Pioneering data-driven solutions for sustainable building management
Sustainable buildings

1. The company

SPG stands at the forefront of technology consulting, specialist recruitment and software development. Under the visionary leadership of CEO Gareth Humphreys, the company has witnessed remarkable year-on-year growth, of 90% per annum, expanding its team to approximately 70 full-time employees.  

Founded in 2019, SPG’s vision is to be the most admired technology transformation team in the industry. 

Gareth Humphreys drives the company with a data-centric approach. He believes in the transformative power of data, especially in the era of Artificial Intelligence (AI). "The old cliché is that data is the new oil, and that’s truer now more than ever. Anybody that wants to do anything with the forthcoming wave of AI needs a solid data foundation on which to develop."

He envisions SPG playing a pivotal role in solving client challenges using data and automation. "We are having those conversations with clients around problems they are trying to solve with data, as well as with automation. That is dependent on data and data quality to function". The integration of AI into SPG's services is set to revolutionise how quickly new products and services can be developed, reflecting a significant shift in the industry's operational dynamics. 

2. The problem

Recognising the urgent need for sustainable solutions in building management, SPG embarked on a mission to explore the feasibility of using automated buildings to achieve significant reductions in carbon emissions and costs while optimising facility or utility usage.  

The heart of their challenge was to demonstrate the practicality of this approach. To this end, SPG worked with the National Innovation Centre for Data aiming to run simulations to predict utility usage to required accuracy, focusing on electricity demand.

The Core Web Image
                                                             The Core, Newcastle Helix. Image courtesy of Newcastle University.

3. The goal

The project charter established two primary objectives: 

  • Short term: The focus was to analyse existing data from the Core building in Newcastle, specifically to model energy usage at a floor level. This investigation was a key component of the NICD skills project.  

  • Long term: SPG aimed to leverage insights gained from this project to develop a commercial product. This product is designed to assist building utility managers in reducing utility costs and carbon emissions, aligning with SPG's broader business goals.


"We wanted to see what we could do with that data. There are lots of solutions out there for seeing trends, but not for looking forward and predicting the usage of a space and the usage of utilities which would then drive carbon output. We wanted to validate if AI could then work out what automatic interventions could take place.” 

Gareth Humphreys, CEO, SPG

4. The results

The project was a collaborative effort between the National Innovation Centre for Data (NICD) and SPG, involving a multidisciplinary team. This team included NICD data scientists and SPG's data engineer, product manager, CEO, and business development professionals. 

 The SPG project revealed a broad spectrum of possibilities in automated building management. Key learnings emerged from the workshop process, highlighting varying stakeholder requirements and expectations.  

Some of the most significant outcomes of this project include: 

  • Mastering methodologies: Mastering various methodologies for data capture, analysis, and testing. The team gained insights into the tolerances permissible in modelling before the results became unreliable.

  • Data quality: The reinforcement of the importance of data quality, understanding which metrics are vital, and recognising the limitations of current methodologies. This understanding has helped in planning future roadmaps for product development.

  • Skills transfer: Conducting a series of ten skill sessions between NICD staff and the  SPG team. These sessions focused on processing, cleaning, wrangling, and visualising electricity demand data from room-based meters in the Core Building. The team applied various time-series modelling approaches to predict 'week-ahead' electricity demand. 

    The project successfully developed models predicting electricity use across various floors of the Core Building. 

    However, it struggled to achieve the goal of ensuring predictions were within 10% error margin over two hours for every floor. This challenge was due in part to the different activities on each floor—office areas were easier to predict compared to exhibition and maintenance spaces. The project also underscored the need to factor in holiday schedules and how full a space was to improve the accuracy of the predictions.
Gareth Humphreys headshot
                                                                               Gareth Humphreys, CEO, SPG. Image courtesy of Gareth Humphreys.

5. Data Science tools and techniques

5.1 Project Focus and Data Preparation 

The initial stage centred on developing time-series models for predicting electricity demand on each floor of the Core building. This involved gathering and preparing data at hourly or bi-hourly intervals. 

5.2 Model selection and implementation

Next, the team employed a range of modeling techniques.

This included traditional methods like:

  • Auto-ARIMA (Auto-Regressive Integrated Moving Average)
  • BATS (Box-Cox, ARIMA, Trend, Seasonal) 

    As well as modern approaches:

  • FASSTER (Forecasting with Additive Switching of Seasonality, Trend, and Exogenous Regressors)

5.3 Evaluation and analysis

Various models were tested, none distinctly outperformed the others in terms of prediction accuracy. The analysis focused on evaluating the models not just for accuracy but also for time efficiency and complexity. 

5.4 Findings and optimisation

The final stage involved identifying the most efficient models.

It was observed that models like BATS were more time-efficient compared to others like autoARIMA, particularly in the aspects of training speed and prediction generation for machine learning models. 

6. Business impact

The collaboration between SPG and NICD has significantly impacted SPG's approach to sustainable building management. The project's success has enabled SPG to apply for funding for future projects, using the NICD report to strengthen their application. The validation of their ideas by NICD's expertise has been crucial, allowing them to avoid potential pitfalls and focus on feasible solutions. 

As SPG prepares for a market launch, they anticipate significant business growth, expecting a turnover of over 1 million GBP in the next year. The project has set the stage for SPG to offer solutions that simplify cost and carbon output reduction for organisations, aligning with their vision of sustainable business practices. 


"We wanted My biggest fear about the project was that NICD would tell us our goals were impossible. Their that it was possible not only gave us the confidence to continue but saved us from pursuing unworkable solutions. The team's expertise and enthusiasm were invaluable. If anything, we would have wanted more time with them.” 

Gareth Humphreys, CEO, SPG

7. Transformative collaboration

The collaboration between SPG and NICD has been transformative, not only in achieving the technical goals but also in broadening SPG's vision for the future. The project has laid a foundation for SPG to explore innovative AI solutions and expand its capabilities, marking a significant step in their growth trajectory. As SPG moves towards launching their solution, the lessons learned, and the confidence gained from this collaboration will be pivotal in shaping their future endeavours. 

With thanks to the North of Tyne Combined Authority Digital Growth and Innovation Programme (NTCA Digital) for funding this project.

To find out more about SPG, visit their website.

You can read more of our case studies and sign up to our newsletter to keep up to date with our latest news, events and developments.



Our Discovery workshop

Our Discovery workshops enable you to explore the potential of your data and understand the benefit you could gain before committing to a full-scale project.