Projets 2019 - Université de Genève
←
→
Transcription du contenu de la page
Si votre navigateur ne rend pas la page correctement, lisez s'il vous plaît le contenu de la page ci-dessous
Projets 2019
Applications Informatiques
29 janvier 2019
Choix des projets
Les projets suivants ont été soumis par des chercheurs·ses de l’Université de Genève ainsi que
d’autres institutions.
Merci de prendre notes des remarques suivantes:
• Les projets sont présentés dans l’ordre de soumission.
• Il y a 40 projets, même si cela vous semble fastidieux, parcourez les tous avant de faire votre
choix.
• Une fois choisis, nous fixerons ensembles les priorités et les contraintes techniques.
• Tous les projets impliqueront une quantité similaire de travail, une fois choisis, nous fixerons
ensembles le cahier des charges et les priorités.
• Le choix des langages, librairies et outils sera également rediscuté une fois le projet choisi.
1Projet 1: Annotation de données MS
Pierre-Marie Allard, Phytochemistry and Bioactive Natural Products, Pharmacy
Context
Mass spectrometry (MS) is a tool of choice for the characterisation and metabolite annotation of
complex biological matrices. We develop computational approaches to proceed to the automatic
treatment of MS data. This annotation process is typically done by spectral comparison between
experimental spectra of the analyzed matrix and experimental or theoretical spectra present in a
database.
Description
In the frame of our metabolite annotation solution developments we are proceeding to the in
silico expansion of structural database of known compounds (passing from 300000 reported
structures to over 10M plausible structure). For this we used a set of open access tools available
at (https://github.com/JamesJeffryes). More precisely we employed the BNICE framework approach
and the pickaxe.py module. The next step is to deploy such structural database as a webapp.
The webapp is allready develloped (https://github.com/JamesJeffryes/MINE-app), but the servers
(https://github.com/JamesJeffryes/MINE-Server) codes are poorly documented and hard to set be-
cause of hard coded paths and poor dependency management. Furthermore our lack of knowledge
in this case is a problem.
Ideally we would like to : i) deploy the MINE servers locally and run the webapp so that we can manage
the previously generated DB ii) integrate our existing in silico fragmentation codes to populate the
structural DB with theoretical spectra ii) tune the existing webapp in order to proceed to batch query
of experimental spectra (single query beeing the only available mnode at the moment)
Looking forward to collaborate !
2Projet 2: Aggrégation de cendres volcaniques
Eduardo Rossi, Earth Science, Science
Context
Particle aggregation describes the formation of larger clusters from an initial population of single
particles. Ash from volcanic eruptions can aggregate, affecting its behavior in the atmosphere. After
the 2010 Eyjafjallajökull eruption, Iceland, it has been clear that dispersion models must adequately
represent aggregation in order to improve their predictions of the expected location and concentra-
tion of volcanic ash in the atmosphere. The aim of this project is to better constrain the process of
aggregation in NAME, the dispersion model developed at the Met Office, UK, and used by the London
VAAC to generate forecasts for the aviation industry.
Description
The aim of the project is to improve the existing scheme for ash aggregation inside NAME. NAME
is written in FORTRAN 90 and it has a specific module for ash aggregation called “Aggregation.f90”.
The first part of the project consists in translating an existing code for aggregation from Matlab to
FORTRAN 90. This code is describing the interaction of particles with two arbitrary internal parameters
(mass and porosity). In this part a good knowledge of both languages is obviously required. Most
of the criticalities are related to the use of built-in functions for Matlab that must be translated
in FORTRAN 90. The code developed in this phase will be tested using a stand-alone version of
the actual volcanic plume that is currently used in NAME. This allows the code to be compiled and
debugged remotely, here in Geneva. Moreover, this step is preparatory for the implementation of
the revised module “Aggregation.f90”. Depending on the time required for the first part, there is the
option of implementing the revised module “Aggregation.f90” inside NAME. In this case there will be
the possibility to travel to the Met Office (UK) to work on the implementation in NAME.
3Projet 3: Retombées de cendres volcaniques
Allan Fries, Département des sciences de la Terre, Faculté des Sciences
Context
Ce projet vise à mieux décrire les retombées de cendres volcaniques. Il s’inscrit dans une ap-
proche multidisciplinaire ayant pour but de caractériser la sédimentation collective des particules
volcaniques qui ne chutent pas toujours individuellement. Suite à une instabilité se formant à la
base du panache volcanique, elles peuvent notamment former des colonnes de cendres qui descen-
dent, dans un courant cohérent, à une vitesse supérieure à celle des particules retombant seules. Ce
comportement a potentiellement un effet sur le temps de résidence des cendres dans l’atmosphère
à prendre en compte, puisque les particules chutent plus rapidement qu’attendu.
Description
Afin de comprendre l’instabilité conduisant à la chute collective des particules volcaniques, des ex-
périences analogiques ont été réalisées en laboratoire. Dans une configuration simple, elles permet-
tent de détecter les principaux paramètres influençant la sédimentation des particules. Des relations
peuvent ensuite être tirées entre les paramètres d’entrée et les caractéristique de l’instabilité. Parmi
les caractéristiques essentielles, l’étude de la longueur d’onde de la perturbation (ou du nombre
de colonnes de sédimentation) est critique pour comprendre l’instabilité. Le but du présent projet
est d’améliorer la détection des colonnes de particules qui se forment lors des expériences. Les
colonnes sont détectées en repérant les pics d’intensité le long d’une ligne. Leur détection dépend
donc de plusieurs critère liés à la détection des pics (largeur, hauteur. . . ). Le projet est divisé en
trois parties: - Faire l’analyse de sensibilité de la détection des pics afin de connaître les paramètres
influençant le plus les résultats. Cela permet aussi de connaître l’incertitude associée aux résultats.
