We added a summarised version below for those who prefer the written word, made easy for you to skim and record top insights! 📝
Additional note from community moderators: We’re presenting the insights as-is and do not promote any specific tool, platform, or brand. This is to simply share raw experiences and opinions from actual voices in the analytics space to further discussions.
Prefer watching over listening? Watch the Full Episode here ⚡️
Introducing Rho Lall | Our Analytics Hero at Your Service 🦸🏻♂️
Rho is a seasoned expert in the field of analytics and data engineering, known for his deep understanding of the complexities of data modelling and architecture. With extensive experience in conceptual, logical, and physical dimensional modelling, Rho has been at the forefront of transforming data into actionable insights that drive business success. We highly appreciate him joining the MD101 initiative and sharing his much-valued insights with us!
In Rho’s own words,
I help analytics and engineering teams clean out legacy logic by building people and pipelines through 3 best practices: emphasis on impact, data modeling for scale, and clean code. Data is messy. I get dirty. As a data analytics engineer, I apply best practices for data migration, refactoring, and query optimization to streamline processes and pipelines. I build on the modern data stack: DBT, Snowflake, AWS, GCP, lookML, Looker, Tableau.
We’ve covered a RANGE of topics with Rho. Dive in! 🤿
TOC
Champions for Data from Business Domains: From Finding to
Evolving as a Data Modeller
Impact of Data Modeling on the MarTech Stack
Championing the Customer’s Voice as a Priority
First Step to Scale Models: Performance Optimization or Plotting Business Gaps?
Clear Distinctions Between and Goals of Logical and Conceptual Models
A Week in the Life of an Analytics Engineer
What Analytics Engineers Don’t Need in their Arsenal
Assessing the Fitness of Data Tools: What Really Matters
Impact of Powerful Code Behind the Scenes
Is Retiring a Tool a Win for the Data Team?
Tools for Lineage Graphs, Dev Explorations, & Stakeholder Engagement
The Gap in Modern Analytics Stacks/Processes
When in War with Legacy Models, Burn the Boat
An Apt Analogy For Your Data Stack
What Role Has the Potential to Impact Analytics Like No Other
Perspective on Semantic Layer and the Future of Dimensional Modeling
The True USP of Data Catalogs
An Analytics Engineer’s Perspective on Data Products
Top Go-to Resources
Acing Work-Life Balance
Champions for Data from Business Domains: From Finding to Collaborating
My recent project was deeply rooted in analytics engineering, specifically moving from Heap to Google Analytics 4. It involved not only a technical shift but also cross-department collaboration with teams across marketing, external analysts, and consultants. We focused on cleaning up historical data and streamlining processes while aligning with the marketing team's goals.
Another major project was benchmarking, which, while simple in theory, was a complex data task that involved metrics modelling and dimensional modelling. I enjoy exploring the complexities of data and solving real-world challenges with it.
Evolving as a Data Modeller
My journey in data modelling has evolved organically, starting from working with spreadsheets and learning Excel functions like VLOOKUP. As I moved into data engineering, I started using Python to handle larger datasets and SQL to manage more complex queries. Early on, I worked with Kimball for modelling, which was a great foundation. Now, I’m exploring Data Vault, which feels like the next step in the evolution of data modelling, giving me new perspectives on structuring data.
Impact of Data Modeling on the MarTech Stack
The role of an analytics engineer in a MarTech stack is dual-sided: technical and analytical. While SQL and Python are crucial, the real value comes from understanding the data’s conceptual layer. It’s about interpreting what data points like “ad spend” or “sessions” mean in different contexts. By collaborating closely with analysts and business stakeholders, we can ensure that the data model is aligned with business needs and flexible enough to adapt when those needs change. Analysts’ insights into best practices, as well as their understanding of marketing terminology, are vital in refining a robust MarTech stack.
Championing the Customer’s Voice as a Priority
Understanding customers' needs is about more than just transactions—it’s about their deeper motivations. One impactful moment was when we discovered a return reason for a pet bed: “my pet died.” It wasn’t an obvious improvement to the customer experience, but it deeply resonated with us. The key is to listen to the customer's voice—understanding their passions, values, and unique needs—rather than simply targeting demographics. It’s about creating a brand that’s built on real connections, not just selling products.
First Step to Scale Models: Performance Optimization or Plotting Business Gaps?
On the scaling side, it’s all about performance—your query either performs or doesn’t. In Snowflake, I focus on optimizing queries by narrowing down issues, like checking joins first. It’s not about admiring code like art; it’s about making it work efficiently and quickly. Optimizing isn’t just about the code but also making it readable and understandable for anyone who reviews it. Whether it’s a bug fix or a business change, the goal is always to improve performance and clarity in a way that saves time and resources.
Clear Distinctions Between and Goals of Logical and Conceptual Models
The logical and conceptual models are part of the data modelling process, and their distinctions are essential in ensuring clarity and relevance in analytics. Conceptual models deal with the broader, business-driven context (e.g., understanding gender through the "sex" field), while logical models focus on the raw, technical structure of the data, like field names and types (e.g., Titanic dataset's fields). The goal is to refine and clean the data to make it clearer and more aligned with business needs, like separating the "sibs P" field to better reflect its data.
A Week in the Life of an Analytics Engineer
An analytics engineer's week is a mix of solitary focus and collaborative problem-solving. The guest enjoys the quiet of Monday mornings, working on code and finishing up tasks from the previous week. While Tuesdays and Thursdays are reserved for meetings and interactions with colleagues, Wednesday is dedicated to deep work and coding. Solving bugs brings immense satisfaction, with the rewarding feeling of fixing something that was previously broken.
"I love fixing stuff that's broken, and that feeling of like it's broken and now it's fixed... that’s a really cool feeling."
