Data Analysis of Meter Readings to Spot Wastage or Anomalies

data analysis
Bobby McHale

By Bobby McHale Environmental Data Analyst at Central Manchester University Hospitals NHS Foundation Trust

Metering is the first and most important part of energy management and is the backbone to an effective energy management campaign. The Carbon Trust states that for many organisations, energy is the second largest expenditure following salaries.

A study completed in 2016 stated that buildings are responsible for 40% of global energy usage, thus accounting for 30% of total carbon emissions. This is an area we can all target to cut down on our usage and help in this global crisis.

First, how do you begin? Well, imagine it is post-Christmas; you look at yourself in the mirror and think there is a bit more on the hips than there was a few weeks ago, so you would go to step on the scales. This is pretty much where we are; you cannot manage what you do not measure if you cannot measure you cannot improve. 

I work in an NHS Trust with an energy spend of £14 million per year and despite ongoing efficiency programmes, this is only increasing due to the price of energy, government legislation and the growing intensity of activity required in a hospital. In the past, organisations would have a basic level of metering just so utility companies would get a vague idea of what to bill. Now, however, large organisations are investing more in complex metering systems as the ‘money men’ are noticing this financial benefit more and more. Getting a good metering system in place is essential however, where to begin once you start receiving the data?

Bill validation

The first form of analysis to be done on meter readings would be bill validation. Many organisations are paying too much for utilities due to inaccurate estimates and errors. A simple analysis of monthly meter readings ensures that companies and organisations will only pay for what they use. This is a fairly simple piece of analysis that can be done using Microsoft Excel. Setting up a simple spreadsheet with columns on various cost data will allow you to see what you are using per month and what it costs, compared with your bills. It is important to include columns for your meter reads, suppliers reads, rates, charges (daily charge, standing etc.), suppliers bill and your estimated bill. Having this system in place will allow you to update monthly with reads and data and apply a quick comparison when the bills come in. At the Trust, we have someone monitoring this data and have included rolling bar charts plotting consumption over time to make it easy to see any anomalies or wastage. This is also a useful way to begin benchmarking.


At the heart of it, benchmarking is just a comparison. Think about it, your laundrette business is going well, you have good custom but your utility bills are costing a fortune. You cannot afford to close early one day, you cannot afford the marketing. How is it that somebody with the same business across town is doing similarly well business-wise, same size building, the same number of machines but his profits are much more? After a little comparison of your meter reads you notice he is using 2/3 the amount of electricity. This simple bit of benchmarking has shown a weakness in your business and an area you can improve on, it shows somewhere you are wasting energy and thus losing money. Being able to compare a building’s energy use against similar buildings allows you to see if wastage is high allowing you to formulate a plan on how to alter this. 


Benchmarking also needs to be done in the form of normalisation. Taking the meter read and consumption alone is irrelevant without adding extra information such as floor area, number of visitors, weather. One of the most prominent methods I use is normalising gas usage against temperature in the form of heated degree days (HDD). 

Plotting gas consumption against the number of HDDs as a piece of regression analysis predicts how much gas should have been used based on the number of HDDs if there is a strong correlation. If I experience a correlation of R2 = 0.95 for the largest hospital building when used in a ‘y=mx+c’ formula, this shows me whether consumption is at a normal level or an anomaly and is essential in providing depth to the data. 

Our intensity of activity and size also changes over time so it is important to analyse the effect of this. The Trust has increased in size by 22% and 10% in terms of occupied floor space and number of patient contacts respectively between 2013 and 2018, and electricity rates have increased by 13% between April 17 and April 18. It would be impossible to reduce energy and costs with these external conditions; however, our normalised data of largely gas and electricity usage (Scope 1 + 2 carbon emissions) shows a 13% reduction in tCO2–  per patient contact. 

Being able to analyse and report this data with context allows us to explain how we are improving but also identify any areas that need improving.

Half-hourly data

Benchmarking analysis and bill validation have been useful tools in saving organisations money and looking towards energy savings, however, new technology allows for much more in-depth analysis of meter readings. Metering technology has improved greatly, particularly since the emergence of half-hourly and smart meters. By law any organisation with a max demand of over 100kW should have half hourly data, however, it would be useful to have it in any case. 