- Etudier les corrélations entre les paramètres d’entrée de la détection des pics pour ne garder qu’un
minimum de paramètres. - Améliorer le code de détection existant.
4Projet 4: Portofolio de projets étudiants
Patrick Roth, Groupe e-learning, DISTIC
Context
Vous travaillerez au cœur de l’équipe e-learning à l’enrichissement de la plateforme portfolio de
l’UNIGE. Le portfolio est une plate-forme Web qui est un outil d’aide à la réflexion et qui permet
le stockage des productions des étudiants. Le portfolio UNIGE est basé sur le projet open source
Mahara (https://mahara.org). La demande de développement provient du corps enseignant pour
leur faciliter l’export des collections pour les évaluations.
Description
Il vous est demandé de développer le plugin d’impression (ou export PDF) pour les collections du
portfolio de l’UNIGE. Actuellement, les impressions / export PDF ne sont possibles qu’au niveau
des pages individuelles qui composent les collections. Il vous faudra donc en partant du code du
portfolio existant implémenter la fonctionnalité de création de PDF contenant toutes les pages de la
collection choisie.
L’output du projet devra être rendue sous forme de plugin. Ce plugin pourra par la suite être proposé
à la communauté Mahara. Notons finalement que vous développerez le plugin en PHP et JavaScript.
Les étapes nécessaires à réaliser le projet sont les suivantes : 1. Prendre connaissance du cahier
des charges et planifier les différentes étapes du projet ainsi que le backlog des tâches à réaliser 2.
Prendre en main le code de la plateforme portfolio universitaire 3. Etudier les fonctions PHP d’export
PDF
4. Proposer une solution conceptuelle d’export PDF sur le portfolio 5. Développer la solution 6. Tester
le module et effectuer les modifications nécessaires
Le projet sera réalisé en suivant une méthodologie agile et le développement se fera sur l’outil de
gestion de version Git.
5Projet 5: Doit Yourself Biology
Gabriela A. Sanchez, Biologie, Sciences
Context
‘Do-it-yourself biology’ (DIYbio) is part of a growing movement where amateurs and enthusiasts par-
ticipate in the making of scientific knowledge and technological artifacts. But little is known about
the size, distribution, growth, and other characteristics of the movement.
‘DIYbiosphere’ (http://sphere.diybio.org/) was developed to gather data about DIYbio initiatives
through participatory means without engaging in ‘extractive research’. Instead, DIYbiosphere is
‘open-source’, creating a decentralized resource for the community to grow and connect with each
other while providing (social) researchers with quantitative and qualitative open data to map and
visualize the DIYbio community. Although the DIYbio community has expressed interest in the site,
the learning curve for using GitHub has proven to be a great barrier to participation.
Description
DIYbiosphere functions similar to a wiki but uses Jekyll and GitHub. The website is already up and run-
ning, designed and coded by UNIGE PhD student Gabriela A. Sanchez. Among the many features and
enhancements we need are: - Create friendly UI for editing entries and posting comments (GitHub
API) - Create Sign-in through GitHub (GitHub API) - Create plugin to geocode new or changes to loca-
tions - Improve mapping functionality (Mapbox API) - Improve search for initiatives (Algolia Search
API) - Create timeline and calendar functionality
6Projet 6: GUI pour analyses spectroscopiques
Tatu Kumpulainen, Department of Physical Chemistry, Faculty of Science
Context
In my work, I use several MatLab scripts to analyze spectroscopic data. The scripts are mostly com-
mand line-based with hard coded options and inputs. The theoretical, analysis, plotting of results
etc are already fully implemented in the scripts. Some time ago, I started programming a GUI for one
of the scripts but due to the lack of time the work was not completed. I still use this scripts often
and the GUI would be a nice addition to the overall usability of the script.
Description
I would like to propose a project for the completion of the GUI. Naturally, the student can start the
work from scratch, but I want to emphasize that I have already gained some experience in this, which
allows me to guide the student during the project. The project would consist of 1) defining the nec-
essary input/output options e.g. read-in of data, analysis options, saving results etc.; 2) designing the
graphics of the GUI with all necessary analysis options using dropdown menus, buttons etc, and the
plotting interfaces for visualization of the data and the results; 3) interfacing the GUI with the already
existing analysis scripts; 4) testing the GUI with real data provided together with the analysis scripts;
and 5) writing documentation (e.g. user instruction) for the GUI. The preferred programming environ-
ment is MatLab because the scripts are already written in that environment and hence the scripts can
be easily interfaced with a MatLab-based GUI. I do not know how experienced the students are with
this environment but several online tutorials/courses exist and personally, I find the environment
relatively easy to work with. I’m open to alternative approaches but the proposed approach would
benefit me personally the most.
7Projet 7: Application Web pour simuler des batteries résidentielles
Alejandro Pena Bello, Energy Efficiency, Institute for Environmental Sciences (ISE), Faculty of Science
Context
Interest in residential batteries to supply PV electricity on demand is increasing, however they are not
profitable yet. Combining applications has been suggested as a way to increase their attractiveness,
but the extent to which this can be achieved, as well as how the different value propositions may
affect the optimal battery technology, remain unclear. We have developed an optimization framework
to determine the best-suited battery technology and size depending on the applications combined,
including PV self-consumption, demand load shifting, demand peak shaving and avoidance of PV
curtailment. We use six different battery technologies currently available in the market.
Description
The idea of this project is to create a web tool interface that allows the user to select different inputs
for the model (e.g. Electricity tariff structure, electricity demand, applications to be combined, type
and size of battery, days to be optimized. . . ) and show different outputs (mainly plots). The student
will develop the web tool interface and adapt the model scripts in order to be able to communicate
with the GUI. The optimization model is fully written in python but integration with other languages
is possible if needed.
