Unifying a Information Stack and Leveraging Self-Serve Analytics with Atlan
The Lively Metadata Pioneers sequence options Atlan clients who’ve accomplished a radical analysis of the Lively Metadata Administration market. Paying ahead what you’ve realized to the subsequent knowledge chief is the true spirit of the Atlan group! So that they’re right here to share their hard-earned perspective on an evolving market, what makes up their trendy knowledge stack, revolutionary use instances for metadata, and extra.
On this installment of the sequence, we meet Daniel Ferguson, Information and Analytics Director at PHMG, an audio branding firm that helps over 36,000 purchasers throughout 56 nations sound their greatest. Daniel shares how PHMG reworked their knowledge stack from fragmented to unified, and the way Atlan has been an important piece in by monitoring lineage, managing studies, and easing workforce onboarding.
This interview has been edited for brevity and readability.
Might you inform us a bit about your self, your background, and what drew you to Information & Analytics?
I was a DJ after which labored as a sound engineer constructing recording studios. Throughout my time as a sound engineer, I discovered myself within the technical and analytical facet of issues. After beginning a household, I needed a change. After promoting my recording studio, my mom, who managed a council workplace, supplied me a job and I began out within the name middle unit, dealing with calls. It rapidly turned obvious that I might do extra than simply calls, so I moved to the database workforce.
I began finding out for a level in Economics and Mathematical Science at The Open College whereas working on the council. Utilizing the abilities learnt on my diploma I began to construct logistic regression fashions to focus on contacts within the name middle I had beforehand labored in. I proposed that, with only one individual, I might obtain the identical outcomes as the whole workforce. My work generated 300% extra outcomes than the workforce’s mixed efforts by optimizing knowledge assortment, addressing lacking info, and cherry selecting the most effective contacts. After that I used to be hooked on the ability of knowledge & analytics.
I then constructed an organization offering knowledge providers to different Native Authorities. Close to the tip of my diploma, a consultancy in Scotland, Aquila Insights, supplied me a place. They labored with purchasers like Sony, Workplace Depot, and RBS, which gave me early publicity to the information occupation. From there, I superior within the area and finally joined PHMG. My journey into knowledge was considerably unintentional, but it surely introduced me to the place I’m immediately.
Would you thoughts describing PHMG?
We focus on audio branding. Consider logos like Netflix or Disney Plus, via sound alone, these manufacturers are immediately recognizable as business leaders in leisure and streaming even when their visible logos usually are not in sight.
We additionally transcend conventional audio branding by creating customized music tailor-made to every group. We’ll take Atlan for instance: What’s Atlan about? What do you symbolize? What’s the kind of rhythm that it needs to deliver?
This connection between music and id is what attracted me to the corporate. We’ve been extremely profitable, working in 56 nations with 36,000 purchasers.
Might you describe your knowledge stack, and the way it got here collectively?
After I obtained right here, we have been utilizing SQL servers with Excel spreadsheets. There have been restricted to no interactive studies, and each knowledge request needed to be raised to the information workforce.
There was a have to modernize the data flowing into the corporate and implement the precise expertise to realize this effectively and reliably. I centered on discovering expertise options that might streamline operations and scale back the necessity for added engineers.
I used to be actually cautious with expertise choice, avoiding options for the sake of it, and never constructing from scratch. Whereas Azure Material presents a complete answer, for instance, it’s nonetheless new and that comes with extra dangers, however is one thing I’m maintaining my eye on. It’s essential to decide on the most effective instruments for the job and guarantee they work effectively collectively. Investing in a seamless course of with these instruments means that you can begin robust and display worth rapidly, with room to evolve as you scale.
In my board proposal, I highlighted two important instruments: Atlan and ThoughtSpot. I defined that whereas we might handle with out them, they’d make a major distinction. I needed governance to turn into embedded in our processes, and that as an alternative of assigning knowledge stewards with out clear route, we offered actionable studies and comprehensible knowledge. With correctly organized knowledge, governance turns into easy, and Atlan streamlines this course of.
I chosen Snowflake for its robustness and cheap pricing, and Fivetran for its dependable pipeline efficiency, which successfully handles our knowledge integration wants.
I carried out PowerBI for govt studies, and ThoughtSpot for our self-serve knowledge wants. I’m a giant fan of ThoughtSpot, as a result of it permits customers to regulate their very own studies, decreasing the necessity for fixed modifications from the information workforce.
