NECSI 2016 summer school course

Between June, 10th and 18th, 2016, I travelled to Boston in USA, for attending the NECSI (New England Complex Science Institute) 2016 summer school course, which was about general theory and models of complex science. The course was held in MIT tang center, where there was roughly thirty attendies from diverse backgrounds (economics, neurosciences, physics, biology, civil engineering, etc.). Besides those from USA, many of us came from different continents (New Zealand, Malaysia, UK, Japan, etc.), so it was rather cool to meet Francesco Castellani from same university of mine! Francesco was a very interesting Italian PhD student, who was studying urban infrastructure vulnerability and resilience against external hazards.

The main course lasted for four days, and was chaired by Assistant Professor Hiroki Sayama, Professor Dan Braha (Fig 1). We were introduced to the basic concepts in complex systems and some specific and most commonly used models for dealing with them, such as cellular automata, networks and agent based models.

Fig 1. Professor Dan Bara and students were discussing about network metrics.

Fig 1. Professor Dan Bara and students were discussing about network metrics.

Then Joseph Norman and Alfredo Morales joined them to show us some real life case studies (data mining and machine learning) using such models. One interesting moment during the course was about the social network analysis. Alfredo demonstrated how to listen to social media, such as twitter and to extract the user information to construct a social network in Python. All attendies were asked to register a twitter account and to tweet for any information he/she wanted using a hashtag. In the end, we got a cool social network (tweeting network) which was ready for visualization and network-related analysis (Fig 2).

Fig 2. Tweeting social network from attendances, visualization in Gephi. Each node is a registered user and edges are based on tweets where other users are mentioned.

Fig 2. Tweeting social network from attendances, visualization in Gephi. Each node is a registered user and edges are based on tweets where other users are mentioned.

During the course, we were also required to form teams to do a project using the tools and techniques we had learned. I teamed with Cristi, Francesco, Alex and John (Fig 3) to finish an interesting project about interdependent urban infrastructure networks and the cascading failures from them to economic sectors. Francesco and I were in the urban infrastructure network research field, while Cristi, Alex and John were economic data users in their daily work. That’s why we had the idea for this project. We used a very simple model of interdependent infrastructure networks (power, water supply and telecom) and simulated different failure scenarios to model the cascading failure and its impact on the economic sectors using some economic models.

Fig 3. Cool team members, from left to right: Cristi (Malaysia), Francesco, Alex (USA), me, John (Canada).

Fig 3. Cool team members, from left to right: Cristi (Malaysia), Francesco, Alex (USA), me, John (Canada).

The course was short but intensive, and it was interesting to learn and team up with people from different backgrounds and with different experiences. I wish all the best for my team members in the future work and it was a pleasure to meet them.

Besides academic work

The course started on Monday, June 13th, while I arrived at Boston on Friday, June 10th, which means I was to lucky to have two free days for some adventures and sight-seeing in the city before the course!

One thing I would never forget was the extremely pleasant weather in Boston. Being at almost the same latitude of Spain created warm and sunny days during my stay there. Since it is at coastal area, there was always breeze, which is really nice in the summer. The city was highly connected by its underground train system, which made my travelling plan even easier. And let’s start!

Quincy market is always the most welcome place for the first time travelers. It is a Roman style architectures (Fig 4) where there are lots of stores for fancy food and souvenirs. You can always find a place to have meals of tasty lobsters at a decent price!

Fig 4. Quincy Market, in the center of Boston.

Fig 4. Quincy Market, in the center of Boston.

Another thing that makes Boston a little different from other US cities is its colonial history. This was the place where the British adventurers landed in the American continent, which was why this region was called “New England”. There is a very special road in the city, which is call “Trial of freedom”, it is a long route (around 5km) made of red bricks on the road (Fig 5). This trail of freedom leads to the coast, and following this road, you will see many historical architectures, which represents the colonial history of Boston (Fig 6). It was a long road, and I was lucky enough to have the time to reach the sea side following this road! (Fig 7).

