vilaweb: Reproducible data analysis for www.vilaweb.cat

The worst tweet. Quantified.

In 2017, Esquire Magazine's Luke O'Neil wrote an article called "How to Know If You've Sent a Horrible Tweet". O'Neil argued that "The Ratio" - that is, the ratio of replies to retweets - is the perfect metric for measuring the badness of a tweet. The higher the ratio, the worse the tweet. When a politician writes a particularly offensive or stupid tweet, very few people retweet it, but very many reply to it.

FiveThirtyEight, a data journalism group in the United States, did an analysis of American politicians' twitter ratio in late 2017. They showed that, despite popular belief, there are American politicians who are even more inept at Twitter than Donald Trump. Hard to believe, but true.

Let's explore "The Ratio" among Catalan and Spanish politicians.

Who are Catalonia and Spain's worst tweets and tweeters?

We gathered tweets for 58 well-known Spanish and Catalan politicians, and calculated tweet-specific reply-to-retweet ratio, as well as a few aggregate statistics. We used all tweets for the three year period from 2016 through 2018, removing those with fewer than 10 interactions.

The below is a "ternary" diagram showing the overall relationship between replies, retweets, and likes. Each point is a tweet from one of the 57 politicians (ie, 279390 dots). A well-received tweet is situated to the right and top. A poorly received tweet is situated in the bottom left.

Most tweets are well received; that is, they have a low ratio of replies to retweets. In fact, of the 279390 tweets examined, only 6986 had a ratio of greater than 1 (ie, more replies than retweets) - that's just 2.5%.

It's hard to make comparisons between politicians on three dimensions, so, let's reduce our analysis to just one. The below chart shows the percentage of all interactions (likes + retweets + replies) which are replies. A high percentage generally indicates a poorly received tweet, and a low percentage generally indicates a well received one.

The top spots are dominated by members of mainstream Spanish parties: the Popular Party and the Socialists. The two worst performers are Miquel Iceta (PSC) and Xavier Garcia Albiol (PP), for whom 19.5% and 19% of all interactions are replies, respectively. At the other end of the spectrum, the most well received politician is Jaume Asens, for whom only 2.4% of interactions are replies.

Which tweet had the highest total number of interactions in the last 3 years from the politicians in question? That honor goes to Oriol Junqueras whose November 2 2017 (the day he entered prison) tweet garnered 43,096 retweets, 71,033 likes, and 5,950 replies (a total of 120,079 interactions).

The below table shows the 10 tweets from 2016-2018 with the greatest number of interactions.

| rank| username | date | tweet | interactions| replies_count| retweets_count| likes_count| |-----:|:------------------|:-----------|:---------------------------------------------------------------------------------------------------------------------------------------------|-------------:|---------------:|----------------:|-------------:| | 1| junqueras | 2017-11-02 | Feu cada dia tot allò que estigui al vostre abast pq el bé derroti al mal a les urnes el 21D. Dempeus, amb determinació i fins a la victòria | 120079| 5950| 43096| 71033| | 2| krls | 2017-10-11 | Demanes diàleg i et responen posant el 155 damunt la taula. Entesos. | 115358| 9039| 43114| 63205| | 3| krls | 2017-10-27 | Catalunya és i serà terra de llibertat. Al servei de les persones. En els moments difícils i en els moments de celebració. Ara més que mai | 111908| 6588| 35452| 69868| | 4| pablo_iglesias_ | 2017-10-03 | Como presidente de un grupo parlamentario que representa a más de 5 millones de españoles, le digo al Rey no votado: no en nuestro nombre | 103184| 11789| 34272| 57123| | 5| adacolau | 2017-10-03 | Ninguna solución. Ninguna mención a los heridos. Ninguna apelación al diálogo. Discurso irresponsable e indigno de un jefe de estado #Rey | 97663| 5226| 36942| 55495| | 6| marianorajoy | 2017-10-27 | Pido tranquilidad a todos los españoles. El Estado de Derecho restaurará la legalidad en Cataluña. MR | 97436| 18155| 30244| 49037| | 7| krls | 2017-11-06 | En llibertat i sense fiança. El nostre pensament és per als companys injustament empresonats per un Estat allunyat d la pràctica democràtica | 97336| 4038| 30283| 63015| | 8| krls | 2018-04-05 | Ens veiem demà. Moltes gràcies a tothom! #LlibertatPresosPolítics 🎗 pic.twitter.com/quG18LKvQp | 87701| 4867| 25411| 57423| | 9| pablo_iglesias_ | 2017-10-01 | Porrazos, empujones, ancianas arrastradas. Lo que está haciendo el PP a nuestra democracia me repugna. Corruptos, hipócritas, inútiles | 87602| 8921| 32895| 45786| | 10| adacolau | 2017-08-17 | Barcelona ciudad de paz. El terror no conseguirá que dejemos de ser quien somos: ciudad abierta al mundo, valiente y solidaria | 82250| 1843| 26242| 54165|

