Data and Society

Dear all,

We'll also briefly mention this in the seminar today (Tue), but I also wanted to put this in writing:

Your project reports should be similar in scope to a "short paper" at a CS conference, with a few more details on the data and data processing, as that was one of the main points of the seminar. In terms of length, a "short paper" typically means 4 pages (plus 1 page for references) in a very dense 2-column format. However, more than the page length, the content should serve as an example. Realistically, it will be _easier_ to write a longer paper. Writing a shorter paper requires even more clarity of message and removal of any redundancy.

Here are some examples of short papers you could look at for inspiration/reference:

Host-Centric Social Connectedness of Migrants in Europe on Facebook, https://ojs.aaai.org/index.php/ICWSM/article/view/22224/22003

Automated Hate Speech Detection and the Problem of Offensive Language, https://ojs.aaai.org/index.php/ICWSM/article/view/14955/14805

Analyzing the Targets of Hate in Online Social Media, https://ojs.aaai.org/index.php/ICWSM/article/view/14811/14661

In these reference papers, the data details might be relatively short. That is one important detail where we'd expect more content from you.

You don't have to follow the structure in any of these papers. For general advice on how to write a paper I'd recommend watching "How to Write a Great Research Paper", https://www.youtube.com/watch?v=WP-FkUaOcOM

The single most common "mistake" students make (IMHO) is that they write a list of "here's what we did". Related to this, there is a tendency to include every possible analysis and diagram to show "we did a lot". - But what is better (and harder!) is to write a story where the key points of the story are clearly communicated and also visually illustrated. This also requires a reflection of what is important and what is not.

Happy to discuss more in person later today. And don't hesitate to get in touch with questions.

Best regards,

Ingmar

Dear all,

sorry again for the live demo not going as smoothly as planned. Hopefully it did not make you afraid of docker, it might first seem a bit like trying to crack a nut with a sledgehammer, but both for your studies and for the job market, making the use of docker a good habit of yours is definitely beneficial.

I fixed the dockerfile of today's live demo: https://unisaarlandde-my.sharepoint.com/:u:/g/personal/tiko002_uni-saarland_de/EWsUoY98s0tMvPlXhOP7GPgBaBmUWET-cc0WmhS3PPAX8Q?e=atRWTk

Feel free to use it for your project submission, but don't feel obliged to use it, there are many ways to ensure reproducibility (please try it among your group members before submitting it to us). If you do, feel free to get back to us in case of any questions.

All the best and see you next week,
Till

Dear all,

Tomorrow, Tue, Jan 16, we'll set aside some time for you to fill out an evaluation form about the seminar.

I'm already sharing the link now in case you'd like to see the questions in advance:

https://qualis.uni-saarland.de/eva/?l=147817&p=qsuqs7

Deadline is 31.01.2024

Thanks and see you tomorrow.

Best regards,

Ingmar

Final project presentations:

Jan 23: (i) Team B (Syed and Neshmia),  (ii) Team A (Shams and Moustafa)
Jan 30: (iii) Team D (Amit, Come, and Umama),  (iv) Team C (Tooba, Angela, and Selvyn)
Feb 6: (v) Team E (Sofia and Darian)

Also at https://docs.google.com/spreadsheets/d/1OAg1fTx7SaDQSSlbXTUb7whgloVXDHM6JdmC9HVaZaY/edit#gid=0

Let us know immediately if we got some dates/teams mixed up.

Happy New Year again!

The (few) slides I used are online at https://docs.google.com/presentation/d/1TROOlIP4DbPEmk8opQwA_wrOAO27wJxpDNmeXxKdJOs/edit?usp=sharing. Also linked under "Material" in the CMS.