8Projet 8: Modèle topographie d’écoulement d’eau de surface
Marion ALCANIE, Earth and Environmental Sciences, Sciences
Context
Le groupe “Crustal Deformation ad Fluid Flow” (CDFF, https://www.unige.ch/sciences/terre/en/research/crustal-
deformation-and-fluid-flow/) au sein du département des Sciences de la Terre et de l’Environnement
étudie, entre autre, la géothermie dans le bassin de Genève. Afin d’identifier les zones à fort potentiel
géothermique dans la région genevoise, des modèles numériques dynamiques 3D sont en cours
de développement. Ces modèles sont construits via Matlab et une boite à outils spécialisée (MRST,
Matlab Reservoir Simulation Toolbox : https://www.sintef.no/projectweb/mrst/). Ils sont basés sur
des données géologiques et géophysiques acquises précédemment sur le terrain.
Description
La topographie est un élément moteur pour les écoulements d’eau de surface, et donc un paramètre
crucial à prendre en compte pour les modèles de terrain. L’objectif est d’intégrer la donnée topogra-
phie dans l’outil spécialisé MRST. Le travail concerne spécifiquement la conception de grille 3D et la
compréhension de l’architecture utilisée par MRST. Une finalité peut être de faire un outil qui soit un
pont entre une donnée géologique de format varié et une représentation numérique importé dans
un outil Matlab préexistant. Nous recherchons un étudiant de Bachelor pour travailler avec une étu-
diante doctorante sur ce projet. Il/ elle peut être curieux des problématiques géologiques, et doit
pouvoir communiquer un minimum en anglais (MRST exclusivement en anglais).
9Projet 9: Analyse de bioindicateurs environementaux
Tristan Cordier, Department of Genetics and Evolution, Science
Context
Our group is working on the development and application of genomic tools for biodiversity surveys
and environmental monitoring. Such tools require robust and accurate informatic pipelines for the
analysis of millions of environmental DNA (eDNA) sequences generated by high-throughput sequenc-
ing platforms. The key element of such pipelines is the taxonomic assignment of DNA sequences,
which allows identification of organisms present in studied environment. The accuracy of taxonomic
assignment plays a major role in the interpretation of eDNA data and the assessment of ecological
status.
Description
The project aims at implementing a new module within an existing pipeline (https://github.com/yoann-
dufresne/SLIM) used routinely in our laboratory (in press in BMC bioinformatics). Such module is
needed to improve the accuracy of taxonomic assignment of marine foraminifera, our target group
of bioindicators. The genetic marker for this group exhibits a high degree of length variation, ranging
from a dozen to 60 base pairs. The challenge is to find a method that would overcome this issue.
The student will compare six published software for the taxonomic assignment of eDNA data. He
will focus on newer approaches that use machine learning and reference sequences database as
training data, for example, using k-mers to detect taxonomic signatures. Such approaches allow
to move away from single, subjective criteria such as sequence similarity, to a more data-driven
approach, hence being adapted to wider use cases. She/he will compare the accuracy of examined
methods by looking at the amount of false positive and false negative taxonomic assignments in
mock and artificial (in silico) communities. The output of the project will be the implementation of
selected method in our pipeline.
10Projet 10: Outils didactiques pour la chimie
Didier PERRET, Chimiscope, Sciences
Context
The Chimiscope (http://scienscope.unige.ch/chimiscope/), member of the Scienscope, is the UNIGE
public platform for the discovery and experimentation of the world of chemistry. With more than
36’000 visitors since its inauguration in Nov. 2011, the Chimiscope has become a reference in the
Geneva’s landscape of education.
At the occasion of the International Year of the Periodic Table of the Elements, the Chimiscope aims
at providing to its visitors of its website a series of games, quizz and tools on the Periodic Table to
increase its visibility and promote the understanding of the basics on the chemical elements.
Description
The games/quizz/tools on the Periodic Table already exist!
They were developed (JavaScript + PHP with MySQL) for the Principal Investigator by a former collabo-
rator (now retired) of the School of chemistry and biochemistry before the Chimiscope was launched.
These codes were hosted on silene3 and operated from the School’s webpages, but they were aban-
doned when the School’s website was reformated under Concrete5.
silene3 will be shut down around end of March 2019, and the codes on the Periodic Table will be
migrated (by Frédéric Radeff) on silene5.
However, nobody in our School is able to adapt and modify the codes (from PHP3 to PHP5), so that
they can be run by the public from the webpages of the Chimiscope (site developed with WordPress
and hosted by Infomaniak).
The project would consist in helping the Chimiscope to give a rebirth to the games/quizz/tools on
the Periodic Table and to make them available to the public on its website. This includes testing their
proper operation on silene5 after migration from silene3, then adapting their use to the specificities
and format of the Chimiscope’s website.
11Projet 11: Application web pour organisation de meeting
Charles Dalang, Physics, Science
Context
Every Friday, the Cosmology group goes for lunch at the RTS cafeteria to take the weekly seminar
speaker to a fancy place. Since we sometimes are a large group of external people (10-30), the chef
of the cafeteria has asked us to send him a weekly email containing the number of people of our
group that will have lunch at the RTS cafeteria, so that they can prepare and still have food for their
employees (thereby avoiding starvation). Otherwise, they will not let us have lunch there anymore.
Since we love that place and have gone there for years, we have charged a member of the group to
collect knowledge of who will be having lunch there and inform the chef weekly. Since this job is
extremely repetitive, we would like to automatize this task.
Description
Every Thursday morning, an email should be sent to all the members of the cosmology group con-
taining a link on which members can click to manifest their presence at the TV cafeteria for the next
Friday. In the afternoon, an email should be sent to the chef of the Cafeteria containing the exact
number of people who have clicked on the link.
12Projet 12: Interface graphique pour l’étude de motifs de reptiles
Athanasia Tzika, Génétique et Evolution, Science
Context
The color pattern of corn snakes is characterised by red blotches on an orange background, but
several morphs exist where either the color or the pattern is modified. We are interested in identifying
the causative mutations of these morphs. We have sequenced, assembled and annotated the corn
snake genome and we use mapping to locate the candidate genes responsible for the different colors
and patterns.