Fig 5. The sign of “Trail of freedom” on the road.

Fig 5. The sign of “Trail of freedom” on the road.

Fig 6. Walking to the sea, following the trail of freedom.

Fig 6. Walking to the sea, following the trail of freedom.

Fig 7. And we reached the destination, the sea side, always nice weather!

Fig 7. And we reached the destination, the sea side, always nice weather!

Despite the exciting sightseeing, I still felt a little regret for not having enough time to explore more about city. While it is a dynamic young city (the same age of USA), it was a city mixed with modern and historical culture and architectures, and a really nice place to live. I wish I can visit this beautiful city again sometime in the future.

Thanks for reading

Qingyuan Ji

ESRI UK User Conference Tuesday 17 May

This year’s ESRI Annual Conference took place once more in the QEII Conference Centre in the heart of Westminster.  It was good to meet up with several Newcastle graduates, each making their mark on the GI industry.  The major ‘take-away message’ from this year’s conference was the increasing ease by which the traditionally ‘clunky’ ArcGIS desktop can be left behind in favour of developing using ArcGIS Pro and ArcGIS Online, products with regular and frequent update cycles and increasingly comprehensive functionality.

From the perspective of ESRI, David Rix (1976 graduate) and Dominic Stubbins (1995) were amongst the welcoming corporate hosts, contributing to an efficiently-run and informative event, well worth the trip from the north!  Nathan Ward (1993) was manning the Leica stall, one of several associated partners exhibiting in the building.  Laura Hanson (2003 graduate), now with Arup, and Rachel Oldroyd (2011 graduate), now with the University of Leeds, were ex-staff members spotted, along with Meredith Williams (staff 2000-2010).

Joining Dominic from the 1995 cohort was Ian Moodie, whilst from even further even back, Rob Knight and Helen Durham (both 1985) are both thriving.  Clive Surman-Wells (1988) and Andy Hopkins (1996) made their presence known also, along with some 21st century graduates – Bruce Ford (2009), Claire Watkins (2009), Victoria Short and Matt Bowerbank (2010), Matthew Bierton (2014) and Jonathan Hallam (2015).

Next week, the re-unions will continue in London at the GeoBusiness event.  There is a formal Newcastle alumni meeting on Tuesday 24 May at 15:00 hrs.  Come along to the Business Design Centre in Islington, if you are in town …

GIS students present research projects Friday 13 May

This week saw the annual ‘Final Year Conference’, the showcase for the Stage 3 Geomatics students to present their dissertation research in a formal venue to their fellow students and to the staff and postgraduates in the School.  A total of 31 Newcastle Geomatics students delivered talks in two parallel tracks with themed session titles including ‘Built Environment’, Spatial Analysis’ and ‘Geovisualisation’.  The 12 final-year Geographic Information Science students’ topics were as follows:

Nada Alabdulwahed used an impressive range of satellite remote sensing imagery from the past 20 years to track the urban development of the city of Al Khobar in eastern Saudi Arabia.  Standard image analysis and classification techniques along with ground truthing and integration with other data on population and land reclamation allowed for accurate quantification of urban growth.

Ed Drummond-Hay researched house price changes across England and Wales, using millions of records from Land Registry, ONS and commercial sources.  His hypothesis was that house price increases in London ‘rippled’ out from the capital to affect the rest of the country after a time lag.  The dynamic and animated representation of such patterns was an important part of the output from Ed’s research.

Emmanuel Egunyu aimed to create an online website resource to promote tourism in his native Uganda.  Using OpenStreetMap data and open source technologies with PostGIS and Geoserver, he reported on the creation of a map-based information portal for those planning to visit national parks and view wildlife in this country.

Alexia Fenn’s study concentrated on the validity and reliability of observations from GPS receivers worn by athletes.  A variation of sport/athletics course shapes, course distances, and exercise/movement intensities were examined.  Walking, jogging, and sprinting over various distances and in a variety of course configurations were examined, with GPS accuracies determined for each exercise regime.

