anyKrowd - How to turn a sea of data into data you can see.

anyKrowd is a tech startup focused on creating an all-in-one application to organise your festival, event, summer bar, nightclub… Whether you need a solution for ticketing, cashless payments, or PoS solutions… they got it covered.

One central element in their application is offering your bar/event/festival customers a personalised experience just with simple clicks in a mobile application. The use of this mobile application, of course, leads to insightful information about consumer behaviour. But tens of thousands of rows in Excel don’t say anything, so that’s where kaffie kicks in to create insight out of raw data.

The Project

This project aimed to check which easy-to-understand data dashboards could be created based on the existing data from the mobile application. In other words: delving and melting gold ores into a beautiful data necklace of insight.

Let’s make this tangible:

Most of my time (Yentl speaking), I’m working on abstract, non-tangible things. So before we head on, I would like to give a fair shot at making this project as tangible as possible.

We start with data, in this case: Excel files full of rows with different transactions that happen in a summer bar:

• Someone orders a drink.
• Someone top-ups their application with money.
• Someone asks for a refund.
• Someone wants to pay with a gift card.
• …

From these raw data transactions, I start by analysing the Excel files to identify which data groups belong to each other. Also, this is when I clean up and correct faulty data. Once we’ve passed the analysis & cleaning phase, we can head off to make clear data dashboards that show big volumes of information in one glimpse.

In short, this process looks more or less like the picture below:

How this process turned out for anyKrowd

Based on the provided data, we made several dashboards in Tableau.

The first dashboard, for a bar, gave an overview of sales per day, sales distribution per weekday, product performance & the top spenders of the bar. It looked somewhat like this, but then with real numbers:

‍

With this dashboard, we could find out the following:

• The effect of events & promotion campaigns on turnover.
• The most valuable clients & what they drink each day.
• The beer & wine kings & queens.
• Adjust staff planning based on weekday performance.

In another dashboard, for a one-time event in a big venue, we worked much more on gaining insights into which staff members handled the most orders in which zone to the granularity of 1-minute intervals:

With this dashboard, we could find out the following:

• Turnover & order distribution during the event.
• Which zone (VIP, bar 1, bar 2, food…) had the most orders & turnover.
• Anticipate bottlenecks & staffing issues based on actual data.
• Identify high-performing staff members.