Description
In order to map the causative mutations, we need to run a series of existing softwares on the Baobab
cluster. We have developed several Python scripts which allow us to refine the results, by identifying
for example if a mutation is within a gene sequence and which gene this is. The input for these
scripts is often a TXT file and the output is XLS formatted. For some of the scripts, there are a few
parameters that the user needs to define. For this project, we are interested in developing a graphical
interface which will facilitate our analysis. The student involved will interact on a daily basis with the
bioinformatician who wrote the scripts and with the biology technicians who will be the end users
of the application.
13Projet 13: Interface graphiquee pour simulation de microscopie
Luigi Bonacina, Applied Physics, Science
Context
We work on deep-tissue imaging based on multiphoton microscopy in the infrared. We would
like to support our results with ad hoc simulation of beam propagation in a scattering tis-
sue. The simulation should be based on a free Python library recently released available at
https://github.com/maweigert/biobeam.
Description
Create a graphical user interface enabling the researcher to enter meaningful physical parameters
(tissue characteristics, laser pulse properties, etc.). Visualize graphically the results obtained using
the commands available in the Python library. Export the results for further analysis in MATLAB or
other data manipulation software.
14Projet 14: Application smartphone pour le calcul symbolique d’unités
physiques
Christophe Berthod, DQMP, Sciences
Context
Dans la pratique quotidienne de leur métier, chercheurs(euses) et ingénieur(e)s sont fréquemment
amené(e)s à évaluer des formules mathématiques contenant des constantes physiques. Le résultat
numérique de chaque évaluation est fonction des unités choisies. Cette opération souvent fasti-
dieuse est propice aux erreurs. Non moins fréquemment rencontrée, la nécessité de convertir cer-
taines grandeurs entre différents systèmes d’unités est souvent source de désagrément. Formelle-
ment, toute conversion de ce type se réduit à l’évaluation d’une formule dans laquelle les constantes
sont cette fois-ci des unités. Un outil de calcul symbolique spécialisé dans les unités serait d’une
grande aide.
Description
Il faut saisir la différence entre les convertisseurs d’unités usuels (ceux-ci n’offrent qu’une liste
prédéfinie d’unités) et le projet proposé qui manipule grandeurs physiques et unités par du cal-
cul symbolique. Mathematica, pour ne citer que cet exemple, offre un module de calcul symbolique
consacré aux unités. Cependant, pour être vraiment utile l’outil doit être simple et rapide, immédi-
atement disponible, que ce soit sous forme d’un widget sur un ordinateur ou d’une application sur
un smartphone. La personne de contact a développé une solution sous la forme d’une page web
(http://mafalda.unige.ch/vuc/converter.html) et d’un widget mac (https://goo.gl/B26EHf). En fonc-
tion depuis 20 ans avec de nombreux utilisateurs dans le monde, cette solution doit être modernisée.
Il lui manque notamment certaines fonctionnalités, comme les puissances fractionnaires, la mémori-
sation de combinaisons, la possibilité de définir de nouveaux symboles, etc. De plus, une version
pour smartphone fait cruellement défaut.
Le projet implique de l’algorithmique, de la programmation et du design. Le problème algorithmique
principal est de déchiffrer l’input (qui représente une formule mathématique). Le travail de program-
mation en C/C++/Objective-C pourra s’inspirer de la solution existante (javascript). Enfin le design
d’une interface intuitive sur smartphone complète le travail.
15Projet 15: Portage d’application sous Androïd
Jérôme Kasparian, Groupe de physique appliquée & Institut des Sciences de l’environnement, Sci-
ences
Context
Sur la base d’une méthode de formation et d’évaluation des étudiants basés sur l’obsrvation en
situation, nous avons développé une app iOS permettant aux enseignants de consigner leurs obser-
vations, de les compiler et d’avoir une vue d’ensemble de la progression de chaque étudiant ainsi
que du groupe dans son ensemble.
Description
Le projet proposé consiste à porter l’app sous Android pour permettre à des équipes munies de
matériel hétérogène de l’utiliser.
16Projet 16: Synchronisation multimodales pour sciences affectives
Guillaume Chanel, Departement d’informatique, Sciences
Context
Dans le cadre de notre recherche en informatique affective nous construisons une plateforme
d’acquisition multimodale synchronisée. Cela nécessite d’enregistrer des informations conjoin-
tement avec des annotations temporelles. Nous avons actuellement un logiciel permettant
l’acquisition de vidéos, à partir de webcam standard, tout en annotant le temps de collecte de
chaque trame.
Description
Notre objectif est d’augmenter le logiciel existant de deux manières. La première augmentation
concerne la synchronization multimodale. Nous souhaiterions intégrer l’annotation temporelle dans
un système existant de synchronization multimodale appellé “labstreaming layer”. Il faudra donc
faire appel à cette librairie externe et faire en sorte que chaque trame soit annotée temporellement
dans le système d’acquisition de labstreaming layer.
La seconde augmentation concerne l’acquisition de la vidéo elle même. Le logiciel existant est
basé sur le framework d’acquisition video gstreamer. Pour le moment la plus part des paramètres
d’acquisition sont spécifié directement dans le code. Nous souhaiterions pouvoir donner le choix
de ces paramètres à l’utilisateur. Il s’agit donc de concevoir un moyen de lister les possibilité
d’acquisition (e.g. les périphériques d’acquisition et les codecs disponibles) sur la machine hôte
et d’implémenter l’interface utilisateur correspondante.
