Due to the circumstances regarding the spread of COVID-19, all Mu Sigma meetups are postponed until further notice! Please keep up to the updates in our Telegram group, we will continue our work as soon as the situation goes well again.
We are preparing a series of meetups to the topics of (1) Reproducibility in statistical science, (2) Operations research, and (3) Causality. Please stay updated, the details will be anounced in the second week of February as latest.
Feel free to join our group on meetup.com for further events: https://www.meetup.com/Mu-Sigma/
... as well our Telegram-chat with free discussions: https://t.me/mu_sigma
Mu Sigma is a group of people that are highly interested in the theory and application of statistical science. It has been organized by several students and alumni of Humboldt University since the spring of 2019. We now have members of different experiences and backgrounds, including leading data scientist from several Berlin companies. The main goal of Mu Sigma is to gain a better understanding of statistics by bringing motivated people together, sharing our experience, and working through papers or other sources of information.
The discussions in our meetings have a rather broad scope but mainly revolve around the following topics:
- Epistemology: What is knowledge and how it should be obtained? What is the role of statistics in that process?
- The proper application of statistical methods in research and working environment
- Newest trends in data science
- Ongoing replicability crisis (especially in social sciences)
- Technicalities of particular statistical methods and software
An excellent way to understand what Mu Sigma is about is to check out the papers and the protocols of our last meetings:
- The list of the papers we've already read is available in our meetup archive
- The list of documents proposed to read in the future is available here
- The detailed protocols or our meetings are available in the following folder
If you have any further questions and like to join us, please read the frequently asked questions (FAQ) below.
- How can I contact you?
- What is your timetable?
- How and where can I attend your meetup?
- What background should I have to participate?
- Does Mu Sigma have a focus on any field?
- Is this another ML or AI meetup?
- Do you organize anything else except the meetups?
- Which programming language do you work with?
- Are you guys bayesians or frequentists? Data or algorithmic modelling?
- Do I need to read the paper before attending the meetup?
- How do I vote/propose to read a paper?
- Why are you called Mu Sigma?
To contact us, feel free to write us a message on musigma.berlin@gmail.com or in our public Telegram group.
We meet every two weeks usually Wednesday with a soft start at 18:30 and hard start 19:00. We do a poll about what to read beginning Friday after the last meeting, its results are announced on the following Monday. We will send a reminder the Monday before the meetup. If you are confused now, recheck this document, the next meeting is at the beginning.
Please contact us on Telegram or via Email and do a short introduction of yourself. After getting to know you, we will tell you the location of our next meetup. We don't publish the place we meet beforehand, but it usually is in the central areas of Berlin.
The attendance is, of course, free of charge.
To be able to follow the discussion, attendees are expected to have at least some knowledge of statistics, programming, and mathematics. A solid college undergrad level should generally suffice, although some papers often go way above this knowledge. Apart from that, we do not have any special requirements except the willingness to participate and learn actively.
It is necessary to note that Mu Sigma was founded by students of psychology, sociology, and economics, among others. Thus, there is a slight focus on the application of statistical methods in behavioural sciences. Nevertheless, we sometimes discuss papers from other fields, such as biology or physics, or "general" statistics.
Except for the point stated above, we do not position ourselves as a meetup that is about a particular school of methods. We instead look how to use all available methods and tools to get the necessary results in data science. We are also not a "meetup" in the classical sense and do not organize networking or hiring events. Nevertheless, there's always a chance that you meet someone who helps you in your professional career later.
So far, the regular meetings are the only thing we organize right now. In the future, we might expand our activities to inviting speakers, providing education, doing collective projects, and so on. If you have any suggestions or offers, do not hesitate to contact us.
There is no single programming language required. As far as our last meetups are concerned, most of the participants work either in R or Python.
Again, we prefer not to go for a single approach but rather discuss them all together. We do like some controversies as long as only the content of the debate is heated. Participants with all views and backgrounds are welcome.
Yes, the discussion of paper is an integral part of our meetup, so everyone should at least skim through the document beforehand. However, you are not expected to understand in details the technicalities of every paper.
We vote over the weekend after the last meetup via Telegram on the proposals found here. To propose a paper add a new file in the proposals folder named [first-author]-[year]-[keyword1_keyword2_..._keyword5].md
(of course, via github pull request, we'll promise to include every proper proposal). Before the vote, we rebuild proposals.md via make
which will include your proposal in the main file. From time to time, we move longtime unaccepted proposals into the archive folder. Your proposal should include concise answers to these questions:
- What is the topic?
- Why is the topic of the paper interesting to the group?
- Why is this exact paper/medium the best for this topic?
- How much preparation time is needed? (If it is >1h, provide an idea of how to tackle it [skimming, over multiple meetups, preparing a presentation, etc.])
- Provide a frictionless way to access the medium (provide APA citation, DOI, link or similar).
Take a look at the existing proposal and use the them as a template. Provide a correct YAML-header.
We didn't want to go with yet another common name like "Berlin Data Science Meetup" and decided to pick something unique instead. We have chosen those two Greek letters just because they are often met in statistics (the population mean and variance, respectively) and sound good in combination. Apart from that, there's no additional meaning to the name of our society.