Analysis of Half Hourly data allows for energy profiles to be built up and anomalies and wastage to be spotted more easily. These profiles will stay constant, however spotting unexpected peaks will point to anomalies and wastage.  Largely, your energy profile will follow an occupancy profile of a building in that the people inside will be using lights, heating and computers etc. A large amount of energy wastage and anomalies can be spotted just by doing this comparison. 

For example, one of the hospitals is just a 9-5 weekday hospital.  A quick analysis of the data showed that we were still using a reasonable amount of extra electricity on the weekend. Whilst it would be expected that some electricity usage would be required for essentials, it was clear from the profile that the HVAC system was operating for the same hours as through the week. A simple change through the BMS has resulted in reduced wastage across 2 days out of 7 (104 days per year!) providing substantial savings. 

This wastage is spotted just through simple analysis and by looking at time profiles, however, should you have the resources to do more thorough analysis you can also look to see whether the overall usage fits with the type of area. For example, knowing that at a certain time the only equipment being used in an area would be computers, TVs, small power equipment, etc.  you would expect relatively low usage. Knowing there is some wastage or incorrect values allows for further investigation. This is where sub-metering and perhaps device level monitoring would be useful.


Having an extensive sub-metering system allows for much more analysis and more wastage to be spotted. Sub-metering smaller areas of the fiscal meter and building up a ‘tree’ as such will show the areas consuming large amounts and the areas benchmarked well or poorly to allow for an energy management campaign which targets areas of higher consumption. The ability to monitor and analyse meter readings for smaller areas will show outliers, improve on billing with any recharges and even highlight maintenance issues. Should the electricity bill be unusually high, analysis of the sub-metered data will allow you to pinpoint the value of the anomaly and where it is arising from. 

It is not only the building data that can be analysed through sub-metering but also equipment level metering which is available through new technology which utilises CTs to measure consumption at device level. This intrinsic level of analysis allows for a holistic view on how everything is running. Monitoring the efficiencies of your high consuming kit can instantly show you where energy is being wasted (through time left on the power it is being run at), equipment that needs repairing or just equipment not being used!  Monitoring the consumption of pieces of equipment allows you to strive towards the holy grail of no wastage and thus reduced energy bills. 

This sub-metering is more likely to be automatic and provide automatic targeting and reporting. The automated reports allow for easier analysis of your data against targets and show any progress. Showing progress is vital as a key driver in driving energy efficiency behaviour change. It is estimated that in our NHS Trust we can save a further 10% of energy through staff behaviour. Reporting this data as a monetary value back to staff is an important piece of analysis as it allows staff to see progress and understand the importance of what they are doing, therefore minimising any wastage. We have numerous programmes running through the Trust targeting behaviour change with feedback stating the importance of seeing the data to inspire staff improvement. 

This technology is also fantastic for any measurement and verification (M&V) purposes.  To be able to implement more and more efficiency projects, a good level of M&V has to be done. Data analysis of meter readings allows for effective M&V showing stakeholders how well an efficiency project has worked.  Analysing the data before and after can show a step down in consumption (if the project worked!) and thus money off the energy bill. Being able to analyse these projects through sub-metering is more effective, however, if the equipment is individually monitored it is possible to put an exact number on it. I have completed a number of lighting efficiency projects where it is now possible to see how much exactly the Trust has saved, giving accurate ROIs and the assurance to push for more projects. 

To summarise, substantial analysis of meter readings is imperative to an effective energy management campaign. To be able to manage, you need to measure. Knowing what you use as a whole is great, however, the ability to dig deep and analyse at building or equipment level allows you to identify wastage and inefficiencies in your operation. The rise of new technology and the prevalence of half hourly data make it simple to compare energy profiles and expected usage thus minimising wastage. It is also important to measure how much energy you have saved; analysis for M&V purposes is imperative to keep getting projects approved and keep up the good work! 

Author’s profile:

Bobby is a Physics graduate who has worked as a data analyst in the Energy & Sustainability team at Manchester University NHS Foundation Trust since 2016. He hopes to develop a strong career in energy management with aspirations to specialise in renewable energy.

Get in touch with us

Any queries? Please send the organiser an email and we'll get back to you, asap.


Log in with your credentials

Forgot your details?