The Statistical Consulting Service at EPFL
The Statistical Consulting Service exists to serve the EPFL community and external entities (companies, government, …) needing statistical advice and help with data-driven problems. We also aim to connect researchers, both internal and external to EPFL, to the expertise of our statistics chairs
A first consultation is free of charge, and during our collaboration we do all we can in order to ensure your success. We can help you identify potential difficulties and opportunities, and aim to provide statistical approaches that are suitable for your problems.
- Research planning and design of experiments: By carefully designing experiments in the early stages of a project, many common difficulties in analysis can be avoided. We offer help with understanding, evaluating and improving your study/project goals, formulating research questions, and anticipating and avoiding data-related issues. We offer help with statistical considerations including choosing a suitable design and sample size, and planning the analysis to link the study design to the suitable analysis tools.
- Data preparation and cleaning: We offer help and advice with data cleaning (quality control, outlier detection, etc.) and preparation (missing values, dimension reduction, etc.). This data wrangling step is essential prior to any further data analysis.
- Data analysis, interpretation, and communication: We offer help with a wide range of data, such as time series and spatial data, or in applications such as in life sciences and industry. We can advise you on the suitability of analyses already performed and discuss possible follow-ups, though earlier consultation is usually preferable in terms of saving time and resources. We can also help with visualisation of the results and drafting the statistical methods sections in your manuscripts, or help in dealing with statistical aspects of reviews.
- Training through crash courses: We offer tailored training in the statistical methods you need, as well as their application in the software R. For statistical training (on theory/methodology/software) please contact us for more information on our courses and their pricing.
Here are some of the external entities we’ve worked with in the past couple of years: Lausanne University Hospital (CHUV), SwissVisio, Loterie Romande, WatchOut Corp, Asterivir.
Internal clients at EPFL include: Center of PhenoGenomics, Integrated Comfort Engineering lab, Laboratory of Integrated Performance in Design, Laboratory of Applied Photonics Devices.
The Statistical Consulting Service will intervene only when the relevant ethics committee (if applicable) has approved your study/project and may decline intervention otherwise. We offer help with statistical needs prior to submitting requests to ethics committees.
We can liaise you with young statisticians and data scientists seeking internships, to help solve data analysis problems in your teams and laboratories, with joint supervision of the intern.
If your company/group proposes internships for master students, we can help with planning the internship and supervise the intern.
Within the limits of our resources we are available to help EPFL students with advice on statistical problems arising in their PhD or MSc projects. We do not help students taking probability and statistics courses in the bachelor and master cycles, and any emails about them will be ignored. Please use this form to book a meeting during our office hours.
To fix an appointment with us, please write an email or use the contact form and make sure to provide a short but exhaustive description of the study. In order to ensure a fruitful collaboration, we will need to know some details about your study, including its (1) design, (2) available data, (3) already performed analysis and (4) your expectation(s). We recommend using the PDF format if possible and including all potentially important details and graphics. See here for an example.