Put data behind every decision

Before the pandemic, many a critical issue had to be solved 'on-site' fuelled by the late-night staples Peanut M&M's and black coffee, with the whole floor plunged into darkness except for the glow of your desk. Today, 'working from home' has certainly made life easier for the SysAdmin. Much of the work can be done remotely, which is great – unless that server just refuses to come back online, forcing you to get in the car and drive directly to the server room.

The modern cloud paints a lovely picture of Fluffy Clouds – an analogy that represents compute for everyone: simple, ethereal, always available. The cloud drives so much of what we do today, not just at work but in our homes too. But the reality is somewhat different for the SysAdmin. There's still a physical computer somewhere, running in a room with lots of other computers, called Data Centres.

These Data Centres can be incredibly vast, like the Vantage Data Centres campus in South Wales, spanning 750,000 square feet (about 10 standard football pitches) and housing, at least, 100,000 servers. Navigating such immense spaces takes lonely hours of effort to physically reach a problem with a "Crash Cart" and maintain these critical machines. SysAdmins often work in super-cooled environments, wearing thick coats, scarves, and gloves to stay warm, all while wearing ear protectors against the deafening hum of thousands of machines.

There's immense stress and countless tales from working in Data Centres, and it's physically demanding work. It's not just command lines and keyboards; it's pure muscle. Picking the wrong cable amongst the "cable hells," a mistyped command, a forgotten 'enter' key, or pulling the wrong power cable – in a Data Centre – might be a small human error, but it can have catastrophic, company-wide implications. The pressure is immense.

You can't just walk in, use your mobile phone, or carry a hot drink with you. A simple 'comfort break' isn't quick either; navigating multiple layers of security and walking considerable distances means a bathroom trip can take you away from your task for a significant amount of time, often requiring re-entry procedures like de-robing, re-robing, and sanitising.

It's an adage often whispered among IT professionals: "Happy Users is a very hard-working SysAdmin." It encapsulates the tireless dedication required to keep the digital gears turning smoothly. Yet, the cruel irony is that when systems are running perfectly, the very same business users can often be heard saying, "The SysAdmins don’t do anything. Why do we need them? Everything is working!" This sentiment, while frustrating, perfectly underscores why SysAdmins truly are the unsung heroes of our digital age.

Increasingly, businesses are dependent on their digital systems and data to drive strategy and, ultimately, profits. Business leaders are reliant on reports, visualisations, and operational processes to be secure, reliable, and quick. The analyst who creates a crucial report gets the credit for their work, not the SysAdmin who tirelessly rescued the data from backup in time for the meeting, a report the analyst may have accidentally deleted.

Of course, like any profession, there are isolated incidents or misguided decisions – the occasional headline about a well-meaning but ultimately disastrous attempt to 'save energy' by turning off cooling comes to mind. However, these rare outliers should never overshadow the overwhelming dedication and meticulous care of SysAdmins who operate with unwavering professionalism.

So, as we celebrate International System Administrator Appreciation Day (or SysAdmin Day) on July 25th, let's take a moment to truly acknowledge these extraordinary professionals. They are the silent sentinels, the diligent guardians, the ones who work tirelessly behind the scenes – often without credit – to keep our digital world running smoothly. They endure the silent watches in vast, fortress-like data centres and perform countless unseen acts of digital salvation.

They may not wear capes, but today, maybe they can.

‍

I recently wrote a blog looking at the data world’s obsession with talking about data like it’s water. It is an analogy worth exploring further: can we think about data quality in the same way we think about water treatment?

Water has to be treated before it can be used and consumed by people across the country, so that it is safe. In a similar manner, as we take data from where it is produced to where it’s consumed, the data should also be treated  to ensure the consuming systems remain healthy and produce reliable outputs.

We can think about data ingestion and storage like a river and a lake. Data ingestion is the river which passes your data downstream, into a lake which is where your data is stored. Just like with water, it is much easier to treat issues before it passes into your data lake. Therefore data quality up-front is crucial. This would involve things such as carefully controlled rules around data entry (free text fields should be used sparingly!) and automated rejection of data received from other organisations if not up to the standards required.

An example of this you will have seen in recent years is when you select your address for most online shopping. You  enter your postcode and get a drop-down list of addresses to select from. This ensures you enter a valid address that matches one the retailer holds in their systems.

It’s no good constantly cleaning the lake if poor quality data keeps flowing in from the river. That’s why having data quality in your ingestion processes has to be your priority. Once this has been established, you can begin to explore sampling the data you hold, and understand the quality of that. This will allow you to take remediative action on your data, for example standardising formats or imputing missing values. This will increase the data’s quality  allowing better decisions to be made off your data.

There is probably a whole other set of analogies that could be explored around the recycling of wastewater and the archiving and disposal of data, but I’ll let someone more creative than me explore that!

It is worth noting that by my limited understanding, water in reservoirs is actually untreated and goes through water treatment on its way to our homes, and therefore the diagram and my whole analogy is really quite false - but hopefully you still found it useful.

If your organisation wants to improve its data quality and make better decisions, get in touch. We’d love to get your data flowing in the right direction.

‍

On my very first day, one of our directors asked if I’d ever used the rubber duck analogy. Amused by the name, I said I hadn’t.

He explained it as a technique where you talk through a coding or technical problem out loud, often to a rubber duck sitting on your desk. By explaining it step by step, you often uncover mistakes or gaps in your logic that you hadn’t noticed before. It’s a surprisingly effective method, especially when working with complex systems.

At this year’s #NHSconfedExpo, I spotted a fun display of rubber ducks and rushed over, assuming it was a tribute to this quirky debugging method. But the ducks weren’t there for that reason.

They were lined up neatly in rows, representing a different saying: “to have all your ducks in a row,” a reminder of the importance of being organised and thoroughly prepared before tackling a task.

Both ideas resonated with me, particularly when it comes to working with data.

Explaining your approach to a dataset, even if just out loud to yourself (or your duck), can help clarify assumptions, reveal flaws in your logic, and force you to think more critically. At the same time, having your “data ducks” in a row - that is ensuring your data is clean, complete, and structured correctly is essential for any kind of meaningful analysis. Without this foundation, insights can quickly become misleading or inaccurate.

Now, I keep my own rubber duck on my desk. It serves as both a conversation starter and a gentle reminder that whether you're debugging code or preparing a dataset, clarity and organisation go a long way.

If you’re curious to try it out but not sure where to begin, here’s a few helpful tips:

Find your “duck”

It doesn’t have to be a rubber duck. It could be a mug, a pen, a houseplant or anything that can be your silent listener.

State the problem clearly

Pretend your duck knows nothing about coding or what you're working on. Start from the beginning and explain what you're trying to do in plain language.

Talk through the steps

Go through your code or thought process line by line. Say out loud what each part is meant to do, as if you’re teaching it.

Listen to yourself

As you speak, you may notice things that don’t add up, or that you’ve misunderstood. That’s often where the problem lies.

Make adjustments as needed

Once you spot the issue, go ahead and tweak your code or plan. If you still can’t figure it out, you’ll have a much clearer understanding when you ask someone else!

Repeat when stuck

Don’t be afraid to do it again and again!

So whether you're knee-deep in code, analysing complex datasets, or just organising your thoughts, take a moment to talk it out. Sometimes, all you need is a duck, a bit of space to think, and the discipline to get your ducks in a row.

‍