17Projet 17: Application de suivi des étudiants de l’école doctorale des
sciences de la vie
Nicolas Roggli, Décanat, Sciences
Context
L’école doctoral en Sciences de la vie des Facultés de médecine et des sciences (https://lifesciencesphd.unige.ch)
développe une application web pour le suivit de ses étudiants. Cette application doit implémenter
les différents points du réglement de l’école régissants le cursus de l’étudiant.
Cette application sera accessible aux étudiants afin qu’ils puissent y ajouter les documents qui leur
sont demandés et puissent voir les deadlines qu’ils ont à respecter. Les directeurs de thèse pour-
ront aussi consulter les dossiers de leurs étudiants. Ces dossiers seront aussi consultable par les
coordinateurs du programme.
Description
Le projet est d’implementer la partie “étudiant” dans une application PHP/MySQL existante, frame-
work CakPHP3. Il faudra respecter les différents points du réglement de l’école pour la gestion des
alarmes, implémenter les tests. Il y a du travail sur le schéma de la base de donnée, sur les dif-
férents vues, sur le UX. Je vous donnerai volontiers plus de détails si ce projet est potentiellement
intéressant de votre point de vue.
Un repo existe: https://gitlab.unige.ch/Nicolas.Roggli/phd-members
18Projet 18: Modèle géographique pour la tectonique des plaques
Christian Vérard, Earth Sciences, Science
Context
PANALESIS is a global plate tectonic model currently developed under ArcGIS® to reconstruct the his-
tory of the Earth over geological times. The model supersedes a former model made at the University
of Lausanne (UNIL). For the latter, a code in .NET language has been developed to convert the maps
(polylines & polygons drawn by hand under ArcGIS.10.1) into 3D topographic surfaces i.e. as Digital
Elevation Models (DEM). In association with our FNS Synergia project (started in November 2018), we
now need to apply that code to the new PANALESIS model.
Description
The aim of the project consists in translating/updating the .NET code made for the in UNIL model
(as Add-In under ArcGIS.10.1) into a Python code for PANALESIS under ArcGIS.Pro (and/or potentially
similar software such as QGIS). The project is a good opportunity for a student to contribute to major
research in Earth Sciences, to use a popular coding language (Python) on an up-to-date platform
(ESRI) widely used in major private companies.
19Projet 19: Interface web pour piloter une camera de telescope à
rayon-gamma
Matthieu Heller, DPNC, Science
Context
The astrophysics group at DPNC is building a camera for gamma ray astronomy. Using the first de-
veloped optical modules from the prototype, we have built a small camera of 144 pixels for outreach
purposes which will be installed at the Observatoire in Versoix. The camera will be operated few
nights per year for outreach operation and potentially for master or bachelor students in the frame
of practical work.
Description
The project we propose is to create a web interface for the operation of the a camera for gamma-
ray astronomy telescopes that would also fit mobile phone use. From the interface, the user could
operate the camera, if given a “expert” account, or simply view live events (when the camera is
operated) or registered events (off-line, e.g. during day). An existing GUI has been developed in
python TK and runs locally on the server that will be installed at the Observatoire of Genève. The
goal of the project is to develop the web interface that would mimic the functionalities of the GUI.
Additionally, the data acquisition software developed in python could be improved to treat the data
in real-time. All the existing code (mini camera control+GUI) will be made available via github. The
data acquisition libraries (written in C#) are very well documented and the developer is the leader
of the electronics group at DPNC.
20Projet 20: Interface web pour l’analyse de séries temporelles
Andreas Suter, Quantum Matter Physics, Science
Context
The laboratory for Muon Spin Spectroscopy at the Paul Scherrer Institute operates the world leading
muSR user facility. These provide Swiss and international users with a wealth of different experi-
mental capabilities with respect to e.g. temperature, magnetic field, etc. The data acquisition for the
muSR instruments, based on the MIDAS framework (https://midas.triumf.ca), registers not only the
necessary muon decay events, but also the time series of all relevant experimental parameters.
The goal of the project is to establish a flexible, easy to use web interface to MIDAS (sever/client)
which allows to present time series and correlations of any recorded parameters.
Description
The aim of the project is the realization of a server/client based application within the MIDAS
framework, which allows to select any available recorded time series parameters and either
plot/export them as time series, or plot/export parameters against each other, e.g. temperature
versus resistivity. On the server side the existing http server (C application) needs to be extended
to provide the necessary facilities to access and propagate the needed data. On the client side
a JavaScript based application needs to be developed which interacts with the server and al-
lows to select the requested data. The client should be able to plot the data as requested and
furthermore allows to export these data sets, e.g. as CSV data. MIDAS already provides most of
the necessary ingredients to fulfill the given task. For details concerning the server side see
https://midas.triumf.ca/MidasWiki/index.php/Mhttpd.
The client side should be implemented along the lines of so called ‘custom pages’ (see
https://midas.triumf.ca/MidasWiki/index.php/Custom_Page). It is possible to split this task
and concentrate on either the server or the client side only.
21Projet 21: Simulation climatiques à la transition Permienne-
Triassique
Jan-Hendrik Malles, Institut des Sciences de l’environnement, Sciences
Context
We need to perform climate simulations of the Permian-Triassic boundary (250 millions of years in
the past) as described in the SNF Sinergia project started in November 2018. The climate model
that we use is the MIT general circulation model, which requires a runoff map as input, i.e. a dataset
describing how each land-point is drained to a point in the ocean. Because of climate oscillations
related to volcanic activity, the sea level changed and consequently the coast line position. Thus, we
need to reconstruct the runoff map for each climatic condition and at different horizontal resolutions.
Description
At present, we construct the runoff map using a laborious sequence of ArcGIS tools and manual edit-
ing steps that, for a given topography and resolution, allow us to find the associated hydrological
basins and thereby the potential river positions. In this way, we can associate each land point to
points in the ocean near the coast line, where water coming from precipitation is discharged to (as-
suming that it is not absorbed within the soil). We need to repeat this procedure several times, since
we need to explore different configurations of the coast lines due to different climatic conditions, or
due to increased numerical resolution.