Will Franklin developed a system to estimate the safety of junctions on the road network in Newcastle.  By calculating a Junction Risk Factor, based on a number of parameters related to road conditions and traffic volumes, he was able to develop a predictive model to determine the optimum location for investment in road network improvements.

Jo Gallagher’s interest in spatially-enabled Twitter data led to a study of the geolocation of football clubs’ fan base.  He found that the ‘bigger’ the club, the more dispersed the distribution of fans – backing up the popular theory that the mean location of a Manchester United fan is some distance from Old Trafford.

Patryck Janicki’s research aimed to develop a geodatabase schema able to hold both CityGML data at Level of Detail 1 and main Industry Foundation Classes (IFC) classes.  The ability of the schema to abide by standards, allow for SQL queries, hold geometry effectively, was to be supplemented by an effective visualisation flowline through export to QGIS.  The outcomes of this project have impact on the integration of BIM with GIS.

This was also the focus of the research of Anthony Morley, who examined current technologies for importing a BIM model into a geospatial environment and assessing interoperability.  This involved using a range of software and data formats, including ArcMap, ArcScene, ESRI City Engine, REVIT, AutoCAD and FME.  The geometric and semantic integrity of the model, and the needs of users were considered.

Amber Kaye-Kenyon chose the analytical capabilities of ArcMap, ArcScene and City Engine (including Buffer, Viewshed and Line of Sight (LOS)) to look at the modelling of an urban sensor network.  In the context of the instrumented city concepts being developed in Newcastle upon Tyne, particularly around the university’s new Science Central campus, ease of use, visualisation and effective monitoring of the sensor network were derived.

Ben Nicholls undertook a very detailed study of some of the rich cartographic generalisation tools in ArcGIS.  Applying these to different types of feature (railways, contours, buildings etc. as represented on large scale Ordnance Survey data), and in different locations (e.g. central urban zones, rural regions), Ben presented an impressive assessment of the data pre-processing, algorithm effectiveness and presentation quality of such procedures.

Josh Watson’s theme was a big data set dealing with cardiac arrests outside hospital.  The spatial and temporal distribution of these events was determined, as a precursor to understanding the nature of the phenomenon and to try to raise low survival rates in the North East.  Clusters of cases were apparent and these were linked to population distribution, to ambulance stations and response times, and to times of day, season and year.

Imogen Weight made a comparative study of several web mapping APIs, assessing data needs, scope of functionality, usability, and final appearance of maps created in Leaflet, Mapbox, and the dedicated Javascript APIs for Google Maps and for ArcGIS.

The remaining students (Surveying and Mapping Science students) pursued research topics in a wide range of other spatial themes, many of which have a GIS ‘flavour’.  Environmental monitoring of coasts, glaciers, urban heat, and floods used GIS tools, as did further investigations of BIM, and further social investigations of education from a GI perspective.

My GISRUK 2016 Experience

So this year I (Kaizer) managed to attend the GISRUK conference in London with Dr. David Fairbairn. I must admit it was quite a good experience to meet other researchers working in the GIS field. I also had a chance to present my poster to a large number of people who never gave me a breather because they kept on coming and asking more questions. It was a good experience, to share with them what my work entailed. Their fascination was mostly on the fact that I was working towards introducing a novel concept of VGI, to a developing country (Botswana). Volunteered Geographic Information in Land Administration has a big role to play in developing countries. This research with its objectives of investigating how such a technology can be used to improve Land Administration Systems in developing countries is very optimistic that it is what the doctor ordered for regions with limited resources, dysfunctional, closed and ineffective land administration systems.

