The newest increasing of the limit tweet size provides for an interesting opportunity to check out the the effects out of a leisure from duration limitations to your linguistic messaging. And interestingly, exactly how performed CLC change the design and phrase usage inside tweets?
The need for a discount off term diminished blog post-CLC. Therefore, all of our first theory claims you to definitely blog post-CLC tweets include apparently faster textisms, such as abbreviations, contractions, signs, and other ‘space-savers’. In addition, i hypothesize that CLC inspired new POS framework of the tweets, with seemingly even more adjectives, adverbs, articles, conjunctions, and you may prepositions. These types of POS classes carry additional info in regards to the situation being described, new referential disease; such as for instance options that come with agencies, this new temporal purchase out of incidents, cities off events otherwise things, and causal relationships between events (Zwaan and you may Radvansky, 1998). So it structural changes and entails you to phrases could be stretched, with terms for every phrase.
Gligoric mais aussi al. (2018) compared pre and post-CLC tweets with a period of just as much as 140 letters. It discovered that pre-CLC tweets within this reputation variety comprise apparently more abbreviations and contractions, and you can a lot fewer chosen blogs. In the present data, i put a unique strategy one contributes subservient value to the earlier conclusions: we did a material research to your an effective dataset of around 1.5 billion Dutch tweets as well as every selections (we.e., 1–140 and you will 1–280), in place of shopping for tweets in this a certain character diversity. This new dataset constitutes Dutch tweets which were authored ranging from , this basically means 14 days before and two weeks after the latest CLC.
We did a broad data to analyze changes in the number out-of letters, terms and conditions, phrases, emojis, punctuation scratches, digits, and URLs. To check on the original theory, i performed token and bigram analyses so you’re able to locate most of the alterations in the brand new relative frequencies out-of tokens (we.e., personal terms and conditions, punctuation marks, amounts, special letters, and you will symbols) and bigrams (we.e., two-keyword sequences). These types of alterations in relative wavelengths you are going to following be utilized to recoup the latest tokens that 
were particularly influenced by the newest CLC. While doing so, a good POS study try did to check on the next hypothesis; that is, whether the CLC impacted the POS framework of the sentences. A typical example of for every single investigated POS classification was showed inside Dining table 1.
Equipment
The information collection, pre-handling, quantitative studies, numbers, token study, bigram study, and you will POS data were did using Rstudio (RStudio People, 2016). The fresh new Roentgen packages that were put is: ‘BSDA’, ‘dplyr’, ‘ggplot’, ‘grid’, ‘kableExtra’, ‘knitr’, ‘lubridate’, ‘NLP’, ‘openNLP’, ‘quanteda’, ‘R-basic’, ‘rtweet’, ‘stringr’, ‘tidytext’, ‘tm’ (Arnholt and you can Evans, 2017; Benoit, 2018; Feinerer and you can Hornik, 2017; Grolemund and you may Wickham, 2011; Hornik, 2016; Hornik, 2017; Kearney, 2017; R Key Class, 2018; Silge and Robinson, 2016; Wickham, 2016; Wickham, 2017; Xie, 2018; Zhu, 2018).
Ages of desire
The fresh new CLC happened to the during the a great.meters. (UTC). The fresh new dataset constitutes Dutch tweets that have been authored within fourteen days pre-CLC and two months blog post-CLC (we.age., regarding ten-25-2017 in order to eleven-21-2017). This era is subdivided on the week 1, week dos, few days step three, and you will week cuatro (see Fig. 1). To research the result of your own CLC i compared what usage in the ‘month 1 and you will times 2′ with the words use within the ‘few days step 3 and you may week 4′. To distinguish the fresh new CLC feeling regarding sheer-experiences consequences, a running analysis was developed: the real difference in code usage ranging from times step one and you can day dos, referred to as Baseline-broke up We. Additionally, the newest CLC might have started a trend regarding the vocabulary usage you to advanced much more users turned into familiar with brand new limitation. That it development could well be revealed from the evaluating day step three that have few days 4, named Baseline-split II.
Moving mediocre and you can practical mistake of one’s profile usage throughout the years, which shows a boost in character utilize article-CLC and you may an extra raise anywhere between few days step 3 and you will cuatro. Per tick marks the absolute start of go out (we.age., a beneficial.yards.). Committed frames suggest the relative analyses: month step 1 having day 2 (Baseline-split I), few days step three that have day cuatro (Baseline-split II), and few days step one and you may 2 that have month step three and you will cuatro (CLC)