The below chart shows the total number of interactions (retweets + replies + likes) between 2016 and 2018 of all politicians examined. Gabriel Rufian is by far the most impactful, followed by Carles Puigdemont.

Which tweet had the highest number of replies as a percentage of all interactions (excluding those with fewer than 500 interactions)? The honor goes to the Catalan Socialists whose May 2017 tweet called for a choice to be made between ballots and lunch for children. The tweet had 508 replies, with only 58 retweets and 45 likes (a reply-to-retweet ratio of 8.7).

The below table shows the 10 tweets with thehighest number of replies as a percentage of all interactions (excluding those with fewer than 500 interactions).

| rank| username | date | tweet | % replies of all interactions| replies_count| interactions| |-----:|:-----------------|:-----------|:-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|------------------------------:|---------------:|-------------:| | 1| socialistes_cat | 2017-06-28 | 💡Més suggerències pel govern després de quedar desert el concurs de les urnes 👇 pic.twitter.com/mzkuAlL2HA | 83.14239| 508| 611| | 2| pablocasado_ | 2016-06-11 | Aclarado el error con @DLasAmericas, borro el tuit para evitar más confusión. Censuro estas situaciones pasen donde pasen. #RevocatorioYA | 82.67288| 1064| 1287| | 3| ppopular | 2017-02-23 | "No hay apellidos ni cargos que valgan para la Justicia en nuestro país: quien la hace, la paga" @ALevySoler #LN24 | 77.88235| 662| 850| | 4| miqueliceta | 2018-03-21 | Els independentistes ni tenen pla de govern, ni candidat de consens, ni els vots suficients http://www.socialistes.cat/ca/noticia/els-independentistes-ni-tenen-pla-de-govern-consensuat-ni-candidat-de-consens-ni-els-vots-suficients-1#.WrKSyRwD8aw.twitter … | 75.34247| 990| 1314| | 5| ppopular | 2018-04-12 | 📽 @martinezmaillo 👉 "Somos un partido de gobierno y de acuerdos, como lo hemos demostrado con los pactos con los funcionarios, sindicatos, la equiparación salarial... Otros, en cambio, lo único que hacen es levantarse de las mesas de negociación". pic.twitter.com/b1X3LCDLTO | 75.00000| 507| 676| | 6| ppopular | 2017-10-13 | 💬 @JavierMaroto: "La reforma de la Constitución es un acto de generosidad de Rajoy" http://www.pp.es/actualidad-noticia/maroto-reforma-constitucion-es-un-acto-generosidad-rajoy … | 73.97820| 543| 734| | 7| miqueliceta | 2018-10-15 | La commemoració del 78è aniversari de l’assassinat del president Lluís Companys és un fet que ens relliga amb la història i han de ser motiu de reflexió sobre quins són els millors camins per preparar el futur del país. pic.twitter.com/bo1DTyJa5Z | 73.53407| 928| 1262| | 8| ppcatalunya | 2017-10-30 | 'Cerrando todas las embajadas en el exterior de la Generalitat nos vamos a ahorrar 25 millones de euros' @Albiol_XG #DUIDespuESP | 73.42939| 1274| 1735| | 9| j_zaragoza_ | 2018-06-28 | El independentismo ha de retomar el camino de la política, hablar, ceder, respetar la ley y si hace falta cambiarla. | 73.12312| 487| 666| | 10| socialistes_cat | 2018-03-18 | Avui molts ens estem aplegant indignats que encara no hagi govern a Catalunya amb la necessitat que ben rapidament acabi la vigència del 155 a través de l'elecció d'un President o una Presidenta. @miqueliceta #SENY18M #CatalunyaSomTots pic.twitter.com/rM23qwPz55 | 72.95720| 750| 1028|

