The Leeds Institute for Data Analytics is pleased to present the next seminar in our series showcasing data analytics.
The seminar will be held in 8.43x, Worsley Building, at 15.30 on Thursday 28th March.
This seminar will consist of six ten minute presentations from our current LIDA Data Scientist interns.
SPENSER – A Synthetic Population Estimation and Scenario Projection Model
SPENSER is a synthetic population estimation and projection model which uses dynamic microsimulation. It provides a comprehensive set of tools for user customisable scenario projections. As a subset of SPENSER, UKsurvey aims to visualise the transitions that occur during the microsynthesis of population, as well as the events that trigger these transitions. The UKsurvey project also serves as proof that SPENSER is not limited to using national census data. Here, Understanding Society data is used to complement ONS census data.
Predicting and warning extreme wind response in bridges using advanced data analytics
Current early warning systems for bridges are poor in their wind speed prediction resulting in unnecessary bridge closures. This project aims to combine data from MET Office’s MIDAS system along with data from the Clifton Suspension Bridge to see if further wind prediction techniques can be coupled with structural mechanics to give more accurate long range warnings. Interpolation and machine learning techniques are used to analyse and forecast the wind behaviour around the Clifton site and predict incoming dangerous wind. Wind speed is currently well predicted at the Filton MIDAS site near the Clifton Bridge; we aim to move prediction to the bridge and integrate structural attributes.
Interpreting Spatial Flow Data – Visualization and Insight
The increasing abundance of urban big data presents opportunities for better understanding of the movement of people. However, issues surrounding visualization of large quantities of spatial flow data present problems in the presentation and elucidation of such data. This project explores the application of various visualization methodologies in order to provide an efficient workflow of spatial data interpretation. Utilising a spatial interaction model to generate flows of migration between Lower Layer Super Output Areas in the UK, visualization techniques are used to better display flow data; in methodologies that can be applied to various spatial flow contexts.
A machine learning approach to understanding the disease trajectories of atrial fibrillation
Atrial fibrillation (AF) is a major cardiovascular health problem: it is common, chronic and incurs substantial health-care expenditure as a result of stroke, sudden death, heart failure and unplanned hospitalisation. The incident prevalence of AF has increased substantially over the last decades. However, the disease trajectories of patients hospitalised with AF are not fully understood. This study will characterise the disease trajectories of patients hospitalised with AF using unsupervised machine learning techniques to classify the outcome events from re-admission. In addition, frailty models will be used to model all-cause mortality, while multi-state models will be used to investigate multiple cardiovascular and non-cardiovascular outcomes. The temporal trends of incident events and anticoagulation uptake will be quantified through trend analysis.
Synergy PRIME – Multi level Modelling, Simulation and Visualization
Travel forecasting is at the core of urban transportation planning and plays a key role in determining the need for new road capacity, transit service changes and changes in land use policies. The Synergy PRIME project is a proof of concept in integrating the data sources and models from the MISTRAL, Surf and HumanDrive projects. It aims to incorporate population growth projections and agent based modelling in a traffic microsimulation in order to assist the development of future transport systems, while using immersive technologies and visualisation to demonstrate and test the potential future design scenarios.
Extracting actionable insights from free text police data
When a crime occurs, large volumes of information relating to the event are being recorded via free text police input systems. These data contain useful information that if appropriately collated, analysed, and understood could lead to significant reductions in harm. Using natural language processing techniques this project aims to demonstrate that actionable insights (e.g. identifying unobserved trends in incident type, context, or geographical locality) can be derived from police free text data. Using topic modelling and vector space model approaches it has been possible to identify a greater level of specificity in crime events when compared to traditional crime categories.
Q&A with all speakers
Networking reception with drinks and nibbles hosted in the LIDA staff kitchen
To book your free place at this seminar please email Hayley Irving with your name, occupation and faculty/organisation.