I was training a group of new Alteryx users the other day and one of the delegates asked a question that I thought it may be worth blogging about. The question was something along these lines:
“What skills do I need to learn in order to be as successful as you are in Alteryx? Is it R, SQL or something else?”
This is an interesting question, and it’s something I’ve thought about a lot over the years as I’ve both trained Alteryx analysts, and also recruited them, how do you improve your skills as an Alteryx analyst?
Well let’s start by quantifying what we mean when we talk about a “successful” Alteryx analyst, definition is always important before discussing the detail. What kind of skills are important. We could measure success by several different metrics:
- Speed – i.e how quickly you can solve a given problem
- Complexity – of the problems you can solve
- Performance – the accuracy of the models you build using the predictive tools, or perhaps the time taken to run a large data process
- Depth – of knowledge of all Alteryx functions and server capabilities
- Honours received e.g. a Top 3 placing in the Alteryx Grand Prix
For the purposes of this post I’m going to focus on the top two, and so focus on how we can get quicker at solving complex problems with Alteryx.
Maybe this will sound familiar: Imagine you’ve been using the tool for a few weeks or so and you’ve built lots of useful modules, mostly using the tools in the “Favourites” category. Things like data blending, filtering, pivoting, cleansing, aggregating are now second nature to you – in fact you nailed those in your first week (remember 86% of people see business value from Alteryx in the first week of their install). However you still come across problems you need help with that you feel should be doable if you knew the right methods.
So where do you start? Well the bad news is that there isn’t a technical skill, like SQL or R, that you can pick up and learn that will make Alteryx easier, aside from using the tool more. The good news is that you might simply change the way you are approaching the problem. Stop thinking that there must be an answer in the tools and start thinking about the problem itself – break it down into some simple steps:
1. What does the end solution need to look like? How does the data need to be structured?
This is the most fundamental part, and it sounds obvious, but until you can’t start a journey until you know where you are going. Take a moment to think about the final structure.
2. What does the data look like now?
Again this seems obvious, but use a Browse tool and look at the data you are using.
3. Break the problem down into steps.
Is there any order to what you need to do? What can be done to break the problem down. Ensure you go back to 1 and 2 above for each step.
4. Put away Alteryx.
Whoooaaaa, back up. Yes, I did just say that. Put away Alteryx. Close it, hide it, do what you need to put it out of your mind. Alteryx is an enabler, but it’s not going to help you solve the problem at this stage.
5. Now how would you solve the problem (without Alteryx), verbalise the steps – say it out loud or write it down or think it through.
Give yourself, and Alteryx, something to work with.
6. Now open Alteryx.
Phew, that’s better – cold sweats while it was closed? You’re getting addicted.
7. Find the tools you need and implement your solution.
Alteryx tools “reward the way you think” by using exactly the same words and descriptions you just used in Step 5. If you can’t find a tool work use different verbs and you’ll find the tool you need.
Sounds like I’m talking rubbish, right? Maybe, but it really does work. Let’s work through and keep this real with a real life example I received in a tweet
So lets work through our steps:
1. End point: We want monthly values in one column and averages of a value in the other. Key question: How do we want the month values formatted “09-2013″, “Sept 2013″ or “01-09-2013″ (we’re in the UK). In this case we want “01-09-2013″.
2. Start Point: we have a csv. It has daily dates and a value. A key question here is: are the dates formatted as dates? No, they’re in a csv file. Tip: don’t forget to consider the datatypes you’re working with.
3. Rough Steps: Import the data -> Format the dates -> Find the Month value -> Average the Value per Month -> Output the Data
4. Forget Alteryx: “What’s Alteryx?” Okay that worked.
5. An English Solution: I want to Import the data from a csv file -> I create a new column to format the date as a date datatype -> I want to Trim the Date to Month level -> I want to Aggregate the data on the new column and average the Sales -> Output the data
6. Open Alteryx
“I want to Import the data from a csv file” So I’m going to choose the Import tool.
“I create a formula to format the date as a date datatype” Hmm I don’t have a create tool, but I do have a Formula Tool. I look down the functions and find DateTimeFormat.
“I want to Trim the Date to Month level” Okay well let’s stay with the formula, there’s a DateTimeTrim function so I will use that with the Month parameter.
“I want to Aggregate the data on the new column and average the Sales” Hmm no aggregate tool. Can I reword? Summarise is another word for aggregate so let’s use that. I can Group by the new column and select the average of the Sales
“Output the data” So let’s use the Output tool.
See how well the words in red line up between the English description and the implementation in Alteryx? Alteryx is built to reward us for how we think, and so makes it very easy to implement those steps in a logical manner using tools named sensibly.
I hope this helps change the way you think about building solutions in Alteryx in future.
The post 7 steps to improving your Alteryx – reward the way you think appeared first on The Information Lab.