Some links:

Details of the chocolate study: https://gizmodo.com/i-fooled-millions-into-thinking-chocolate-helps-weight-1707251800

Seminal paper on "Why Most Published Research Findings Are False": https://journals.plos.org/plosmedicine/article/file?id=10.1371/journal.pmed.0020124&type=printable

Good video on how to write a research paper: https://www.youtube.com/watch?v=1AYxMbYZQ1Y

A couple of "semantic", i.e. AI-powered engines for searching academic literature: https://consensus.app/, https://elicit.com/, and https://typeset.io/

Wiki article about the replication crisis in general: https://en.wikipedia.org/wiki/Replication_crisis

Bonferroni correction to guard against p-hacking: https://en.wikipedia.org/wiki/Bonferroni_correction  (this requires being "honest" though about listing all the things that were tried)

Related: https://en.wikipedia.org/wiki/Publication_bias

Gold-standard for meta reviews: Systematic reviews, https://www.cochranelibrary.com/

Details about the (false) claims that some alcohol is better than no alcohol: https://theconversation.com/alcohols-health-benefits-hard-to-prove-but-harms-are-easy-to-document-98813  (due to the fact that non-drinkers often don't drink because they are sick)

Dear all,

On Tue, Jan 2, we will _not_ have a regular, in-person seminar.

Instead, I offer to be online on MS Teams from 10h30-11h30:

https://teams.microsoft.com/l/meetup-join/19%3ameeting_ZDFiNGNjYTUtZmU1Ni00OWRmLThmNDEtZThiNmViMTk2OWMz%40thread.v2/0?context=%7b%22Tid%22%3a%2267610027-1ac3-49b6-8641-ccd83ce1b01f%22%2c%22Oid%22%3a%224f1c03f9-1c68-48b5-a99f-2ad082b5fa1b%22%7d

If you have questions re project planning or other things, don't hesitate to join this online meeting. Note that this is optional though.

If possible, please let me know in advance if you plan to join, so that I don't bother in case nobody is attending. In case the MS Teams link should not work, please send me an email at the time of the meeting and we'll figure something out.

I hope you've had a nice Christmas holiday, and I wish you a good start to the new year.

Best regards,

Ingmar

Dear all,

On Jan 23, Jan 30, and Feb 6 we will have the project presentations. We will have two presentations on each of the first two dates, and one on the last date.

Team A (Shams and Moustafa) have expressed a desire to present on the first date (Jan 23).

@Other groups: Please let us know if you have a preference for a particular date, in particular if that date is _not_ Feb 6. (I assume that most people will want to present last.)

While we cannot guarantee to satisfy your preference, we'll certainly make an effort.

Please let us know your date preference no later than Jan 2. We will announce the data assignment on Jan 3, so that each group has at least three weeks to prepare.

Thanks again for your interest in your seminar, and have some nice holidays!

Best regards,

Ingmar

Dear all,

Please remember to submit a 1-page project plan report by *tomorrow*, Tue, Dec 12, either via the CMS or by email.

You can still pivot later if necessary, but you should show that you and your team have thought about which problem to study, which data to use, and which analysis method to use. At least some awareness of prior related work is required. Of course, all of this will still be "shallow" at this point as you've not yet actually collected or analyzed any data for the project.

On Tue, Dec 19, you will then present your project plans and we will provide feedback to help you work on the project over the coming weeks.

Best regards,

Ingmar 

If you have not yet registered in the LFS then please email  studium@cs.uni-saarland.de  with me (= iweber@cs.uni-saarland.de) in CC. Note that you will have to email from your official student email to avoid impersonation.

Please do so *today* and by tomorrow at the very latest. This was originally due by Nov 27.

Dear all,

As mentioned in the first two lectures (see e.g. https://docs.google.com/presentation/d/1bSA6H0ZBpcIW7brYfLg89HXIxNFN3-vKVijS6UYMMvA/edit#slide=id.p21) you have to register in the LSF (https://www.lsf.uni-saarland.de/) to eventually obtain credits for the course. This was required by Nov 27, but if you do so *today* then that's still ok with us.

Best regards,

Ingmar

If you're taking this course as a 7-credit point seminar then please pick a paper and date for the presentation (from the author's perspective) as soon as possible:

https://docs.google.com/spreadsheets/d/1OAg1fTx7SaDQSSlbXTUb7whgloVXDHM6JdmC9HVaZaY/edit

For this, please leave a comment in the corresponding column, one for the paper and one for the date.

Once we have all the dates/presentations selected, we can then have a second round for choosing a paper to discuss/review. That is only possible for papers that are also presented (on the same day).

You can ignore this message if you're taking the course as a 5-point pro-seminar instead.