This week kicked off with two days of learning about Alteryx Macros with Phil Lowe. In short, Macros are a workflow packaged into a single tool. The purpose of this can be to provide interactivity to a workflow, make a complex process simple to use, or allow a common series of steps to be reused. There are four types of Alteryx Macros:
- Location optimiser.
On day 1 we focussed on standard macros (as simple as just combining several steps into a single tool, and batch macros. Batch macros allow you to take your data as one input (e.g. details of orders of luxury soap), take a series of batch values as a second input (e.g. soap box sizes), and then run your data through the workflow once for each member of the batch (e.g. for each size of soap box, calculate how many boxes would be required for each order).
On Tuesday we kept on learning about Macros, and encountered iterative macros. Iterative macros essentially take your data input and then iterative in loops until a condition is met. For example an iterative macro could be used to calculate interest gained on an initial investment over a given number of months.
We also experimented for the first time with web scraping. After signing up for the NY Times API we all built a macro that allowed us to enter in a search term and a maximum number of pages to search, and then returned the titles and taglines for all articles on the NY Times website that met the criteria.
Later on in this day we also spent some time discussing how client projects work ahead of our first project next week. This was valuable and covered the key events, roles, presentation tips, and more.
On day 3 we dipped back into Tableau and this time officially entered Advanced territory. This took the form of table calculations and more complex LOD calculations. As expected, these topics were just an absolute joy to undertake and not at all mindboggling. In particular, we had the need to take care around the order of filter operations as for LODs the FIXED operator filters at a different level to the INCLUDE & EXCLUDE operators.
My pneumonic is “ED’S Chilli Fries DIE MeTriC. No, it doesn’t make sense but it does help me remember the order as “Extract – DataSource – FIXED – Dimension – INCLUDE/EXCLUDE – Measure – Table Calcs”.
Day 4 breached into new Alteryx territory as we were introduced to Machine Learning and Statistical Modelling with Benedetta. We learnt that there are two types of machine learning: supervised learning and unsupervised learning. Supervised learning gets broken down into regression (used to measure actual values) and classification (used to score or group values).
We also learnt some of the basic modelling tools, how to compare models, and how to score them. I think the true intricacies of how the predictive tools work in Alteryx is still a little beyond me but I’m sure with bit more practice and some more lessons I’ll get there.
This week’s project went down a different route from the past few weeks and instead was based around Alteryx! The task was to recreate as many table calculations from Tableau as Alteryx Macros. At first this project didn’t seem too complex until you remember that table calculations are based entirely on what fields are included in the view, which means having to configure in the options for grouping in Alteryx. After spending slightly too long trying to find a way to allow a user to group by as many fields as they wanted I instead set an achievable goal of allowing users to group by two fields. I also focussed entirely on the running table calculations, which was lucky because the five running table calculations (SUM, COUNT, AVG, MIN, MAX) took up my entire time!