*From the “Data but Make it Personal” series – Part 3
On understanding structure before trying to make things fit
The other day, when I was trying to get some Alteryx practice by doing one of their Weekly Challenges, I was building a pretty straightforward flow. Or at least, it should have been.
All I wanted was to join two data streams – clean, ordered, properly prepped (or so I thought). But then came the dreaded red exclamation mark, and this delightful little message:
“String fields can only be joined to other string fields”
Of course, I’d tried to join a field formatted as string with another formatted as an integer. Rookie mistake – but one that somehow still sneaks in, even after weeks of training. I fixed it quickly enough (thank you very much Select tool), but the error stuck with me. Not because it was technically complex, but because it felt strangely familiar.
I’ve made that mistake before – sure, in Alteryx, but also in life.
Trying to connect two tables – or two people – that just aren’t using the same format. Expecting someone to “just get it” when they process the world differently. Disregarding the fact that we’re not aligned not because one of us is wrong, but because we’re formatted differently.
So, let’s talk about data types and the importance of structure.
What are data types, anyway?
Data types define the kind of values a field can hold and the operations that can be performed on it. In other words, they define the kind of information you’re working with, and how it should behave. And they matter. A lot.
Trying to do math on a string? Error.
Trying to sort dates when they’re stored as text? Chaos.
Trying to join a string to an integer? The story of my life.
Talking specifically about Alteryx, it supports five general types of data:
- Strings. Text data that includes letters, numbers, or symbols, always treated as text even when it looks like a number.
- Numerica Data. Data used for calculations, including whole numbers (integers), decimals, floats, and doubles, each with different precision and storage requirements.
- DateTime. Data that captures dates, times or both, following strict formatting rules.
- Boolean. Logical data with only two possible values: True or False.
- Spatial. Data representing geographic objects like point, lines, or polygons, used for spatial analysis.
In data, understanding types means understanding how to treat the information.
In life, it might be the same.
Types of People, Too
That failed join, and the ones I’ve had since (because a girl keeps stumbling over the same rock), got me thinking about conversations – the kind that just don’t land. Not because of disagreement, but because of something subtler: a mismatch in how things are expressed or understood. A type mismatch.
I’ve felt it both personally and professionally. You walk out of a meeting knowing everyone was speaking the same language – technically – but somehow still not connecting. You couldn’t quite get your point across, and the other person’s message felt … scrambled.
Setting the technicalities aside, a lot of this miscommunication comes down to expectations – and those expectations are in part shaped by our own “type”.
Just like in data, people come in different formats. Here’s how I like to think about it – not as a hard rule, but as a helpful lens:
- Strings are expressive, full of context, emotion and nuance. Sometimes messy, but always rich with detail.
- Booleans are direct: yes or no, true or false. Clear, decisive, and sometimes rigid.
- Floats are mostly steady with subtle shifts after the decimal that can make a huge difference.
- Integers are structured, clean and predictable. You get what you see.
- Nulls aren’t empty, they’re undefined.
I’m not trying to reduce people to labels or oversimplify personalities. Even if we tend to lean towards one type, we all have a bit of every type in us, and we use them differently depending on the situation. What I’m trying to highlight here is how important it is to understand someone’s structure or mindset at a specific moment. If we get that wrong, we might apply the wrong logic – and when that happens, the connection (or join) might not work.
Quick Fix: Changing Data Types in Alteryx
Back to data. Here are two ways to fix type mismatches in Alteryx:
- Use the Select Tool. This is your go-to. Drop it at any stage of the flow (although I recommend doing it sooner rather than later) and make sure all fields have the appropriate type.
- Use the Formula Tool. You can use functions like ToString([FiledName]), ToNumber([FieldName]), or DateTimeParse() – for messy date formats.
As an additional tip, any tool that has two input anchors gives you the possibility to change data types for the output.
Understanding Before Joining
I wish all type mismatches in life could be fixed as easily as they are in Alteryx. A quick Select Tool, a formula tweak– problem solved.
But we don’t come with drop-down menus. We don’t walk around labelled as [V_WString] or [Int64].
And yet, as data consultants, we are going to be working with a lot of people – clients from different industries, backgrounds, levels of hierarchy (and levels of patience). Understanding the technical side is expected. But understanding people? That’s a skill we need to develop intentionally.
That’s why we need something data tools don’t have: awareness, empathy, and curiosity.
A willingness to listen actively, pause and ask:
“Am I expecting this person to work like I do?” “Are we actually speaking in the same format?”
Because sometimes, when a conversation doesn’t land or a connection feels off, it’s not about disagreement – it’s a type mismatch waiting to be fixed.
So, like any good analyst would:
Check the structure. Understand the type. And only then try then join.
