Share your favorite coding skills and tools with your friends and colleagues in friendly, no-pressure work-alongs.
Get together to work on your coding projects, help each other out and share your work.
Meet new people in your field, organization or community - and find out what we can do when we work together.
Session leader: Dmytro Perepolkin
Choosing best model for your machine learning problem may be daunting task. H2O AutoML simplifies the process of setting up the complex machine learning modeling pipelines to just a few lines of code. You will learn how to make your own stacked ensemble in the first 30 min of the workshop. For those who want to stick around for the rest of the evening, we will go a bit deeper into specification of individual models (GLM, GBM, Neural Networks, etc) and understanding of model performance.
After attending this workshop, learners will be able to set up their own machine learning pipeline in R from simple data preparation to specification of the classification or regression model, training, evaluation of the model using validation set, and prediction on the test data. Users will be also introduced to the concepts of ensembles and stacking and will be able to specify and run automatic machine learning models, analyze and pick best models and use them for predicting the outcomes.
This workshop is repeating the session held at recent UiO Research Bazaar.
About the speaker: Dmytro is a passionate data educator and data analyst at Equinor. He comes from econometrics and decision analysis background and has particular passion for spreading data literacy and probabilistic thinking in the society. Dmytro co-organizes Oslo UseR! Group and is frequently seen at Caprentry@UiO events at University of Oslo.
Session leader: Valeria Vitelli, Lene Norderhaug D., Aurora V., Athanasia Monika Mowinckel, Isabelle V.
Bayesian methods for rank and preference data - from recommendation systems to cancer genomics
Ranking items is crucial for collecting information about preferences in many areas, from marketing to politics. The interest often lies both in producing estimates of the consensus ranking of the items, which is shared among users, and in learning individualized preferences of the users, useful for providing personalized recommendations. In the latter task it is particularly relevant to have posterior distributions of individual rankings, since these can provide an evaluation of the uncertainty associated to the estimates, and thus they can avoid unnecessarily spamming the users.
I will present a statistical model which works well in these situations, and which is able of flexibly handling quite different kind of data. The Bayesian paradigm allows a fully probabilistic analysis, and it easily handles missing data and cluster estimation via augmentation procedures. Interestingly, this Bayesian framework has also proved to be useful for genomic data integration, since typically heterogeneous microarray data are available from different sources, and their combination allows both to gain statistical power and to strengthen the biological insight.
Valeria Vitelli holds a PhD in statistics from Politecnico di Milano, Italy. She was a postdoc at Ecole Centrale Paris for a year, within a research group funded by Eléctricité de France working on big data problems in the energy sector. She then moved to the University of Oslo, where after a 5 years postdoc period funded from the Norwegian Cancer Society she became associate professor in September 2018. Her experience spans over several areas of mathematics and statistics, including functional data analysis with applications in physiology, machine learning (describing people mobility in dense urbanized areas from mobile phone data), and finally statistical genomics of cancer.
Session leader: Dmytro Perepolkin and Andrea Chi Zhang
A personal website or blog is a great place to record, share and communicate your thoughts and learning process, however the whole procedure might seem complex to those without relevant experience. An R package developed by Yihui Xie, blogdown, makes it easy to build a website using R Markdown and Hugo. Andrea will talk about what is blogdown and why you should use it, and walk you through the basic steps to publish (for free!) your website using Netlify and rbind.io.
About the speaker: Andrea (Chi Zhang) is doing her PhD at Department of Biostatistics, University of Oslo, working on Electronic Health Record data. She is interested in time series data related to health, and is eager to learn new tricks in R and share what she has learnt in her newly built website using blogdown. Outside work she is an active supporter of non profit organisations.
Session leader: Radovan Bast, Anne Fouilloux and Sabry Razick.
The aim of this course is to demonstrate to and familiarize the workshop participants with best practices and tools in modern research software development. The main focus is on professional tools for efficiently developing and maintaining research software. Since most research code is developed in a collaborative setting, we will discuss tools and workflows which facilitate this process. Most of the content is also relevant to a single developer.
If you are writing code that is used in research, then this course is for you. If you develop research code and you know all the tools already, join us as a helper! It’s fun, and you always learn something new about a subject by teaching it.
Did you see the recent announcement, that registration is now open for the CarpentryConnect Manchester conference? The event is an opportunity for members of The Carpentries’ global community of instructors and anyone else with an interest in helping researchers to improve their computational skills, to get together for a few days to exchange ideas and learn some new things. With an exciting programme (still under development, no spoilers!) including sessions focussing on developing new and existing curricula, increasing the sustainability of research software, methods and tools for better teaching, and more, the conference will have as much to offer those who are newcomers to The Carpentries as it does to those who have attended a workshop, helped or taught at one or two, or been involved for years.
Still not sure what to expect? Following the success of the inaugural CarpentryCon in Dublin last summer, several of the attendees wrote about their experiences describing some of the workshops and breakouts, reflecting on their experiences, and generally celebrating the chance to bring such a diverse and dispersed community together (https://carpentries.org/blog/2018/06/carpentry-con-report/). The organising taskforce of CarpentryConnect Manchester, chaired by Aleks Nenadic and including several organisers of the event in 2018, has every intention of creating a sister event with a similar atmosphere of creativity, inclusivity and optimism, and a love of using sticky notes for almost everything.
We hope that the structure of the programme (https://software.ac.uk/ccmcr19/programme), with bursts of lightning talks to spark conversation, long breaks to create space for deeper discussion and networking, and an emphasis on workshops, tutorials and breakout sessions, reflects the importance placed on community building and professional development that I have been so inspired by at previous events of both the Research Software Engineers and Carpentries. The conference offers a chance to learn about and discuss the latest developments and challenges in computational teaching and research software development, to access training in relevant skills, tools, and resources, and to get to know other members of this warm and welcoming community. For more detail on the programme, keep an eye out as we announcement our keynote speakers over the coming weeks.
We’d love to see plenty of new faces at the conference, as well as bringing together as many Instructors, Trainers, Mentors, and Maintainers, as possible from the UK, Europe, and across the world. So, if any of the above sounds interesting to you, we’d love for you to go ahead and register. If you do so before 10 April 2019, you’ll be able to take advantage of the Early Bird discount.
Our call for proposals is still open (https://software.ac.uk/news/carpentryconnect-manchester-2019-call-proposals).
You can submit an abstract or your suggestions using the submission form until 1 March 2019. The notification of the outcome of your application will be sent out by 1 April 2019.
We look forward to seeing you in Manchester in June!
University of OsloInterests:
University of OsloInterests:
Marketing and Social Media Coordinator
The National Library of NorwayInterests: