Click here to watch this seminar
Presentation 1: Inferring dwelling occupancy patterns from high temporal resolution water metering data
By: Jacob van Alwon
Abstract: High temporal-resolution water consumption data – collected at the dwelling level via smart metering – could offer a non-intrusive means of inferring dwelling occupancy patterns. Non-intrusive approaches to infer dwelling occupancy (periods of time when a given dwelling is occupied or unoccupied) may have a range of benefits, dependent upon the temporal scale. Our interest is in identifying dwellings that have long-term occupancy patterns associated with tourism, such as second homes or holiday lets available for short-term rental. Specifically, we report on a 6-month internship that seeks to develop and recommend analytic tools suitable for extracting dwelling-occupancy features associated with tourist dwellings from household level-water consumption data. Those features include periods of unoccupancy, seasonal occupancy patterns and an absence of consumption trends that may be associated with habitual residential routines. We identify and present a series of approaches that offer potential in distinguishing tourist dwellings from their residential counterparts, and in inferring occupancy patterns for those tourist dwellings. We illustrate these techniques using household level data for a sample of households in Devon and Cornwall provided by our data partner, South West Water. This analysis forms part of a larger externally funded ‘WatPop’ project, which we use to highlight the potential of these analyses.
Presentation 2: Evaluation of Data Granularity Within the Non Household Water Market
By: James Hulse
Abstract: In 2017 the Non Household water market was reformed. Water utilities (Yorkshire Water, Thames etc.) were no longer selling their services direct to customers but became wholesalers, with retail services now being provided by new intermediary companies in competition to provide this service, much like the energy sector.
The industry regulator, Market Operator Services Limited (MOSL) has identified poor data (re: consumption, bills) as the key factor constraining market efficiency, hindering the ability of customers to easily switch retailers based on their needs. Furthermore, wholesalers are also concerned about good metering intelligence as this informs their water conservation works (conservation/efficiency advice; losses from pipe leaks).
We look to develop Non Household water use benchmarks in order to inform water saving targets and conservation advice, and identify pipe leakage. Specifically we look to enhance customer segmentation by considering premises consumption. Assess the benefits of granular data with regards to enhanced segmentation, detection of pipe leakage and determination of nightline use. Recommend a method to devise water saving targets and assess aggregated savings should these be met.
Presentation 3: Identifying social bottlenecks for post-Brexit agricultural policy adaptation
By: Matthew Walker
Abstract: As the UK transitions out of the EU, EU agricultural policies are being replaced with UK government policy, and DEFRA, who is the partner on this LIDA project, is keen to know how various argi-environmental schemes are adopted in England. The adoption of policies varies between individuals, for a range of different reasons. Here I investigate factors effecting the likelihood of a policy adoption in agriculture in general, and in particular focuses on identifying social interaction between farmers. To do this, I am using large spatio-temporal datasets on farm subsidies across England from the last 10 years to identify spatial and temporal patterns as well as other confounding factors. In this talk some preliminary results are given for a region of the UK, highlighting some of the influential factors in determining scheme adoption.
For more information, click here.