Having standards based systems has worked diligently for western countries and a large number of developing countries have tried to follow suit, with less success. The reasons being that they are simply not affordable to these countries, while in other cases they lack capacity to carry them forward after they have been implemented. As such, alternative solutions like VGI, which aim to develop affordable, fit for purpose, more open and participatory solutions are worth investigating. Our belief is that they are the answer to these countries especially in rural areas where there is little or no land information to help rural communities secure tenure. In addition, the participatory nature of VGI helps individuals become aware of activities in the communities and they are better poised to manage their land and its resources. With VGI, there can be clarity of who owns what, where, and in what proportions. Conflicts can also be reduced amongst neighbouring villagers in the community.

Reflections on GISRUK 2016

The 24th GISRUK conference took place last week in the University of Greenwich.  After significant Newcastle input to the 2015 GISRUK in Leeds, this year saw a smaller presence – just David Fairbairn and Kaizer Moreri attended, each with a poster, each on VGI and the fidelity and value of such citizen-sourced data for applications in national mapping and in land registration systems; the overall attendance and size of programme was certainly smaller than previously. 

The venue was certainly impressive, Greenwich (a UNESCO World Heritage Site) looking stunning in the spring sunshine.  The university’s vice-chancellor, David Maguire – GIS guru from his time at ESRI and his continued joint authorship of ‘the book(s)’ – welcomed us, thanking the main organisers, Zena Wood and Mike Worboys from the Greenwich GIS Research Group.

There was the usual variety of sessions and papers, with some unusual focus on gazetteers, on novel systems and applications (e.g. an innovative tourist guide to Lancaster), and on VGI issues and uses.  Those ex-Newcastle GISRUK stalwarts from Maynooth introduced us to some interesting personalities, responsible to addressing GIS concepts long before GIS was ever thought of …

Workshops and challenges, and an appeal to informality which benefits the large number of MSc students brought along to GISRUK every year, are what can be expected from this annual meet-up.

The three keynotes, well-scheduled throughout the programme, were the highlights: Ross Purves (University of Zurich) demonstrated the enduring value of his long-standing research on tags, semantics, ontologies, and full text retrieval and analysis; Nye Parry (University of Greenwich) took us on a tour of music (and dissonance) which had a spatial aspect to it; and Jeremy Morley (an external examiner at Newcastle in a previous life), demonstrated, amongst other things, the response of Ordnance Survey to changes and opportunities in GIS technologies, in a wide-ranging and thoughtful presentation.  Amongst his assertions was that fundamentally GIS has not changed in the past 30 years, and that all developments have been incremental ‘add-ons’.  It was ever thus, of course, with Ordnance Survey itself being recognisable, even today, to the apocryphal cavalry officer for whom its maps were created from 1791.  The main debate of the conference, and perhaps one its main outcomes, was about the nature of change in GIS, and whether GIS needs a revolution or continued evolution.

Dr Rachel Gaulton to co-organise Royal Society international research meeting

Dr Rachel Gaulton, School of Civil Engineering and Geosciences, with colleagues from Salford University (Professor Mark Danson, lead organiser), University of Massachusetts Boston, and University College London, has won support from the Royal Society to hold a Theo Murphy International Scientific Meeting at The Society’s Chicheley Hall. The meeting will bring together key researchers from around the world to discuss “The terrestrial laser scanning revolution in forest ecology” and amongst the sixteen invited speakers are researchers from Australia, United States, Finland, Netherlands and the UK. The meeting will lead to a special themed issue of the Royal Society’s inter-disciplinary journal Interface Focus.

Forest Structure

Complex forest structure is difficult to quantify from traditional field inventory but can be characterised in 3-dimensions using TLS

Terrestrial laser scanners, or TLS for short, provide detailed three-dimensional measurements of forests with unprecedented accuracy, by firing millions of laser pulses up into the canopy. These measurements are set to revolutionize the way in which ecologists measure forests, allowing changes over time to be characterised, and will help scientists to understand the role of forests globally in carbon storage and to monitor the impacts of climate change.  Newcastle University Geomatics Research Group has a long track-record of research at the forefront of TLS processing and application, including recent NERC-funded work on dual-wavelength laser scanning for forest health monitoring and the meeting will provide a key showcase for this work, whilst helping set the wider agenda for the future of the field. The meeting will take place on the 27-28th February 2017 with the programme released soon.

