arXiv Twitter Feeds (Mathematics)

Twitter Accounts (Mathematics)

Twitter, in principle, permits automated feeds for public or community benefit (source: Twitter official site):

Automating Tweets by streaming external information

Feeds for community benefit: With express user consent, you are permitted to create feeds for what Twitter deems to be community benefit, or to provide information to a niche group of users such as local weather feeds or transit information, as long as the posts comply with the Twitter Rules.

Twitter also specifically mentions cases like blog feeds. Furthermore,

Automating multiple accounts

... Automating multiple accounts for what Twitter deems to be community benefit is permitted.

Therefore I create the following Twitter accounts and a bot that daily tweet abstracts (summaries) of new submissions into the various mathematics categories of the arXiv e-print archive.

Some academicians and journals actually recommend more use of Twitter for their colleagues; unfortunately, there aren't too many Twitter accounts worth following for researchers in pure or applied mathematics. As arXiv is definitely one of the most important resources used by mathematicians, I am hoping that easy availability in Twitter will make it easier for those who are more attuned to the ambience of Twitter to streamline it with their professional activities also. This may be beneficial for Twitter also, attracting a more serious user base.

Mathematics Subject Classification arXiv code Twitter Handle
Algebraic Geometry math.AG @rXiv_math_AG
Algebraic Topology math.AT @rXiv_math_AT
Analysis of PDEs math.AP @rXiv_math_AP
Category Theory math.CT @rXiv_math_CT
Classical Analysis and ODEs math.CA @rXiv_math_CA
Combinatorics math.CO @rXiv_math_CO
Commutative Algebra math.AC @rXiv_math_AC
Complex Variables math.CV @rXiv_math_CV
Differential Geometry math.DG @rXiv_math_DG
Dynamical Systems math.DS @rXiv_math_DS
Functional Analysis math.FA @rXiv_math_FA
General Mathematics math.GM @rXiv_math_GM
General Topology math.GN @rXiv_math_GN
Geometric Topology math.GT @rXiv_math_GT
Group Theory math.GR @rXiv_math_GR
History and Overview math.HO @rXiv_math_HO
Information Theory math.IT
= cs.IT @rXiv_cs_IT
K-Theory and Homology math.KT @rXiv_math_KT
Logic math.LO @rXiv_math_LO
Mathematical Physics math.MP
= math-ph @rXiv_math_ph
Metric Geometry math.MG @rXiv_math_MG
Number Theory math.NT @rXiv_math_NT
Numerical Analysis math.NA @rXiv_math_NA
Operator Algebras math.OA @rXiv_math_OA
Optimization and Control math.OC @rXiv_math_OC
Probability math.PR @rXiv_math_PR
Quantum Algebra math.QA @rXiv_math_QA
Representation Theory math.RT @rXiv_math_RT
Rings and Algebras math.RA @rXiv_math_RA
Spectral Theory math.SP @rXiv_math_SP
Statistics Theory math.ST @rXiv_math_ST
Symplectic Geometry math.SG @rXiv_math_SG


All of the above is also available at once as a: Twitter list.

The first column of the above table is, btw, the URL of the UC Davis Front for each category and the second column a link to the original Cornell site.

The profile images and banners are screenshots (cropped/modified) from the relavant arXiv page.

The accounts all follow each other. This is merely for ease of users to find all others from one.

Follow these accounts to daily get the arXiv submission information for the respective categories into your Twitter feed. To avoid duplicate Tweets, each paper is Tweeted only from the handle corresponding to its primary category. If it belongs to more categories, then the tweet from the primary category's handle is re-tweeted. Thus to avoid seeing duplicates, or even to avoid multiple categories, it is enough to turn off retweets for the corresponding handles.

Such Twitter feeds created by various individuals for arXiv already existed and not just for mathematics; some were even formally studied. The existing feeds meant for specifis subject areas, for example the ones operated by So Okada, have had generally a followers count of below two or three hundreds after operating for several years. This is in contrast to the thousands of followers of the MathPapers, bioRXiv, or the bot @OverheardOnAph mining comments from TeX sources of arXiv preprints for astrophysics, after shorter operation periods. It remains to be seen how popular this new attempt will be. The required code was written in Python from scratch and later while google-searching to test how it fared, I discovered that others had already published Twitter bot source codes for similar purpose to Github, years ago.

