A Year of Corona Language | Deloitte Netherlands


A Year of Corona Language – A Dutch Perspective

A speech analysis of the Dutch Prime Minister’s opening statements at corona-related press conferences

From the beginning of the outbreak in The Netherlands, Mark Rutte has been in charge of the coronavirus crisis management. To this end, he has given several press conferences. The content of these conferences is the subject of the current analysis, where Natural Language Processing (NLP) techniques are used to summarize and analyze what Rutte has said from 9th of March 2020 until 23rd of March 2021.

Corona Press Conferences

On Monday March 9, 2020, Prime Minister (PM) Mark Rutte gave an opening statement at a press conference following the first crisis meeting held by the Dutch cabinet regarding the recent outbreak of COVID-19 in the country. Throughout the rest of 2020 and the beginning of 2021, Rutte proceeded to give a number of such statements in relation to the developments and management of the virus in The Netherlands.

Here, several NLP techniques are used to analyze Rutte’s language in these opening statements. From such an analysis, one can learn about various aspects of how Rutte addressed the public during the pandemic as well as how the pandemic itself manifested in The Netherlands. For instance, if one looks to which words Rutte used most frequently in his statements, one can better understand which topics Rutte placed the most importance on when addressing the press and the nation. Or for example, if one looks to the overall sentiment of Rutte’s statements over time, one can discern in which periods of time the situation worsened in the country and in which periods it ameliorated.

The current article is ultimately an example of how one can use data science to better grasp world events.

“You often hear, ‘The best time of life is always now.’ And that is often true, but at the moment I am inclined to say that the better times are still ahead of us. – Mark Rutte (23/03/21)

Length of Statements

The number of words in Rutte’s press-conference statements, or in other words, the length of his statements, varied randomly between 607 and 2426 words. Additionally, the average length of his statements was approximately 1446 words.

Additionally, the TV-statements which were given by Rutte from the so-called Torentje on the 23rd of March, 2020 and the 14th of December, 2020, varied considerably in length (1483 and 2426 words, respectively). Het Torentje is the tower adjoining the parliamentary buildings where the PM’s office is found. Speeches delivered from this location are rare and can thus be considered special occasions. As one can see, Rutte addressed the public with nearly an additional 1000 words during his TV-statement 9 months into the pandemic in comparison to the beginning.

Most Frequently Used Words and Combination of Word

There are several notable observations found in Rutte’s top 20 most frequently used words throughout all the press conferences (some of which can be seen on the dashboard). These included words such as “blijven”, “mogelijk” and “thuis” (“to stay”, “possible” and “home”) or “houden”, “regels” and “maatregelen” (“to keep”, “rules” and “measures”). Indeed, throughout the last year, Dutch citizens have repeatedly been called to action by the PM to stay at home as much as possible and to keep to the rules and measures.

Rutte’s top 20 also comprised of the words “anderhalf”, “meter” and “afstand” (“one-and-a-half”, “meter” and “distance”). In fact, his most commonly used combination of three words was “anderhalve meter [afstand]” meaning 1.5 meters [distance], which was the standard distance Dutch citizens were asked to keep from each other. A further interesting remark is that one of the dates where “meter” and ”afstand” are more actively used was on the 24th of June – the first conference since the reopening of bars, restaurants and terraces.

Furthermore, Rutte frequently talked about “onderwijs” and “school” (“education” and “school”) as the closing and reopening of schools has indeed been a key concern in The Netherlands during the pandemic.

The most frequently used words thus paint a clear picture of the topics Rutte placed the most importance throughout the entire pandemic.

Words Defining a Given Statement

One often assumes that the words which best define the topic of a particular speech or statement are the most frequently used ones. Contrary to this belief however, it is usually the less frequently used words that are the most topic-defining. The trick to uncovering these words is finding the ones which have a high frequency in the given statement relative to all the other statements. This measure for topic-definition of words takes a value between 0 and 1, where a higher value implies that a word is more topic-defining.

The most topic-defining word of the very first corona-related conference was “Brabant”. This is the province where the virus first broke out. From the two statements given 9 months later in December 2020, the word with the most contextual importance was “kerst” (or “Christmas”). As one might expect, on the 20th of January, 2021, the most important word with regarding the context of the Rutte’s statement given during that conference was “avondklok” (“curfew”), which was in fact three days before the curfew was implemented. In the following press conference, on the 2nd February, the most topic-defining word was “variant” (variant) referring to the British variant of the virus that had at that point started spreading in The Netherlands.

It therefore becomes clear that by looking for these words in each one of Rutte’s statements, one can uncover the dominating context of the press conference on certain dates.


The approximated sentiment of Rutte’s statements did not vary greatly, with their polarity ranging roughly between 0.0 and 0.2 (where a polarity of -1 and 1 indicate extremely negative and extremely positive sentiment, respectively). It is however intriguing to see how the variation in sentiment corresponds with how the pandemic unfolded in The Netherlands. For instance, one of the most (relatively) positive statements was given on the 19th of May, 2020 – approximately two weeks before the first reopening of bars, restaurants, etc. Moreover, it is interesting to observe the steady decline from last September, when the number of infections had started to increase very rapidly, to the 3rd of November, when there was an all-time high number of infections.

The approximated sentiment of Rutte’s statements did not vary greatly, with their polarity ranging roughly between 0.0 and 0.2 (where a polarity of -1 and 1 indicate extremely negative and extremely positive sentiment, respectively). It is It is interesting to note that no relationship was found between the length of the PM’s statements and the overall sentiment of his statements. That is to say, that whether Rutte’s language was more positive or negative was not associated with a shorter or longer statement length.

Future vs. Past Looking

Analyzing the distribution of verb tenses in Rutte’s statements further revealed that the PM spoke overwhelmingly in the present tense (nearly 80% of verbs). The remaining 20% of verbs show that the PM might be slightly more future looking than past looking as roughly 13% of verbs were future tense and 7% were past tense.

A Year of Corona Language

More than a year has passed since the news of the first coronavirus case in The Netherlands broke and the virus still continues to dominate headlines as well as political and public discourse in the country. The present article has shown how the use of NLP techniques on a portion of this discourse, namely, the Dutch prime minister’s opening statements at corona-related press conferences, can reveal key points of the management and development of the virus in The Netherlands over the course of the last year.

More information?

Do you have questions or would you like to receive more information about the article A year of Corona Language? Please contact Simona Nita using the contact details below.

Did you find this useful?