SALCA forest image

Forest laser scan from the Salford Advanced Laser Canopy Analyser

 

Ordnance Survey PhD Workshop

February 9th and 10th was the third of what has become an established annual event for Ordnance Survey (OS); inviting all sponsored PhD and postdoc researchers to present their latest research at the Ordnance Survey head office in Southampton. Whilst many know OS mainly for their paper maps, there is a wide range of research that is being undertaken to keep OS at the forefront of mapping technologies. This diversity in research was presented over the two days in four themed sessions. The opening session was ‘3D’ and was opened with my research on automatic reconstruction from a dense image matching dataset. Other 3D research presented included automatically adding texture maps and semantic information to 3D building models (Jon Slade, Cardiff University), why people need 3D (Kelvin Wong, University College London) and real-time urban dataspace modelling (Oliver Dawkins, University College London).

As well as giving oral presentations, a panel session was held at the end of each session where the session’s presenters and OS employees working in the relevant field were asked questions on the overriding theme of the session. This generally led to some really interesting and insightful discussions about how the panel saw the discussed research fitting into the OS remit. It also became apparent that parts of the research presented in the session as well as the other session had several overlapping themes, opening the opportunities for future collaboration. A poster session was also held after lunch each day which allowed OS employees who may not be directly involved with the research to view what studies are currently being undertaken, whilst giving the authors an opportunity to network and discuss their results.

The final session of day one was themed ‘Data Analysis’. The first two presentations from Robin Frew, University of South Wales, and Rebecca King, University of Southampton, addressed spatial usability issues and temporal modelling, respectively. The final two speakers of this session were first year PhD students presenting an overview of their research; Nick Bennett, University of Southampton, and Judit Varga, University of Nottingham, will be investigating similar topics in the use of data mining to map events and update mapping at different scales.

The second day opened with the theme of ‘People and Places’, the social sciences side of Ordnance Survey research. This covered a wide range of research included mapping vernacular geographies (Katherine Stansfeld, Royal Holloway, University of London), how people identify with narratives and place (Iona Fitzpatrick, University of Nottingham) and how different users interact with geo-spatial technology (Mike Duggan, Royal Holloway, University of London).

To close the workshop the final session theme was ‘Machine Learning’. Presenters Ce Zhang (Lancaster University) and postdoc David Young (University of Southampton) discussed their research utilising deep learning for supervised and unsupervised image classification.

As well as giving us the chance to see the various research projects undertaken by OS as well as getting to see our external supervisors, it also gives the chance to see how fellow PhDs are progressing through their projects and share an tips or advice of how to overcome the next hurdle. As I am now approach the writing up stages of my research, this will unfortunately be my last OS PhD workshop. I would therefore like to take this opportunity to thank Ordnance Survey for the data supplied and their continuing support throughout the progression on this research.

Andrew McClune

Final year PhD Researcher

LiDAR & hedgerows at the British Ecological Society Conference

I attended the British Ecological Societies 2015 conference over 4 days in December in the beautiful city of Edinburgh complete with Christmas jumpers and a ceilidh. It was a great opportunity to present my work – using LiDAR to model and measure the effects of cutting and rejuvenation management on hedgerow structural condition as wildlife habitat; to over 1500 delegates and to engage with the more ecology focused side of my PhD.

