How can you make data quality training more interactive?

Training is an essential component of our work in IT industry. A few ideas that we have implemented at Addend are: 1. Add people as shadow resources to different projects. 2. Maintain a live list of internal projects that team members can do. These projects are managed just like client projects. 3. People share their projects in monthly team meetings and training so that all other team members are also aware.

1. Put data quality in the perspective of daily life. (like groceries or electricity) 2. Create personas for people in the office, those that are in the more operational might have a different use of data quality versus marketing or upper management 3. Create company wide data quality issue hunts (gamefication) Engagement is key for change in data literacy in the company

Getting it right is critical, I recommend these steps in a hands-on exercises, Identifying: 1. Data structured/semi, unstructured & format 2. Data volume generated daily 3. Data behavior options 4. Data source extraction capabilities 5. Data schemas, master data/golden records 6. Master data, Hierarchical structures 7. Geolocation data 8. Relationships between your data objects 9. Data quality issues in source data Afterwards, performing these actions: 1. Perform a distinct count and percent 2. Missing or unknown data 3. Perform a min/max/avg string length 4. Checking relationships and fixing issues 5. Checking if data fields are formatted correctly 6. Defining how to correct data quality issues Do it using Python or paid tools.

Use stories. Stories are a powerful way to teach that helps people to absorb and retain the lessons fully and longer. Pick common scenarios (real or hypothetical) and walk people through them to highlight the issues, interventions, proactive and reactive measures, and more. Make them identify and solve these data quality issues, which helps with application and reinforcement.

Where possible, I make DQ training interactive by interviewing/soliciting existing quality issues from the trainees prior to the training session. I then work with the training provider to adjust/modify their courseware. The investment proves to be invaluable for the trainees.

I think the most important element for having engaged data quality training is the use case being selected. Taking real world examples would really help in generating the interest in the audience to solve the problem. The use case selection plays a critical role as it should include all the parameters required to focus in data quality .

1) Create a very specific scenario that emulates a current business process or business problem. Ensure you provide an overview of the process, the different roles and responsibilities against personas in the process, and step through the tools/actions/activities. 2) Create Test data and similar data environment/landscape 3) Use the real tools, but in test or training environments 4) Step people through the scenario/use case, and get them to apply the real world scenario through the tools and "production like" data. Viola, they should be ready to apply in a real world scenario.

The way to generate interest in learning is to bring the problem to the real world and solve it through a use case. However, real-world data tends to be "dirty" and imbalanced, making it very challenging to apply a successful case to a beginner. It is best to start with pre-tested data and cases ready for learning before venturing into the real world.

This is something which has work very well for me when I take the data quality training. I start with the real life example how some enterprise, impacted due to data quality. Like a reporting in terms of statutory requirements had issues in data, leading to enquiry. Real life example helps the people to understand the importance of data quality.

Real-life scenarios in data quality training provide tangible examples that participants can relate to, enhancing understanding of concepts. Icorporating real-life scenarios helps in the following areas, 1. Relavance 2. Engagement 3. Application of Concepts 4. Problem-Solving Skills 5. Retention / Easily Memorable 6. Behavioral Change By incorporating real-life scenarios, data quality training becomes not just theoretical but a valuable tool for improving participants' skills and decision-making abilities in handling data effectively.

One thing I have found useful was a "Drop In Time". This is a scheduled 'consultation period' where it could be an hour or half an hour a week that people can dial into a teams call, drop by an office or desk, and always find someone is there to support from a training or problem solving perspective. Teams channels, slack, other forms of communication are always great but sometimes people want a human and that human does not want their work interrupted all day!

To make data quality training more fun, work together by doing activities in groups, talking about problems, and solving them as a team. Keep asking for opinions during the training, so everyone can share their thoughts about what they're learning and how it's going. Working together like this helps everyone get involved and learn from each other, making it easier to use what they've learned about data in real situations.

Enhancing data quality comprises two key facets in its pillars: predefined checks ensure quality, and addressing issues is the second focus. Bridging these requires fostering collaboration. Encouraging users to provide feedback on their datasets enables a system to translate insights into enhancements, fostering a shared understanding of unknown issues. This loop, exemplifying collaboration's impact, becomes a valuable training tool, showcasing how collaboration optimizes data utilization.

The effective integration of collaboration and feedback in data quality training is fundamental in creating a more engaging, supportive, and practical learning environment. This approach not only enhances the acquisition of technical skills but also develops essential soft skills like communication, teamwork, and critical thinking. My best practice: Design structured activities that encourage collaboration, such as group projects, peer review sessions, or team-based challenges. These activities should be aligned with learning objectives and encourage active participation from all members.