We would like to automatize this procedure by using Python scripts that call the different procedures
in ArcGIS (or any similar program such as QGIS). Especially interesting for the programming task
might be the steps within this procedure that have to be done manually until now, as e.g. changing
the topology of geographic data to obtain a closed coastline.
22Projet 22: Analyse de processus métaboliques
Santiago Codesido, Analytical Sciences, School of Pharmaceutical Sciences
Context
Metabolomics is the study of the metabolism via fingerprints of low molecular weight compounds,
i.e. metabolites. It is the natural continuation of genomics and proteomics needed to complete
the picture of chemical processes in the cell. Unlike the latter two, its experimental target is not
made of sequences of similar building blocks, and the wide variety of compounds studied results
in a wide variety of experimental techniques required. All these must be understood and processed
as a whole. Having bioinformatics tools for data treatment, analysis and visualization becomes a
necessity, particularly with the ever-increasing throughput of available experimental data.
Description
In this context, a tool for the simultaneous analysis of multiple data blocks has been developed
(Network PCA). A development implementation currently exists in MATLAB. We want a port to an
open language/framework, together with user-friendly (GUI) tools to import the data and display the
results. The proposed project involves two parts. The student will design a basic unit test suite for
the port of the core algorithm (to be done by a member of the lab), estimated to be around 15% of
the 80h project time. The remaining 85% of the work should be directed to the development of a GUI
for data import, validation, and visualization.
Language/framework choice can be discussed, as long as it is open and well-established. In particu-
lar, it should have a good ecosystem of data analysis libraries, and be easy to run by end users. With
the requirements for GUI and data plotting, the natural candidate for this -and our preferred choice-
is Python.
Network PCA is still a matter of intense research for us, and modifications to the core algorithm
can be expected. Therefore, continuous integration will be a necessity to ensure a smooth coupling
between ongoing research and code development.
23Projet 23: Visualisation interactive de l’offre de formation en Suisse
Bastien Chopard, Informatique, Science
Context
Développer une carte interactive de la Suisse qui indique les possibilités de formation en informa-
tique. Le but est d’aider les étudiants à choisir au mieux l’université suisse qui leur convient
Description
Il s’agit de développer en javascript une carte de Suisse qui permet d’afficher dynamiquement des
informations relatives aux formations universitaires existantes (p. exemple en informatique), selon
des critères prédéfinis (type de cours, niveau d’encadrement, interdisciplinarité,. . . ).
24Projet 24: Capture de position pour TP de physique
Maggio-Aprile Ivan, DQMP, Sciences
Context
Dans le cadre des Travaux Pratiques de Physique, des expériences de mécanique des corps solides
étudient les trajectoires de mobiles afin de vérifier les lois de conservation de grandeurs physiques
(quantité de mouvement, énergie, centre de masse). L’analyse de ces trajectoires est actuellement
basée sur des systèmes de capteurs optiques ou par traçage direct d’impulsions électriques sur
papier calque. Les techniques modernes d’acquisition et d’analyse des images peuvent permettre
de simplifier la détection des mouvements et d’améliorer grandement la précision des mesures.
Description
Il existe plusieurs moyens de capturer la position d’un ou plusieurs objets en mouvement. Dans
un premier temps, il s’agira d’évaluer quelle est la technique la plus appropriée pour effectuer
l’acquisition des images et la détection précise des positions. Par le fait que l’arrière-plan de la
scène est statique, une simple webcam et la différentiation des images pourrait suffire à la détecter
les mobiles individuellement. Néanmoins, d’autres outils tels les capteurs de mouvements utilisés
par de célèbres consoles de jeux (Kinect de Microsoft-Xbox ou Wii remote de Nintendo par exem-
ple) sont également à envisager. L’essentiel du projet consistera ensuite à interfacer le dispositif
sur du matériel à définir selon les besoins (Raspberry, PC, Mac), puis de développer le code dans
l’environnement (openCV, Windows SDK. . . ) et le langage (Python, Java, C, C++. . . ) appropriés. Le but
du projet est d’identifier la position individuelle de un ou deux mobiles en mouvement simultané,
et d’atteindre une résolution spatiale de l’ordre du millimètre afin d’estimer les vitesses avec le
maximum de précision.
25Projet 25: Interface Web pour contrôle environemental
Iaroslav Gaponenko, Department of Quantum Matter Physics, Faculty of Science
Context
Properties of surfaces at the nanometer scale are highly dependent on the environmental conditions
- in particular temperature and relative humidity. The control over the former is usually achieved by
the use of thermoelectric or resistive heaters. The latter is however more challenging, as it required
to control the content of water in the atmosphere. To address this, a standalone humidity controller
was developed in our laboratory and is based on a computer-controlled mixing of dry and wet gas
flows.
Description
The project is to design a desktop or web-based UI that will interface with humidity controller over the
network through the exposed JSON and REST API. A preexisting user interface prototype programmed
in C# can be used as the basis for the work. Required functionality includes:
• Command and control of the humidity controller settings (relays, control loops, parameters)
• Display and saving of a time-based graph of the acquired parameters
• Web querying of the meteorological data (temperature/humidity/pressure) at the user location
• Scripting engine for user-programmable measurement sequences (for example Javascript)
26Projet 26: Visualisation de changement climatique
Iaroslav Gaponenko, Department of Applied Physics, Faculty of Science
Context
The last decades have seen a rising concern over climate change induced by human activities. Whilst
predictions range from the bearable to the catastrophic, a clear consensus has been reached as
to the inevitable increase of temperatures worldwide. An alternative interpretation of this global
temperature increase is the equator-to-pole climate displacement, giving rise to the existence of
climate twins – regions of similar climate separated in time. The Nonlinearity and Climate Group at
the University of Geneva studies the physical models behind climate displacement, focusing on the
different scenarios and computer-generated predictions.