What Analytics Engineers Don’t Need in their Arsenal
Analytics engineers should prioritize learning SQL, especially for querying, joining tables, and using window functions. Tools like Snowflake, DBT, and Mage AI are useful, but they are secondary. You don't need expertise in machine learning frameworks like TensorFlow or Keras. The real skill lies in wrangling messy data, understanding how data flows, and developing hands-on experience with it. Getting comfortable with tools and data sets is key, not necessarily mastering every tool available.
Assessing the Fitness of Data Tools: What Really Matters
When assessing tools, focus on how well they fit your project and your ability to use them effectively. Learning to query large data sets, not tool experience, is critical for landing a job. Tools like Snowflake, DBT, and Mage AI are helpful, but you don’t need to be an expert in them to start. The goal is to understand how to apply data tools in real-world scenarios. Specializing in tools becomes important as you progress in your career.
Impact of Powerful Code Behind the Scenes
Clean code contributes to a tech stack's effectiveness by ensuring clarity and maintainability. It should be so clear that comments are unnecessary. Mark Romero taught me the importance of writing code that speaks for itself, which is crucial for long-term understanding, especially when revisiting old code. Additionally, reviewing others' code broadens your perspective and helps improve your approach. It's about consistency and clarity—like the debate on leading versus trailing commas, where consistency wins. Writing clear code with proper whitespace and explanations, even for complex logic, is key for future adaptability.
Is Retiring a Tool a Win for the Data Team?
Retiring a tool can be a huge win for the team, especially when it reduces costs. It’s essential to periodically review the tools in your stack and determine their value. If a tool can be replaced or retired without major disruption, that’s a success. A modular approach helps, as it allows tools to be swapped out more easily. Retiring expensive tools not only saves money but also contributes to a more agile stack. Ultimately, supporting the broader team’s success—whether in marketing or other areas—ensures collective wins, regardless of individual contributions.
Tools for Lineage Graphs, Dev Explorations, & Stakeholder Engagement
DBT and Atlan play significant roles in lineage graph visualization. DBT, in particular, offers an evolving and developer-friendly interface with a dedicated explorer tab for non-technical users like analysts and data-savvy decision-makers. This tool makes it easy to communicate technical details in an accessible way, even using it for presentations to engage stakeholders interactively.
The Gap in Modern Analytics Stacks/Processes
The primary gap in modern analytics stacks is communication. Data teams often have valuable insights, but the key challenge lies in effectively conveying those insights to decision-makers. Analytics professionals must learn to communicate in ways that resonate with others' perspectives, even if they fall outside their comfort zone. It’s crucial to adopt a persuasive approach and recognize the importance of clear communication in data-driven decision-making.
When in War with Legacy Models, Burn the Boat
Dealing with legacy models involves tackling spaghetti-like systems with inconsistent data models. The solution starts with a "lift and shift" approach—moving data to a modern tool and focusing on specific areas to optimize. The key to success is the willingness to delete old code once it’s no longer in use. This requires a disciplined, project-managed effort, where even a part-time project manager can guide the process of untangling complexity and ensuring clean, optimized systems.
An Apt Analogy For Your Data Stack
A data stack is like an attic. It's where you store things that you don’t necessarily need every day but want to keep. Over time, different people start filling it up without organization, and the newer items get closer to the door while the older items are buried. This leads to a cluttered space. The solution is to categorize, build shelves, and organize everything, making it searchable and easier to access. You also get rid of what you don't need, transforming chaos into a well-organized library.
What Role Has the Potential to Impact Analytics Like No Other
The role of someone who understands both deep technical skills and the business side could be transformative for analytics, especially with tools like AI and Snowflake improving. However, it's a lot to manage, from AWS to Google Analytics 4. A balance is needed between technical know-how and business understanding, which will be key to success in driving key KPIs like ROI or ARR.
Perspective on Semantic Layer and the Future of Dimensional Modeling
The semantic layer often means different things to different people, but it can be simplified as a "metrics layer." It's where metrics are modelled, pivoted, and aggregated for use in tools like Tableau or Power BI. As for dimensional modelling, fewer tools will be needed with the rise of integrated solutions. However, while tools like Looker are sticky, they’re not always perfect. The future is likely to bring more specialized roles and tools, and the skills for these will evolve over time.
The True USP of Data Catalogs
Data catalogs, particularly tools like Atlin, simplify conversations for analytics engineers, not by directly resolving piled-up requests but by facilitating cross-team collaboration. Atlin, with its seamless integration and ability to cater to non-developers, played a pivotal role in enabling these discussions and gaining business buy-in.
An Analytics Engineer’s Perspective on Data Products
The concept of a data product is fluid, but it can be viewed as anything that adds value through data—like how a beard oil comes with information that enhances the user's experience. Data products go beyond raw data and include the contextual services that support the end-user, whether in business or consumer products.
Top Go-to Resources
LinkedIn and local meetups are key resources for staying updated in the data community. LinkedIn hosts a vibrant, supportive data network, while local meetups offer the opportunity to engage with peers and discover new tools like DBT. Starting a community in your area, like a low-key data happy hour, can also foster valuable connections.
Clip Acing Work-Life Balance
Maintaining work-life balance isn't about avoiding work but about enjoying it. Just as Stephen King integrates his love for writing into his daily routine, work is a passion. Balancing it with quality family time—without distractions like phones—helps maintain harmony. A balanced life also means taking breaks and unplugging when needed, like during a storytelling festival.
📝 Note from Editor
The above insights are summarised versions of Rho’s actual dialogue. Feel free to refer to the transcript or play the audio/video to capture the true essence and details of his as-is insights. There’s also a lot more information and hidden bytes of wonder in the interview, listen in for a treat!
Guest Connect 🤝🏻
Connect with me on LinkedIn 🙌🏻
Share this post