What is Alteryx
Alteryx was founded in 1997. It is used to manipulate geospatial data and integrate them with demographic information. SRC, the predecessor to Alteryx, was robust enough that users started manipulating non-geospatial data in the tool.
In 2006, the product was released under the Alteryx name, this time with tools that were clearly designed for manipulation of geospatial and non-geospatial analysis. In 2018, Gartner named Alteryx as a leader in the Magic Quadrant for Data Science and Machine Learning.
Alteryx is a platform that allows me to bring in different data sources. So far, I haven’t met a data source that I couldn’t apply to Alteryx. With its GUI environment, Alteryx allows me to manipulate data. I can see the results of my transformations.
Impressing the Senior Vice President
For a previous employer, monthly reporting felt like juggling bowling balls. It was always the same analysis. Each month, I got Excel spreadsheets with my data. The data had to be transformed, and then packaged into the same report. The process repeated the following month.
Since I had to bring in historical measures, the Excel spreadsheets were bogging down. I had a confusing array of spreadsheets that I had to sort through each month. It took longer each month because there was more to juggle. It wasn’t a hard task. Prior to Alteryx, I spent at least a week creating the reporting. I often got lost in the data. I forgot various transformations in the complicated process. It was frustrating.
With Alteryx, I spent less than two days a month on the whole thing. Everything was saved, so I could quickly go down a check box, go from one step to the next. It took me two to three weeks to build the workflows, but that was time well spent.
The real litmus test came when a senior vice president called me for some on-the-fly analysis. He wanted to compare two markets over a specific timeframe, looking at selected product lines. With him on the phone, I pulled up my built workflows. Filtering the data, I was able to tell him the historical pattern. To understand the drivers, I dropped a regression tool onto the canvas. In less than five minutes, I had achieved the insights that he needed.
“I expected this would take a week or so,” he said. “How did you do it?”
“Alteryx,” I replied.
I can see what I’m doing
There is an open-source, free statistical software available – called R. When I use R, I feel like I’m solving a Rubik’s Cube blind folded. I have to know how to respond to the yellow corner on the right edge. I can’t experiment. Alteryx allows me to see what I’m doing. In Alteryx, I can see my data at every stage of analysis.
Recently, a professor assigned us some analysis of survival rates from the Titanic disaster. As professors in grad school are wont to do, he gave few instructions. I needed to figure out the survival rates for first-class passengers, and then assess those rates for women and children. Did women and children have higher survival rates? I’m sure there’s a way to do this R, but I couldn’t think what it would be. I pulled the CSV file into Alteryx. I didn’t have a clue how I was going to do this. I fiddled around, trying out various methods. After watching a few more YouTube videos that the professor had linked to the syllabus, I started to see the light.
With Alteryx, I could stumble my way through the data. I didn’t have to know exactly where I was going. I could also see exactly what I was doing at each stage. Browse tools showed me the impact of my formulas and filters. Like a Rubik’s Cube, I could see the sections coming together as I moved things around.
(And to answer your question, yes, first-class made a huge difference in surviving the sinking of the Titanic. First-class passengers had a 60 percent survival rate in the aggregate. Lower classes had only a 30 percent survival rate. Below you can see the most interesting thing I found!)
Robust Statistical Tools
Alteryx has robust tools for statistical analysis. As a part-time grad student of Economics, R is my tool of choice for school. When I started using Alteryx I was skeptical. R allows me to do a wide variety of testing. Alteryx tools were prepackaged. Alteryx was the wrapper around R. Would I have enough control? The answer came last year.
Another project had me study the potential impact of child tax credits on fertility decisions. I started out doing the analysis in Alteryx. I planned on figuring out the data in Alteryx, and then translating it back into R, once I knew what I was doing. Research hit a dead-end, as research does, so I went to my professor. After applying a Dickey Fuller Test, I was sure we couldn’t do it in Alteryx – but I was wrong. I built out the Dickey Fuller Test in Alteryx and continued my merry way.
One night before class my classmates and I were discussing our results. I flipped open my laptop and showed my classmate my workflow. There were a few jaws that hit the ground, and more four-letter words that flowed.
“Wait, you don’t even have to walk us through the transformations,” they said. “I can see every transformation you did! I can see it!”
For graduate school, I turned in an Alteryx workflow. And I got an A on the project! Alteryx has enough statistical creds to do graduate-level economics.
(And no, child tax credits don’t impact fertility decisions. People have children for a wide variety of reasons. And a tax credit you get once a year – based on a complicated, hard to understand system – doesn’t factor in to the decision to have a child.)