Description
In the framework of this project, we are seeking to develop innovative ways of visualizing both climate
displacement and climate twins, based on the generated displacement fields. Specifically, a web
application will be developed in order to allow: - Generation of trajectory and retro-trajectory splines
based on the yearly displacement maps (2D matrices in CSV format) - Identification of climate twins
for specific locations - Overlay of climate trajectories (and retro-trajectories) on the map of Europe
27Projet 27: Application Web pour une plateforme de spectroscopie
Emmanuel VARESIO, Mass Spectrometry Core Facility, Sciences
Context
The mass spectrometry core facility (MZ 2.0) is devoted to help researchers answering structural and
quantitative questions covering a wide range of environmental and life sciences applications by using
mass spectrometry (MS) technologies.
For each analysis request a sample record is created, data is acquired and processed, and a certificate
of analysis is sent back to the scientist. These services have different costs that are invoiced to the
lab principal investigator periodically.
Due to the growing demand, a relational database has been designed to replace the numerous Excel
spreadsheets.
Now, a web interface to interact with the database is required.
Description
The facility IT hardware consists of a Synology RS2416RP+ with packages like Web station (web server)
and Docker (hosting the MySQL and PHPMyAdmin containers).
Database has been designed according to the facility needs to track customers, lab principal investi-
gators, samples, type of analysis, reporting, instruments, maintenance, costs, etc. The structure of the
database was done with the help of Yann Manet (system administrator and developer at the School
of Pharmacy).
Currently, the database is populated by importing manually CSV extracts from multiple Excel files.
The project would consist in: 1. Front-end: building a front-end web-based interface for data input
by the facility members (instead of maintaining Excel spreadsheets), using Bootstrap framework (no
web-design necessary), 2. Back-end: data treatment required to check data format validity before
populating the MySQL database, 3. Back-end: the interaction between the database content and MS
data processing using PDF PHP Class would allow the generation “on-the-fly” of the certificates of
analysis.
Programming languages should be: - Server side: PHP oriented object (POO) to construct “names-
paced” Class to be compatible with future extensions (e.g., accounting or authentication modules).
- Client side: JavaScript (if possible oriented object) based on Jquery library (possible integration to
Node.Js in the future).
28Projet 28: Conversion de format de données pour la chimie physique
Hans Hagemann, Physical Chemistry, Science
Context
The SNF requires on the long term a data management with universably readable data. Many instru-
ments provide output files in proprietary formats which are not suitable.
Description
Develop macros to convert systematically FT-IR data from Bruker and other companies to ascii files
allowing for the addition of a set of metadata to provide in the same operation means to retrieve
rapidly any given measurement
29Projet 29: Interface graphique pour la génétique des populations
José Manuel Nunes, GENEV - Anthropology Unit, Sciences
Context
Population Genetics
Description
The project aims to build an OS independent graphic interface for a (set of several) command line
computer program (eventually unified into a single executable or shell script), currently running only
under GNU/Linux, in view to extend the program to a standalone, multi OS, version.
The programs use user provided text files (and existing databases) as input, they have several distinct
options, and they produce diverse textual and graphical outputs. A simplified and limited interface
to the programs already exists on the website hla-net.eu/tools. The current pipeline is mostly in-out,
but at least one step includes a, possibly interactive, validation of the input files.
The development would happen in the UA GNU/Linux servers, where the programs are installed
(with their dependencies) and are currently used. There are several regular users of the programs
within the group that can help in testing and provide advice for the design of the interface. The
contact person is the main developer of the programs and can adapt the current program options
and design.
The result of the project would be a standalone graphical version of the programs. If time allows,
cross compiled versions for MacOS and/or Windows would also be produced.
30Projet 30: Base de donnée pour la physique quantique
Jean-Daniel Bancal, Physique, Sciences
Context
Bell inequalities are powerful tools to identify quantum effects. They play a crucial role in certify-
ing the quantum advantage in tasks such as quantum cryptography and quantum computing. But
knowledge about these mathematical objects is so spread out in the literature, that just knowing
whether a given inequality was already discovered previously is currently a quasi impossible task. A
few years ago we started to work on a platform to list up all known Bell inequalities. A first prototype
version is available online at http://www.faacets.com . But this platform still misses one of its core
components: a robust database.
Description
The aim of this project is to identify the database structure that would be best suited to the faacets
project. It will need to satisfy a number of constraints, including: - indexability: ability to search and
retrieve items - use of standard tools: the database should rely as much as possible on standard tools,
with a maintenance foreseen for more than 10 years - ability to add elements - unique numbering:
each inequality should receive a single reference number - ability to check the consistency of the
database - ability to reconstruct the entire database from raw data - version control: ability to trace
changes to the database
The first part of the project will be to explore how existing databases of scientific or mathemat-
ical objects address (or do not address) our desired requirements. Here are some examples of
such databases: - http://www.internationalgenome.org - https://oeis.org/ - http://www.lmfdb.org/
- http://brauer.maths.qmul.ac.uk/Atlas/v3/ - http://www.liegroups.org/
After identifying the best choice of technology, the second part of the project will be to realize a
prototype implementation demonstrating the various capabilities.
31Projet 31: Application mobile pour l’enseignement de la physique
Luis Darmendrail, Physics & IUFE, Faculty of Science
Context
The use of mobile devices in the teaching of physics and research.
Description
Development of a basic application for noise measurement with GPS tracking in Google Maps and
save the information on a server.
32Projet 32: Visulation de noeuds et de tresses
Bonnier, Sciences, Mathematics
Context
Hopf alegebras and knots/braids in general (do no need to know this objects)
Description
we could draw a knot/braid and the app can recognize some moves (called Reidemeister moves) and
give a TikZ code. And the same thing for Hopf algebra diagrams.
