Julie Le Gallo received her PhD from the University of Bourgogne in 2002 and is now full Professor in Economics at AgroSup Dijon, University of Bourgogne Franche-Comté (France). She is member of CESAER (Center of Economics and Sociology applied to Agriculture and Rural Areas).
Her research interests lie in the fields of spatial statistics, spatial econometrics, microeconometrics, land and housing markets and economic geography. In spatial statistics and econometrics, she has worked on exploratory space-time data analysis, cross-sectional and panel data models including endogenous variables, scan tests with applications in hedonic housing models, land use models strategic interaction models and growth and convergence models.
She has published over 70 articles in peer-reviewed scientific journals including American Journal of Agricultural Economics, Research Policy, Economic Geography and Journal of Urban Economics, 20 book chapters including review chapters in economic and econometrics handbooks (h-index = 20 in Web of Science). She is a fellow of the Spatial Econometrics Association, a co-editor of the peer-reviewed journals Spatial Economic Analysis and Papers in Regional Science and member of numerous editorial boards, including Journal of Spatial Econometrics. 

Chris Brunsdon is currently Professor of Geocomputation, and Director of the National Centre for Geocomputation at Maynooth University. Prior to this he was Professor of human Geography at the University of Liverpool in the UK. He has degrees from Durham University (BSc Mathematics) and Newcastle University (MSc Medical Statistics, PhD in Geography). His research interests cover various aspects and applications of spatial statistics, data science and spatial analysis. In particular, he advocates the use of open source software, and reproducible methods, mainly via the R Statistical Programming Language. In terms of subject areas, he has interests in crime pattern analysis, house price modelling, health geography and environmental applications.

Intro to sp – reading and working with data, basic visualisation

Instruction: Kateryna Zabarina
Date: June 21, 2022 9:00 – 11:00
Software: R

In our everyday life, we meet spatial objects all the time – bus stops, transport lines, various buildings, rivers etc. All of them can be expressed using geometric figures – point, line, polygon and to each of them, we can add a feature (number of bus lines, square of building, length of the river etc.). This workshop introduces to a multifaceted world of spatial data in R package sp. You will learn how to read and change data, create them from scratch, and visualise them during it. Among others, we will calculate the distance between point and line and between points. The workshop will cover instruction and practice sessions.

Simple features – modern and convenient way of manipulating spatial data in R

Instruction: Maria Kubara
Date: June 21, 2022 11:30 – 13:30
Software: R

Using R as the main tool for spatial data analysis is a convenient choice. Well-developed environment, an increasing number of available methods and ongoing support of the R community can make your qualitative research much more comfortable. However, in order to make the best of the possibilities R provides, it is beneficial to stay up to date with the modern analytic approaches. Many spatial analysts are still using the sp:: package as their primary analytical tool for spatial data in R. Yet, this package is slowly being deprecated and replaced by a much more computationally effective package sf::.
Package sf:: transforms spatial entities into simple features, which enables storage of spatial data in data frames. Such an approach allows for much more convenient data manipulation. Moreover, it allows for easy utilising a vast tidyverse environment, including ggplot2 graphics, on your spatial datasets.
The goal of this workshop is to show how to manipulate spatial data in R, utilising sf:: package. Moving from spatial points and lines to polygons, we will cover most of the possible spatial structures which are used in regional analysis. It will also be shown how to go from simple features into spatial weight matrixes of different types, opening the road for further statistical and econometrical analysis. Data analysed during this workshop will include point collections in csv files as well as NUTS2 and NUTS4 shapefiles. Data will be read, transformed, and visualised using ggplot2 approach.

Spatstat – manipulating point (and line) data in R

Instruction: Kateryna Zabarina
Date: June 21, 2022 14:30 – 16:30
Software: R

Nowadays, we can observe the growing popularity of point data, which are also the most common kind of spatial data – locations of enterprises, cities, train stations etc. Some points are located randomly, but another does locate due to underlying behaviour or impact of terrain features so that such behaviour can be captured with point process models. Spatstat package in R allows for various operations in R – from statistical analysis to modelling. During the workshop, you will learn how to create a point pattern and work with it, what is the difference between marked and unmarked patterns, how to spot clustering and regularity and how to create a simple point process model and check its validity.

Clustering of spatial data: points, GWR coefficients, raster data

Instruction: prof. Katarzyna Kopczewska
Date: June 22, 2022 9:00 – 11:00
Software: R

Spatial machine learning (ML) has developed recently quickly and established analytical approaches being the natural alternatives to traditional spatial statistics and spatial econometrics. Unsupervised learning, part of machine learning, concentrates on clustering data in space. The workshop presents the modern solutions to consider space and spatial relations, especially practical implementations of clustering in space: k-means for clustering, block sampling, spatial cross-validation, detection of heterogeneity and anisotropy in spatial drift models etc. It will discuss different methods of clustering of points and values as well as refer to clustering of geo-referenced coefficients from Geographically Weighted Regression. It will also cover comparing different partitioning methods and present methods of rasterising point data to obtain common geo-reference.

Data-to-grid aggregation and basics of satellite imagery processing

Instruction: Ewa Dobrowolska
Date: June 22, 2022 11:30 – 13:30
Software: R

Every research requires well-prepared data to work with. However, sometimes the need for extra variables arises, especially in urban analysis. In such a case, one can use online sources of free data. With many options available, such as OpenStreetMaps providing information about the natural environment as well as the location of services and amenities, or Sentinel-2/Sentinel-3 files reporting various features, from multi-spectral data to climate monitoring, it is not hard to overcome the problem of missing variables. However, in spatial analysis sometimes simple data collection is not enough. Many researchers decide to divide the studied area into polygons and consider aggregated data instead of groups of points. One of the most efficient approaches is studying grid-leveled aggregation due to small and equal areas of the cells.
The workshop presents many methods of aggregating different types of data into grid cells: assigning the total number of points located within the given cell, calculating the total length of lines (such as roads or pathways), or the percentage of the cell’s area covered with a polygon (ex. a park or a forest), and more. Additionally, the workshop will cover the topic of simple manipulation of satellite imagery data: using the Sentinel-2 mission satellite and extracting the NDVI index values for area limited with contour map. As a result, one gets an application-ready self-made database, providing valuable insight into the urban features.

API interface and geocoding of spatial data

Instruction: Mateusz Kopyt, PhD
Date: June 22, 2022 14:30 – 16:30
Software: R

Analyses, not only spatial ones, require having an appropriate data resource for research. Acquiring such data can be done using external applications. More and more data providers currently enable their collection using the so-called API interface. The R program, along with the corresponding add-on packages, can be a tool that allows you to download data directly using this interface. The advantage of such a solution is that we do not involve any additional software for data collection and processing, and, above all, we can download the data dynamically, i.e. every time we are sure that we receive the latest data offered by the service providing it.
In the case of spatial analyses, it is actually essential to have information about the location of a given observation. In many cases, such information in the data set is provided descriptively (e.g. an address). In such a situation, it is necessary to assign locations to specific coordinates in the system used in a given study (most often geographic) by the so-called geocoding. Using available external services, the R program allows you to automate this process, obtaining satisfactory results relatively effectively.
The proposed workshop will aim to present examples of R communication using the API interface with selected data delivery services and their basic processing, in particular from the popular geoJSON exchange format to a form that allows further analysis. In the second part of the workshop, the possibilities of data geocoding will be presented, emphasising services based on open data and the OpenStreetMap service.