(image credit: BES on twitter)

(image credit: BES on twitter)

The conference saw talks and posters from across a range of ecological disciplines including using various remote sensing techniques to measure and monitor forests from the global to the local scale. This included Markus Eichhorn from The University of Nottingham who has invited me to present my work to his research group in the new year and was keen to hear of other relevant work going on in Newcastle CEG. I would encourage anyone in CEG whose work is based around forest ecology to consider attending next year’s meeting as their was a decent sized community of forest modelers in attendance.

poster (2)

Ensuring I could effectively communicate the potential of my work to such a broad range of scientists attending the conference was a big challenge and exciting opportunity. The conference was a great reminder for those working with emerging technologies or applying existing technologies to new challenges that communicating to the end users of your work is key. In my case this includes agricultural ecologists and policy makers with a range of interest and familiarity with remote sensing and LiDAR. I was asked everything from “What is LiDAR?” to questions about how LiDAR could add to a huge range of ecological monitoring challenges, stressing the real need for communication and sharing of ideas across disciplines. I also believe attending BES was a great way for my work to stand out and to really engage with what is novel about my project, as the only presenter (that I heard of) attending from a civil engineering discipline I was in a really unique position.

Lyndsey

Visualisation workshop

A number of us last week attended a workshop on visualisation, organised by Nick Holliman, professor of visualisation in the school of computing scince. After successfully having abstracts accepted David Fairbairn gave a presentation on ‘using geovisualisation in the decision theatre’ and Neil Harris presented his and Craig Robson’s work entitled ‘spatio-temporal network simulation and failure visualisation’ (as presented at GISRUK earlier in the year).

The aim of the workshop was to firstly share the work going on in the visualisation of data in research around different groups and departments across the university. This was achieved through a series of presentations giving an opportunity for people to present their work on visualisations as used in their research. This also gave the opportunity for them to raise questions about their work with regard to how to improve the visual outputs as well as pose thought provoking questions in the area of visualistion. This session provoked some productive discusions on the area of visualisation in gereneral in repsonce to the material presented.

The afternoon session focused on two questions; (1) what tools and methods are commonly used for visualisation in research and (2) what is required for the future of visualisations in research. Breakout groups were used to facilitate disscusions around these questions as well as full group discussions allowing key points to be discussed by the wider group.

The day proved useful in many ways by combining a range of people with varying thoughts and interests with regard to the visualisation of data. By the end of the day future meetings and collaborations were being discsused along with ideas on how to continue such meetings/workshops in the future.

 

ITRC @Newcastle – autumn 2015

As a research group a small number of us have been involved in the ITRC project over the past five years, namely Stuart Barr and Dave Alderson. Craig Robson joined the ITRC ranks in January, pausing his PhD work to help complete the final phase of the research required by autumn 2015. ITRC (Infrastructure Transitions Research Consortium) has been investigating the future of the UK’s national scale infrastructure with regards to how it must develop to meet ever changing demands and how climate change amoungst other factors will affect the resilience of those networks we rely on.

Our role in the project has centered around the development of the tools which would allow the five year project to be completed and included, but was not limited to, the development of the central database for all data for the project, and the support tools which would enable the analysis to be undertaken and results reported. It is on this later point where most of the past 6 months have been spent; developing a reporting tool for the presentation of the results from the long-term infrastructure planning aspect of the project.

The developed reporting tool allows users to view results from the each infrastructure sector (e.g. transport or waster supply), or view cross sector results, a set of similar metrics computed for each sector allowing for direct comparisons between them on there performance. For each sector a range of model outputs can be viewed from the level of emissions produced, to the running costs per year to the cumulative capital investment required. Results are shown not only at the regional level, but where possible at the sub-national level through the government office regions for example where the models output data at this granularity. This allows the tool to show data not just in charts, but also through maps, allowing new insights to be learned which may not be identified through non-spatial results. More detail on the tool, along with images and the like will be provided in a specific post at a later date, but for now a small selection of images below exemplify the tool.

Drawing1

With the end of the project looming near an event was organised at the ICE in London on the 15th October where the key results and impacts from the project could be disseminated to a wider audience with those key members behind the research all being present to answer questions and discuss their work. At the event we were available to demo the reporting tool and discuss the complexities behind the database for those interested, while a set of slides were used to give a overview of our work. More generally two videos (below) were produced giving an overview of the project and the one on the resilience of the UK’s national infrastructure.