Use real or realistically simulated data sets to give learners practical experience. This helps them understand the complexities and nuances of working with data in real-world scenarios. Furthermore expose learners to a variety of data quality tools and technologies. This not only broadens their skillset but also prepares them for the diverse tools they might encounter in their professional life.

The use of real data for the training of data quality skills is one of the best ways to do this. The training should focus on the types of analytics that should be applied, and not necessarily on the programs and code to be used.

A twist on this is to make the exercise part of all new hire training, perhaps starting in departments that handle personally identifiable information or company sensitive data. Publish dashboard results to executives and make data handling guidelines readily available to reinforce the adherence to quality standards as part of all project post mortem or project close out exercises.

During these sessions all attempts should be made to break the data, find the "holes" and execution of the "rainy-day" or exceptional scenarios Acid test of the underlying process and code lies in these boundary conditions which are highly improbable but aren't impossible. A library of such scenarios should be created and the Testing and QA process should assure the data and the underlying code to stand the test of all these exception scenarios. This in a way would also help refactor and bolster the code and fool proof the data to be fully fit for purpose for business decisions, insights and strategy

We indeed learn best by trying our work hands-on ourselves. What needs to be considered, though, is how long to let your wards struggle with the solution and when to step in and help. Unfortunately, there is no ready-made answer, you need to go by feeling.

I personally use a lot of personal anecdotes and/or cultural references. This help people remember what I explain better and relate better to the topic at hand.

Storytelling!!! Take the learner on a journey through poor data quality. Once they see and feel the impact of poor data quality, they will have a vested interest in improving data quality. Our brains are hardwired to remember stories, partly because there is an emotional connection to a story. Once you have taken the learner on journey via storytelling, they are much more likely to recall and implement what they have learned. Training in data quality, may seem like potentially boring topic. If you can show the learner how it impacts their daily life, you can not only make the training more interesting you can turn them into a data quality champion.

The power of storytelling lies in its ability to connect with the audience on a personal level. Develop stories that are not only informative but also relatable to the learners' experiences. For example, narrate real-world scenarios where data quality played a crucial role in business decisions, highlighting both successes and failures. These stories should be contextually rich, allowing learners to see the direct impact of data quality on various business aspects, such as marketing, sales, or finance. Incorporate a mix of visuals (like infographics and diagrams), audio (such as podcasts or interviews with experts), and videos (like tutorials or real-world case studies).

Storytelling aided with visuals or other multimedia has helped put data insights into context for and inspire action from the audience.

A 24/7 online learning platform is a convenient and effective solution for participants with the same learning objectives but different work schedules across continents. The course can comprise several weekly video sessions with slides, transcripts and a weekly test. There is a deadline to complete each test within a time limit once started. The students acquire hands-on skills by signing on to a lab system with practical exercises. An online forum enables trainers and students to communicate and share knowledge. A webpage is dedicated for answered FAQs. After the course, there is a final exam to attempt within a specified duration once started before a deadline. A certificate is issued with a passing score based on weighted tests and exam.

A combination between theory and practice. Explain first what defines the data quality. For each defintion provide one exercise to ensure the reader geys hands on experience.

I follow the below steps for training session: - Create couple of dummy datasets with DQ issues across different dimensions like completeness, Integrity, Accuracy, Consistency, Timeliness etc. - As we progress, request participants to figure out DQ issues in the given dataset - Then, ask the participants to suggest remediation strategy and approach to fix the issue

Data quality can be a housekeeping item. This is why it can take a backseat. It is essential to help people visualize the consequences of bad data quality. This can be highlighted through case studies and scenarios. There are famous cases like NASA losing a 125 Million $ Mars orbiter because data units were recorded incorrectly. Amsterdam’s City Council lost €188 Million Due to Housing Benefits Error In 2014, when Amsterdam used software programmed in cents instead of euros, it sent out €188 million to low-income families instead of €1.8 million and was forced to ask for the extra money to be returned.

These aspects will help highlight the impact and urgency of the veracity of data and the need to ensure the right quality of data: 1. What all across the process and data value chains will be impacted owing to poor Data quality 2. What could potentially be the cost of low Data quality borne by the enterprise every day 3. Highlighting the fact that Data quality is an indicator of clarity in the organisation's processes, robust system and data stewardship and process ownership 4. The defining advantages of Data being fit for purpose and the perils of having issues in the data This is a non negotiable discipline in the Data value chain and with the advent and explosion of AI in the modern world, veracity has assumed paramount importance.