33Projet 33: Editeur de réseaux de Petri pour la vérification de logiciel
Dimitri Racordon, Computer Science, Sciences
Context
The project will touch software verification by the means of a high-level extension of Petri nets.
Description
We developed a high-level extension of Petri nets to support higher-order functions, which allow
us to better model functional programming applications. We now aim at creating a tool to create
models in such formalism, and to perform various simple checks on produced models, such as state
space exploration.
A base project for the tool has already been written in React, using Javascript. The tool already
implements a basic GUI editor, as well as an interpreter for the formalism.
The student will be expected to complete the development of the GUI editor and/or work on verifi-
cation related aspects.
34Projet 34: Analyse d’imagerie cellulaire
Carles Blanch Mercader, Biochemistry, Science
Context
Cells are autonomous genetic machines with the capability to build multicellular structures. Un-
derstanding how the interplay between cell mechanics and genetics orchestrates this process and
ultimately leads to organ formation such as the gut or the spine, is one of the primary goals in
biophysics. In this context, it is crucial to engineering robust and high-fidelity methods to measure
accurately physical variables, but at present, these are lacking in many biological systems of interest.
A key problem in this context is to represent, process, and analyze data in three spatial dimensions.
Description
In vitro studies of single cell layers reveal interesting dynamics: cell layers that form on the inner
surface of elastic capsules buckle after the surface is completely filled and the cells continue to
grow; rolled up cell layers increase their thickness; cell layers adhering to a circular domain rotate
and eventually grow into the third dimension. Time lapse confocal fluorescence microscopy of the
three-dimensional cell aggregates yields large amounts of data that pose a challenge for storage,
processing, and analysis. In this project, we propose to transform pixel based microcopy data for the
systems mentioned above into an adaptive particle representation (Cheeseman et al, Nat. Comms.
9, 5160 (2018)), which compresses the data and facilitates the extraction of cellular features for the
whole cell assembly.
In a second step this data will be processed to extract the positions and sizes of the cell nuclei,
the positions of the cell membranes, the orientation of actin filament bundles and others. These
quantities will then be used for quantifying salient features of cell assemblies like cell shape or
cytoskeletal organization. Furthermore, markers of cell differentiation will be fluorescently labelled,
which permits to link these features, i.e., mechanics, to genetics.
35Projet 35: Outil graphique pour l’analyse de réseau electrique
Ruchi Gupta, University of Geneva, Institute for Environmental Sciences (ISE) and Department F.-A.
Forel for environmental and aquatic sciences (DEFSE), Faculty of Science
Context
Energy storage (ES) can be integrated with intermittent solar and wind technologies to form flexible
hybrid systems and help them supply electricity on demand. The techno-economic performance of
these hybrid systems however depends on a large number of parameters, including the wind and
solar production profile, costs and operational parameters of ES technologies. In this research, we
implement a method to simulate the performance and determine the levelized cost of hybrid systems
for various supply strategies depending on the scale (residential, utility and bulk). This method is
implemented for Switzerland but can be replicated for other geographies.
Description
Levelised cost of energy is the most widely used techno-economic indicator to compare the cost-
competitiveness of RE and ES technologies. In this study, we extend the definition of the levelized
costs to hybrid systems incorporating a renewable energy (RE) supply and an energy storage (ES)
system. The levelized costs of hybrid systems (LCOHS) calculator will aim to calculate LCOHS and
examine which ES technologies can support RE technologies to supply firm electricity (also referred
to as a constant electricity output) depending on the supply duration and scale of deployment. The
calculator will be a graphical interface where the user will be able change the input assumptions
(e.g. costs, capacity, operational parameters of RE and ES technologies). The user can use the slider to
control or enter values directly in the input data sheet. The calculator will return the LCOHS expressed
in EUR per kilowatt-hour (EUR/kWh). This will enable the user to compare various hybrid systems
among themselves and also with conventional technologies (e.g., gas, hydro and nuclear) at various
scales and supply durations. Policymakers, researchers, investors and project implementers can take
the information from this LCOHS analysis and facilitate better policy and investment decisions for
future RE hybrid systems.
36Projet 36: Visualisation de données astronomiques
Krzysztof Nienartowicz, Astronomy, ESA Gaia Data Processing Centre in Ecogia, Versoix
Context
Gaia is a cornerstone mission of the scientific program of the European Space Agency aiming at
making the biggest and most precise catalog of the Milky way ever. The University of Geneva plays
a major for the Processing and Analysis of Variable stars. It requires high performance computation
and Big Data storage capacities to classify and characterize more than 1.6 billion stars of the Gaia
catalog to be delivered at three subsequent Data Releases.
Gaia Geneva team expect to reach the petabyte milestone in 2020. DPCG designed and implemented
both a distributed database cluster based on Postgres-XL and a compute node cluster (Java based)
and number of auxiliary tools to deal with this exceptional mission.
Description
The main goal of the project is to help us on our software visualization tool named VariDashboard.
It is written in Java and is based on the Vaadin Framework that enables to develop web applications
in Java.
The VariDashboard main feature is to access and visualize the results of the computations performed
on the compute cluster by querying the database cluster. We can browse the different execution runs
and visualize time series of observed stars, as well as results of period search and modeling or light
curves. It is also possible to display some skymaps, histograms (1D or 2D) or other of any statistical
property computed on the stars. There is also an embedded groovy engine used for creating custom
plots. We plan to integrated this tool with Apache Ignite distributed execution engine for fast data
analysis.
The main task of the student would be to perform a version upgrade of Vaadin from 7 to 8. For this
he will have to understand the build system based on maven, analyze the impacts of upgrading all
the required dependencies and fix or update the code and graphical components accordingly. The
other task would be to create a new graphical component enabling to browse the catalog hierarchy
and attributes of Gaia data or another extensions to be discussed with the student.
37Vous pouvez aussi lire