For orchestration, I exploit Airflow to handle pipelines, and DBT with GitLab for our code repository and CI/CD processes.
Why was Atlan an excellent match? Did something stand out throughout your analysis course of?
In my earlier group, I attempted utilizing open-source with DataHub, however its upkeep and growth required important funding. Atlan stood out as a result of it’s plug-and-play, routinely constructing out miners that reveal beforehand unknown insights. It identifies and explains scripts we weren’t conscious of, saving time and decreasing technical debt from having to manually overview intensive code.
Atlan lets us monitor and monitor what we’ve constructed, together with knowledge lineage and property. It’s invaluable for reviewing studies while not having to ask for code particulars—simply navigate via Atlan to see the report’s historical past. New workforce members also can perceive report development via Atlan.
For me, Atlan was a key piece of the puzzle.
I researched Collibra, Alation, and Atlan extensively, and Atlan was the clear alternative. It felt designed for medium-sized enterprises and required minimal engineering effort. Given our state of affairs, it was essential to combine it from the beginning, quite than as an afterthought. This allowed us to study and develop Atlan alongside our current programs, quite than attempting to power it into our pre-built setup.
I all the time make it some extent to satisfy with management groups at occasions to gauge their angle and dedication, and I don’t know of some other gamers which are doing it in addition to Atlan. I used to be genuinely impressed by Atlan’s management workforce — not solely their ardour for the product but in addition their dedication to addressing my challenges and enhancing our state of affairs.
How are you planning to harness Atlan to boost your knowledge stack? What thrilling use instances and targets do you keep in mind?
We’ve invested in a knowledge vault mannequin for our knowledge warehouse, which feeds into an operational knowledge retailer, what I name the information mart. All our studies and metrics are constructed from this knowledge mart. In Atlan, we outline tips on how to assemble every part, so as soon as a metric is outlined, we will write the SQL to extract it from the mart.
We then create curated tables for shopper providers and gross sales organizations, enabling them to self-serve by way of ThoughtSpot. For detailed insights into the development and rationale of those metrics, we retailer that info in Atlan, which turns into our catalog.
As new individuals come on board, I be certain that there’s no want for a handover. By default, we doc our processes as we go and construct programs that depart clear breadcrumbs for others to observe. Atlan performs an important position on this. We direct new workforce members to Atlan to assist them perceive how every part is constructed and what it’s constructed from. Atlan doesn’t simply spill out the code, it highlights the important thing objects, their utilization, and their significance.
One other main venture entails making a complete glossary inside Atlan, serving as our single supply of reality. This setting permits enterprise customers to entry all company metrics and look at studies from Salesforce, PowerBI, and ThoughtSpot, all linked round key KPIs.
We’re additionally at the moment refining our knowledge lineage and mannequin descriptions. As we create new knowledge fashions, we replace descriptions incrementally quite than in bulk. This ongoing effort helps be certain that our knowledge fashions are well-documented and simply comprehensible.
Do you’ve got any recommendation to share along with your friends who’re beginning out in managing and organizing their knowledge property successfully?
Companies all the time speak about being knowledge pushed, however they don’t discuss in regards to the property that really drive the information. We wish info to circulate in our group, however info can not circulate if it’s not organized persistently. And for me, instruments like Atlan are making it considerably simpler for us to prepare and talk what knowledge issues.
Don’t get me flawed, Atlan isn’t a silver bullet. It gained’t repair poor group inside your knowledge warehouse. Nonetheless, it does present a centralized place to outline and assess your processes, serving to you establish which of them are efficient and which of them want enchancment.
Atlan helped us decide the place to begin by figuring out our most important tables and specializing in what was necessary. As an illustration, we discovered one desk vital for every part we constructed, permitting us to prioritize it. We then assessed our studies and found that some we thought have been necessary have been related solely to particular studies, not the broader context.
As we get delicate knowledge, we will additionally instantly flag it. If we get audited, we will merely pull up Atlan and say, “Hey, that is what now we have. That is how we handle our knowledge. That is what our knowledge property are.” So, for these dedicated to being data-driven, they should take care of their knowledge property and perceive what their knowledge property are.
Picture by Adi Goldstein on Unsplash