Visualization of climate variable correlation: wind blowing across Europe?
Reference professor
Giovanna Venuti
AREA DI RICERCA
Geospatial Data Analysis – Earth Observations
KEYWORDS
Reduced order models, weather data, renewable energy forecast
TECHNOLOGIES
– Languages: Python
– Suggested SW/libraries: e.g., Pygrib, Geopandas, Rasterio, Tensorflow, Django
DESCRIPTION
Development of a stand alone tool able to retrieve, extract and process open-source weather reanalysis data (ERA5) and population data (GHS). The tool is meant to derive and plot maps of the spatial correlation of variables of interest for renewable energy forecast (Temperature, Solar Irradiance, Wind velocities, Precipitations, etc.) among different European areas, at different time scales. In order to deal with the large dimensionality of the weather datasets, a reduced order basis in space and time will be computed and adopted to project historical data, either with a linear (Proper Orthogonal Decomposition) or nonlinear (Autoencoders) approach.
Streets of Women (G)
Maria Antonia Brovelli
maria.brovelli@polimi.it
RESEARCH AREA
Geospatial Information
KEYWORDS
OpenStreetMap, GeoAnalysis, GeoVisualisation
TECHNOLOGIES
– Languages: Python – JavaScript
– Suggested SW: Postgres-PostGIS, Leaflet
DESCRIPTION
The project requires to develop a Web app to enable analyzing and displaying streets named after female role models for Italy. The app will consume OpenStreetMap (OSM) data to detect the streets of interest and compute city-wise summary statistics. The recognized female names will be searched using the Wikipedia API, and the percentage of positive findings will be included in the summary statistics record. The app will include a dedicated map-based client to visualize and query the data.
The project is related to the activities of an international association (Geochicas: https://geochicas.org/) of women belonging to OSM. Their aim is to close the gender gap in the OSM community, where it is estimated that women account for 3% of people who collaborate globally. One of their projects is “Streets of Women”.
Land cover AI-classifier with OpenStreetMap (I/G)
Maria Antonia Brovelli
maria.brovelli@polimi.it
RESEARCH AREA
Geospatial Information
KEYWORDS
OpenStreetMap, Machine Learning, Land Cover Maps
TECHNOLOGIES
– Languages: Python – JavaScript
– Suggested SW: Postgres-PostGIS, Leaflet, Overpass API
DESCRIPTION
OpenStreetMap (OSM) provides increased data availability for multiple purposes, such as validating land cover maps. This project requires to consume OSM data to match land cover classes by using OSM data labels. The extracted data should serve to calculate label counts, object areas and types within each spatial unit of a selected study area. This information is used to train a machine learning model to classify each unit according to predefined categories.
The project is strictly connected to an ESA CCI + project related to computation and assessment of high-resolution global land cover maps.
Google Earth Engine App for desert monitoring (I/G)
Maria Antonia Brovelli
maria.brovelli@polimi.it
RESEARCH AREA
Geospatial Information
KEYWORDS
Earth Engine, Desert, Earth Observation
TECHNOLOGIES
– Languages: JavaScript and/or Python
– Suggested SW: Google Earth Engine, Node.js
DESCRIPTION
Google Earth Engine (GEE) is promising to develop the next generation of data intensive Web mapping platform. This projects requires to exploit the GEE API to create a map-based Web app to search and browse time series of optical satellite imagery (NASA-Landsat 8 and possibly ESA-Sentinel 2). Imagery will refer to the Emirates Desert. The app should be designed for further integration of processing functionalities, such as imagery classification, already partially available from the GEE API store.
The project is partially connected to an ESA CCI + project related to computation and assessment of high-resolution global land cover maps.
Scientific data visualization
REFERENCE PROFESSOR
Barbara Pernici
barbara.pernici@polimi.it
RESEARCH AREA
Information Systems
KEYWORDS
Big Data, data and process modeling
TECHNOLOGIES
– Python, Postgres
DESCRIPTION
Visualization of numeric data sets, starting from data available in a Postgres database and some analysis functions. Implementation of model-based exploration visualization interface modules.
Tweet geolocation: quality improvement based on generic geographic terms
REFERENCE PROFESSOR
Barbara Pernici
barbara.pernici@polimi.it
RESEARCH AREA
Information Systems
KEYWORDS
Information extraction from social media, microservices
TECHNOLOGIES
– Python
– OSM with Nominatim
– Postgres with PostGIS
– REST, Jason
DESCRIPTION
Exploratory projects are proposed for the use of information from social networks and / or crowdsourcing to collect information to improve the quality of information in the maps for emergencies (earthquakes, floods). Focus will be on information useful to improve the maps generated (semi) automatically immediately after the events starting from satellite data with the EMS system of the European project Copernicus http://emergency.copernicus.eu/, see https://www.e2mc-project.eu/