The "honor" of "worst" tweeter (ie, the person whose replies-to-retweets ratio is highest) goes to Miquel Iceta. He is followed closely by Andrea Levy Soler, then Xavier Garcia Albiol, Ines Arrimadas, Pedro Sánchez, and Pablo Casado. Though our dataset had nearly as many sovereigntists (26) as unionists (30), the top spots in terms of the Ratio all went to unionists.

The twitter "Ratio" (the ratio of replies to retweets) is generally higher among pro-union politicians than pro-independence politicians. Socialists and members of the PP have the highest "Ratios" (replies to retweets). The most impactful tweet, as well as the most impactful tweeter, are both members of ERC. Despite having a far smaller base than Spanish national politicians, Catalan-specific and Barcelona politicians dominated the top 10 most "impactful" tweets (measured in terms of total number of interactions).

The "Ratio" is a fun metric to analyze, but of questionable scientific value. It reflects the "badness" of a tweet, but "badness" is a function of both (a) the tweet itself and (b) the engagement of the audience. It is a metric grounded in the "wisdom of crowds", but the extent to which crowds are wise - or not - is not entirely clear.

What is clear from this analysis is that among Spanish and Catalan politicians, unionists' tweets are particularly poorly received. Miquel Iceta had the highest ratio of replies-to-retweets, and his party (the PSC) had the tweet with the highest percentage of replies. Sovereigntists, on the other hand, had the most impactful tweet (Junqeras), the most impactful tweeter (Rufian) and - on average - much lower ratios of replies-to-retweets than unionists.

The unionist-independentist divide is not unlike the Republican-Democratic divide identified by the analysis team at FiveThirtyEight. The below shows tweets from the two main American political parties.

If we compare this with tweets from Catalan and Spanish politicians, we find a similar pattern: like the American Republican party, Spanish unionists (red) tend to have a higher replies-to-retweets ratio and greater likes than retweets, whereas Catalan independentists (blue) - like the American Democratic party - tend to have a lower ratio of replies-to-retweets and more retweets than likes.

Do twitter Ratios really matter? No. But perhaps they are relevant insofar as they touch on an underlying - and only infrequently discussed - subject in Spanish and Catalan politics: who cares more?

Much ado is made of the 50-50 union-independence split in Catalonia. But in the case of gridlock, the eventual victor is not necessarily decided by the marginal changes in majorities/minorities, but rather by the extent to which the constituency is passionate about the topic. In other words, just because the population is split evenly on an issue does not mean that both sides of the split care evenly about that issue - and the side that cares more will generally win.

In the case of Catalan independence, who cares more about the issue - those in favor or those opposed to independence? Measuring "care" is hard, but those in favor of independence are more interested in politics than those opposed, and interest is a good proxy for care.

By the same token, those in favor of independence exhibit higher levels of engagement in politics than those opposed. These two factors, combined with the fact that unionist politicians in Catalonia receive generally lower ratings than their independentist counterparts, suggests that those in favor of independence care about the subject more than those opposed. This may explain why unionists' tweets have such worse "Ratios" than those of independentists.

So, are Miquel Iceta's tweets so poorly received because he is a particularly bad tweeter? Probably not. Rather, his "Ratio" is so high because the population which is interested and paying attention is largely opposed to his political views (and opponents are more likely to reply than to retweet), and the people who agree with him (ie, PSC supporters) themselves say that they are not very well-informed about what is happening in politics.