the estimates accurate, achievable in a sprint implementation? Are there differences between different data sprints and can we understand why? For a set of data tasks to be implemented in a data up sprint the discovery phase maps the data required to the data sources, and identifies any gaps and mismatches. When evaluating the discovery process, we need to evaluate the degree of automated discovery achieved and the amount of manual intervention required. How accurate are the data mappings resulting from the discovery process? Have the data items been mapped to the business terms? What percentage of the required data was not available? Are process changes needed to support that? How accurate was the estimate for the discovery process? The data classification phase classifies the data items discovered in the previous phase and assigns to business terms. In assessing the classification process, questions need to be asked, such as whether classifications have been created for business meaning, confidentiality, and retention period. What percentage of the data items were automatically classified? What was the accuracy of the data classifications? The data quality phase defines the quality metrics for an organization, assesses the quality of data, remediates any issues, and monitors that quality. The DataOps process should have at a minimum, quality dimensions defined and agreed and a dashboard in place to monitor the quality. Any assessment of the process should understand the degree of automation achieved in gathering the data quality KPIs and monitoring them. Likewise, average time to remediate in quality exceptions should be reviewed and the overall quality trends examined in the organization. The managed policies phase is responsible for assigning appropriate governance policies to data items and data categories. When assessing the effectiveness of the managed policies phase, we need to measure if data policies have been assigned to all categories. What percentage of the data items have policies assigned? How much automation versus manual intervention is required when managing the policies and assigning them to data cache reason data items. The objective of self-service is to enable data consumers to find the right data and to process that data in the way they want to use it. Reviewing the self service process involves reviewing how easy and accurate it is to find the right data. What percentage of data is of interest and is readily accessible? How clean and documented is the data? In other words, how easy is it for data consumers to find and work with the data without needing any further assistance? All of the above require a platform and tuning fit for purpose. The data movement and integration phase is responsible for moving the data between data sources and targets using the appropriate tooling. Any review of the process needs to review the tuning selected and how effective it was? What architectural choice was made when designing the data movement and integration? What's the right pattern for data movement applied? How accurate were the sprint estimates? The objective of the complete and improved phase is to monitor and assess the quality of the data for a specific sprint and put assets delivered as part of the data sprint into production. The improvement complete process requires KPIs to measure the effective of the sprint, the involvement of stakeholders, identification of problems, and associated resolutions. Key questions which must be asked when reviewing the improved process are; how effective and accurate is the KPI dashboard? How involved were the stakeholders in identifying bottlenecks and suggesting improvements? DataOps is not a once off process. It is an ongoing process to implement DataOps across all parts of the organization, applications, and source systems. It involves long-term strategic goals and short-term implementation initiatives. A critical component is continuous measurement, monitoring, and improvement of every stage of the DataOps process. DataOps looks to break down silos across IT operations and software development teams. It encourages line of business stakeholders to work with data engineers, data scientists, and analysts so that the organization's data can be used in the most flexible, effective manner possible to achieve positive business outcomes.
