Skip to main content


Digital city festival


01 Apr 2021

The power of data science and visualisation for creative marketing campaigns

Luigi Raw, Senior Data Engineer, Enjoy Digital
The power of data science and visualisation for creative marketing campaigns
Luigi Raw, Enjoy Digital's Senior Data Engineer, takes you through the ways to use data science intelligently and efficiently to achieve success in your creative marketing campaigns.

No doubt you’ll have heard of data visualisation, aggregation, cleansing, engineering, machine learning, artificial intelligence and many similar buzzwords. But have you got a grasp of them?

When people hear ‘data science’ they immediately think it's complicated and expensive. But this is a misconception. There are many ways data science can be used on creative projects that will offer a significant ROI.

Let's take a look at a couple of ways you can use data science to support a successful campaign.

Before getting into the thick of data science, it's important to understand two key things: what you have, and what you need.  Why these two? Let's look at them.

What you need

First, it’s important to understand what you're hoping to gain from a campaign. This could be as simple as being able to better understand your key audience groups, or being able to predict and forecast sales and revenue. 

By identifying what you hope to achieve, you can then take the step to…

What you have

To achieve your goal, you need to review what data you already have that can feed into the project - and outline anything you feel is missing.

This could be anything from business and marketing data, through to external datasets such as typical house prices or weather. 



1. A social media marketing campaign

Imagine you’re running a social media marketing campaign, and want to understand the ROI for leads generated through these channels. 

Upon review, you identify that you’re not effectively tracking the leads through the channels, creating a break in the user journey.

What you'd need to do is review and implement additional tracking to bridge those gaps - allowing you to calculate ROI based on spend, leads generated and average lead value. 

2. Prediction of regional performance in real estate

Here's a more comprehensive example - you’re an estate agent looking to identify key areas to target with sales and marketing activity across the UK, with a view to opening a branch.

To achieve this, you know you have access to your portfolio of houses sold by location from which you can understand average prices, listing-to-sale times, frequency, and so on. 

But it would be great to understand other external factors such as average salary, proximity to local amenities and locality to transport hubs. This is where external datasets can help - you can use several open-data platforms or paid sources to grab this data. 


Let’s take a step back. Now we understand what we have and what we want, the next part is fun. 

Data Engineering 

Data Engineering is the foundation for all data science projects and a useful tool for creative campaigns. 

It can be as straightforward as connecting to a dataset, and mapping out dimensions and measurements, through to ETL (extract, transform, load) processes and workflows.

The transform stage of this process - involving cleaning, standardising and streamlining the data - is key, as it allows the data to be processed in a way that is useful. This includes removing elements that are not required, adding in new elements, standardising elements throughout, and creating unions and joins. 

Let say, for example, you have a datasheet of users which includes a column for date joined and one for postcode. In the data, you notice that the some columns are incorrectly formatted:

To ensure this data can be translated into a successful campaign, this is a good place to clean the data and ensure consistency. For the date column, you just need to align all the dates to the same format. The postcode, however, may need some further consideration. 

To use it for granular analysis, where the full postcode is required, you'd first need to make a consistent format, let's say ‘LS2 8PY’. This would work for most of the entries, but what about the incomplete ones? Instead, we may have to pull a list of all UK postcodes and the street name, and create a lookup. That means that where the postcode is incomplete in the user data, it can find the street name in the external data and return the full postcode.

Data is the foundation of a successful campaign

Ultimately, “The output is only as good as the input”. 

If the data is inconsistent and incomplete, you’ll struggle to further use and analyse that data down the line. Data is generally the foundation of any successful campaign - creative or otherwise. Only with careful consideration of all these stages will your ROI reach its full potential.


Digital City Festival takes place from April 12th - 23rd as a truly digital experience. Will you be part of it?