Consumer Data Research Centre

Geo+Data London: 2 – Round Up

Thanks to all those who made it out for our second Geo+Data London event. For those of you who were unable to attend, we have asked our speakers to provide a short summary of their presentation as well as a copy of their slides. Unfortunately we are unable to share Mastercard’s slides due to commercial sensitivities.

Justin van Dijk: What’s in a name? Storing surname geographies for both visualisation and analysis

In the United Kingdom and many other countries, surnames are regionally clustered, reflecting the spatial extent of local, socio interaction, regional cultural milieus and settlement histories. As such, using longitudinal data sets, surnames and their geographies can be used to, for instance, study population change. For the past months, we have been working on cleaning, analysing, and visualising such surname geographies. At the same time, we are working on a website on which people can visualise their own surname geographies. Ideally, we would create a single database that could serve both purposes.

We are creating surname geographies using a Kernel Density Estimate (KDE) – which is a technique to capture spatial variations in point densities. However, a KDE is calculated over a regular grid that is placed over an area of interested – the entire United Kingdom in our case. Depending on the size of the grid cells, this leads to files that are relatively large in size. With thousands of surnames for multiple years, the pre-calculation and storage of all these geographies becomes infeasible. At the same time, if we were to only extract certain information from the regular grid we are losing information. This leaves little margin for errors in our analysis

In this talk, I will introduce a method that I have used to store hundreds of thousands of grids with a minimum of information loss. By carefully deconstructing the grids during the storage process, and by on the fly reconstructing the grid for visualisation, the same database can be used for both advanced analysis and rapid display of these geographic data.

Justin van Dijk is a post-doctoral research assistant in the Geospatial Analytics and Computing Research Group in the Department of Geography at UCL.

 

Eliot Marcus: Smart Insights from geo-aggregated transaction data

Mastercard transaction data is a rich source of location intelligence insights. Eliot reviewed some of the opportunities and uses of this data, and some of the location intelligence products that Mastercard offers. He also discussed the company’s strict data privacy controls and their impact on Mastercard’s work.

Eliot Marcus is Partnerships Europe Manager at at Mastercard.

 

Geo+Data London: 2

The CDRC and the Association for Geographic Information (AGI) are pleased to announce the second Geo+Data London event. Our events aim to provide an environment for those working with geographic data to meet, share, and learn from experiences and best practices as demonstrated by industry and current academic research.

For our second event, we have brought together speakers from Mastercard and University College London (UCL) to speak on transaction data and the geography of names.

The event will be held in Room G22 of the Pearson Building at UCL on Tuesday 4th December, from 6pm to 8.30pm. See here for a map of the location.

Please register your attendance through Eventbrite. Booking is free, but it is mandatory. Seats are assigned on a first come, first served basis.

Schedule of events:

18:00: Registration
18.25: Introduction and Welcome
18.30: Eliot Marcus (Mastercard): ‘Smart Insights from geo-aggregated transaction data’
19.00: Dr Justin Van Dijk (UCL): ‘What’s in a name? Storing surname geographies for both visualisation and analysis’
19.30: Reception and networking
20.30: Event close

We look forward to welcoming you to this event.

About our speakers:

Eliot Marcus is Partnerships Europe Manager at at Mastercard.

Justin van Dijk is a post-doctoral research assistant in the Geospatial Analytics and Computing Research Group in the Department of Geography at UCL. Prior to this he obtained a Master’s degree in Human Geography and Planning from Utrecht University (the Netherlands), and received a Ph.D. degree in Transport Economics from Stellenbosch University (South Africa). His primary research interests are grouped around the analysis and visualisation of large scale spatial data, urban mobility, and geographic information systems in general.

 

Geo+Data London: 1 – Round Up

For those of you who were unable to come to our first Geo+Data London event, we have asked our speakers to provide a short summary of their presentation as well as a copy of their slides.

Balamurgan Soundararaj: Estimating real-time high street footfall from Wi-Fi probe requests

In the past decade, Wi-Fi has emerged as the most extensively used technology in providing internet access to mobile devices in public spaces, resulting in multiple Wi-Fi networks being available at almost every location in dense urban environments. Modern mobile devices with Wi-Fi capability regularly broadcast a special type of signal – probe requests – to discover these available Wi-Fi networks, and the devices switch between them seamlessly. Though this provides us with an open, passive, continuous, and wireless source of data available at any urban location which can be used to understand the number of people present in the immediate surrounding in real-time and with high granularity, we also face two major uncertainties in such data sets. First is the field of measurement, which is impossible to delineate precisely; and second is the randomisation of MAC addresses by devices to protect the privacy of the users. In this talk, I introduced a proposed methodology which solves the former by classifying reported signal strength using k-means clustering algorithm, and solves the latter by a novel graph based clustering algorithm. Thus enabling us to estimate pedestrian footfall at these locations from just the Wi-Fi probe requests with considerable accuracy and without infringing on the privacy of the users involved.

To see Bala’s beautiful presentation, please click here. You will need to use the arrow buttons on your keyboard to click through it.

Bala Soundararaj is a a PhD student in the Department of Geography at UCL. To learn more about Bala, please click here.

 

Alastair McMahon: Smart Cities and Smart Transportation

Alastair is Analytics Director at Telefonica UK. To learn more about his work, please click here.

To see a copy of Alastair’s slides, please click here.

 

See here for a news item about the event. Our next event will be on Tuesday 4th December and features speakers from Mastercard and UCL. We will release full information and registration details shortly.

 

Geo+Data London: 1

The CDRC and the Association for Geographic Information (AGI) are collaborating for a six-part series of events entitled Geo+Data London.

These events aim to provide an environment for those working with geographic data to meet, share, and learn from experiences and best practices as demonstrated by industry and current academic research.

Event 1:  WiFi technology and smart cities, 2 October 2018

Speakers:
Telefónica UK and University College London (UCL) discussed the significance of WiFi technology and smart cities.

Schedule of events:
18:00: Registration
18.25: Introduction and Welcome by Tim Marston (Carto).
18.30: Alastair McMahon (Telefónica UK): ‘Smart Cities and Smart Transportation’.
19.00: Bala Soundararaj (UCL): ‘Estimating real-time high street footfall from Wi-Fi probe requests’.
19.30: Reception and networking
20.30: Event close

About the speakers:
Alastair McMahon is Analytics Director at Telefonica UK.

Balamurgan Soundararaj is a PhD student in the Department of Geography at UCL working with CDRC footfall data produced under the SmartStreetSensor project. Prior to this he worked in a Knowledge Transfer Partnership between UCL School of Construction and Project Management and Transport for London looking at the use of network analysis in project management. His primary research interest is in the analysis and visualisation of large scale complex data relating to spatial and networked phenomena.