Four reasons I switched over 90% of my work from Python scripting to Alteryx

DucksInARow 400px

Alteryx helps you get your data ducks in a row.

In the middle of the Highly Pathogenic Avian Influenza outbreak during 2015, I found myself getting completely overwhelmed.

Many data processes were needing to be automated, epidemiological analyses needed to be completed, and I needed to remember everything I was doing for writing reports later.

I ended up writing a lot of Python language scripting code, and that code quickly became a headache. I needed to maintain, update, and run the code—and often I was writing so much code I’d forget which project I was working on—while working on GIS and cartography products, among other tasks.

I needed a better way to handle the data, and that better way come through Alteryx software.

What is Alteryx?

Alteryx is software that lets a user build data transformation, blending, and analytic workflows visually. That means instead of writing lines and lines of code, you can simply drag tools onto a dashboard and create a workflow that looks a lot like a flowchart.

There are four primary benefits that Alteryx brings to a data person:

Benefit 1: Save time
Benefit 2: Reduce mistakes
Benefit 3: Simplify collaboration
Benefit 4: Unite tasks

Let’s dive in to each benefit, starting with saving time.

Benefit 1: Save time

In my first computer programming classes, we’d have to create a flowchart of how the program worked before writing a single line of code. It was a great learning experience. 

Creating the flowchart made us think about the logic of how things would flow before worrying about the details of the code. But then came the coding part, and that’s when you’d end up trying to figure out all those details, squash bugs, test output, etc., etc. 

Dealing with data is no different in it’s need of logical progression. 

When involved in a project that requires data transformation, blending multiple data sets, and creating analytical output, you currently have to look at the individual data sets, sketch out a plan on how to accomplish the task, then convert that plan into code. While each step takes time, converting the plan into code is by far the most time consuming.

Alteryx removes the last part of converting the plan into code. Alteryx lets you build workflows visually, and you can take the plan directly into the software without dealing with code. 

Especially because dealing with code introduces mistakes, which takes us to benefit 2.

Benefit 2: Reduce mistakes

Hand coding data transformation, blending, and analytics can introduce many mistakes:  variable names must be typed correctly, data joins must be coded correctly, and variable types must be correct. Any mistakes in the coding can introduce errors in the process that create incorrect results.

Alteryx makes many of these mistakes impossible … 

Alteryx provides a list of variables to choose from, so you can’t mistype a variable name. Data merges are performed by selecting from available data, and the data types need to agree. To merge data, a quick glance at the data source tells you exactly what the variable types are so that you can convert them before ever running a single process. No mismatched data types will happen.

And when you want to check your logic, Alteryx provides a quick and easy tool to look at the data set at any point along the workflow. You can quickly check to make sure the workflow is doing what you planned at any point along the process.

That takes us to the third benefit of never having to explain code again.

Benefit 3: Simplify collaboration

When collaborating with analysts and others, it can be difficult to describe exactly what you are doing or did. Handing collaborators pages and pages of code does little to facilitate that collaboration. Instead, what often happens is you create a flowchart to explain to others what the code is doing.

Alteryx builds the explanatory flow chart from the beginning of data processing so you never have to draw the flowchart out again.

Alteryx creates visual workflows, which you can pull up and use quickly explain the process to anybody. There’s no need to convert code to a flow chart. Even better, the workflow flowchart becomes your documentation. You can save or print the flowchart out and put it into the project documentation for later reference.

But for me, maybe the biggest benefit of Alteryx has been uniting many varied tasks all into one workflow process.

Benefit 4: Unite tasks

One of the great things about Python has been its versatility. It can handle many data file formats and do many, many different things with the data. That’s an important point because often data processing involves many different data tasks, such as:

  • Importing data from a data source a software understands.
  • Cleaning data to standardize variables, filter out incorrect data points, and conditionally fix data points.
  • Converting data from one file format to another, such as spatial shapefiles and Microsoft Excel.
  • Converting data with merging, filtering, and calculations.
  • Spatial analyses, such as buffering, overlaying, and calculating distance or direction.
  • Statistical analysis, such as summarizing, calculating correlations, and calculating frequencies.
  • Outputting data into a certain file format.

If you try to accomplish these tasks without scripting, these tasks can quickly become very time consuming and labor intensive … especially changing data between file formats so that one software package can use the output of another software package. 

Alteryx handles much of that time-consuming conversion process by performing data processing in a single environment. The data filtering, merging, and calculating functions within in Alteryx allow it to handle most processing traditionally done inside a database. 

The spatial analytic functions within Alteryx mean there is no need to export data out, import data into a GIS, convert the data to a spatial file format, and then perform the analysis. The analysis is handled inside the workflow.

Plus, there is no need to export data out and import the it into a statistical package to perform analysis because Alteryx has many statistical functions. The analysis can be performed right inside the workflow.

And Alteryx has input/output covered, too. It is able to read and write many different file formats, so it is a simple matter of importing and exporting to the data file format that makes the most sense for any work that needs to be performed outside the workflow. You can output a spatial file to create a map or output the data into an Excel spreadsheet for sending to a customer.

Even better, if you need a report as output, Alteryx can create well-designed reports as the output.

Alteryx can do more than 90% of my work tasks that I used to have to do in Python

That doesn’t mean I’ve abandoned Python. It simply means that I have a tool that has improved my productivity, left me with a lot less code to write and maintain, and lets me accomplish tasks that may have needed several different software packages to accomplish inside a single environment.

But what about cost?

Yes, compared to free Python, Alteryx is very expensive. But the time savings in less code, fewer mistakes, ability to more easily document and discuss processes with collaborators, and not having to bounce between multiple software packages more than makes up for the cost. 

In the end, it’s how you want to spend your time. I’d rather spend less time on writing code and more time collaborating to get work accomplished. It’s more valuable work to me, and more valuable work to my employer.

And in the end, it’s getting the work done that matters

Having Alteryx from the beginning of the Highly Pathogenic Avian Influenza outbreak would have made my life a whole lot easier. I would have had clear flowchart processes to explain things, I wouldn’t have had a boatload of code to try to keep straight, and I wouldn’t have painted myself in the corner as the single point of failure for many projects.

If you’re scripting in Python, give Alteryx a try and see what you think.

© Philip Riggs 2014-2016  Privacy Policy