What is new with my feeds here is that they compile the article abstracts in LaTeX and embed them as images into the tweets. These can be enlarged and read with a click on the desktop browser, or a touch on the mobile or tablet apps. The tweeting time is about half an hour after arXiv's submission availability time which is 8pm EDT on Sunday to Thursday.


To keep with the Twitter ambience, feedback may be provided via tweets to @NotionsandNotes: .

For e-mail communications, address is george @ this site's domain name (without www).

A number of small bugs were already caught and fixed. Notification of more will be appreciated.

Pro Forma

Creating these many accounts involved some nonzero work. Automated tweeting of multiple streams is generally frowned upon by Twitter as there are a lot of abuses. After all the effort, both imaginative and repeatitive, the thought of Twitter deeming this abuse and summarily banning gives me the shivers. In brief, each account requires a separate e-mail address and more importantly a phone verification, and one phone can verify only at the most 10 accounts, and in case of suspicion of automation or abuse, Twitter asks to re-verify the phone by an SMS verification code. I have gone through all these steps for each account after the first one and want to avoid the effort turning out to be futile. Therefore, I am writing a piecemeal compilation of checks of their various conditions as an assertion/attestation of good faith.

Permissions of arXiv

The arXiv site asserts: arXiv supports and participates in the Open Archives Initiative (OAI). arXiv is a registered OAI-PMH data-provider and provides metadata for all submissions which is updated each night shortly after new submissions are announced.


Metadata for arXiv articles may be reused in non-commercial and commercial systems.

Indeed, arXiv also provides in detail their APIs for fetching this information which many third parties have used so far for a number of projects. arXiv is aware of these and acknowledges and invites more. Moreover various apps created by individuals for phones and tablets to access arXiv submission info.

The twitter ids above do not include the full name 'arXiv' and therefore does not conflict with arXiv name copyrights if they ever decide to operate Twitter accounts by themselves. Each author submits the papers to arXiv for irrevocable and worldwide dissemination under Creative commons licenses. Of the submissions that are multi-page articles, we only tweet the abstracts that are at most one paragraph and is specifically intended by the author for public dissemination as a short summary. Therefore no rules are violated and implicit permission exists for everything done here and all the requirements of the Twitter rules in regard to content and rights of arXiv (trademark) and the copyright holders (the authors).

Twitter automation Rules

This effort does not violate any Twitter rules and guidelines to our best knowledge. Ultimately, Twitter has to make a decision but we check the following gotchas relevant to us, quoted from the Twitter Help Center:

It seems Twitter actually invites and encourages bots by developers when these are not harmful or spam: We love developers and are constantly amazed by the applications and services created on the Twitter platform. We hope our effort qualifies under their criteria.

With regard to user permissions, this app does not access any user accounts and only tweets into pre-fixed accounts as listed above. Therefore specific user permissions are not an issue, except for permission from arXiv, which is verified above.

Automating multiple accounts

Creating and/or automating serial (or multiple) accounts for overlapping use cases is prohibited. For more information about serial or multiple accounts, please see the Twitter Rules.

Though the multiple accounts above are 'overlapping' in the sense of using arXiv feeds, they are also distinct, as they correspond to published papers in distinct and separate categories. We hope our use is allowed with this clarification. Twitter also states in the next paragraph:

Automating multiple accounts for what Twitter deems to be community benefit is permitted.

The arXiv itself and everybody who publishes to arXiv are serious researchers who make their work free for community benefit, and intend it to be disseminated as much as possible. These automated Twitter accounts are nothing but a continuation of this good intent to avail of the visibility and amplification features of Twitter and therefore should fall under the community benefit criterion of Twitter.

Automated Retweeting

Automation of Retweets often leads to spam and other negative user experiences; therefore, Retweeting in a bulk or automated manner is prohibited. Automated Retweeting is permitted for apps or accounts that Twitter deems to provide a community benefit.

We have argued for community benefit in the previous paragraph. Twitter rules prohibits: Retweeting spam updates, repeatedly posting duplicate links, etc. There is some very limited automated re-tweeting, but it is only when a few papers fall into multiple categories and then, it is tweeted in the primary category and re-tweeted in any additional categories. We believe this a very logical thing to do and far from increasing spam, it actually decreases redundancy.

Regarding other prohibitions of Twitter, such as automated favoriting, deleting, direct messaging, following, unfollowing, replying, mentions, tweeting to trending topics, automation based on completed user actions, etc., none of it applies to our case.

© 2015 -- 2016 George S. of