Michelin stars 2020 announced– how accurate was our prediction?

13 January 2020

Article written by Jurriaan Nagelkerke, Principal Consultant

Today was the moment of truth for many restaurant owners, and for us as well. The announcement of the Michelin stars and the moment we would now how well our tekst analysis model performed.

Winners: congratulations!
For starters, we would like to congratulate all restaurants who were awarded with their first Michelin star:

  1. Eden from Valkenswaard
  2. Marrees from Weert
  3. Versaen from Ravenstein
  4. Beluga Loves You from Maastricht
  5. Rantree from Maastricht
  6. Noble Kitchen from Cromvoirt
  7. Zout & Citroen from Oosterhout
  8. Graphite from Amsterdam

Congratulations on your extraordinary performance! To many restaurant owners, being awarded as Michelin star is the cherry on the cake, a recognition for many hours of labour and the blood, sweat and tears that went into receiving this star. You must be so proud!

Leading up to the announcement, our prediction was widely read and shared (here, here and here). Whereas Michelin relies on the judgement from an expert about the quality of the food and the working methods from the chef, we use the reviews from actual visitors. This total of 400,000 reviews contain valuable information which is hidden inside the text. Hidden, since extracting value from textual data requires expertise, creativity and a nice dash of passion for our trade.


Stir the pot, where did we go wrong?

Out of 8 restaurants who were awarded their first Michelin star, none made it to our own ‘Top 10’. Off course we were a bit disappointed and we hope that we didn’t give the top restaurants false hope with our prediction. Based on the reviews from customers, our predicted stars were well deserved. Various media who were evaluating our selection, thought our predictive model and our selection weren’t that far off. For example, based on our selection, AD reported:

“The list is not so farfetched. DeliCees from Cees Timmerman has been tipped for years to receive a Michelin star. Also, 6&24 from Saskia de Kuijer and chef-kok Rik van de Laar have been mentioned to have a chance for a Michelin star in due time. Hofstede Meerzigt in Zoetermeer has already had a number of chefs with Michelin stars working their kitchen, so that might have helped.

According to food critics, The Marque has been deserving of a star for years. Restaurant De Portier in Nijmegen is also mentioned often as a contender, even though chef Ruth Bisseling has been saying for years that he doesn’t believe it will happen.”

Then how… the explanation
Out of 11.000 restaurants who were awarded with a score by us, the restaurants Noble Kitchen and Rantree made it into our top-25. We excluded Beluga Loves You because of a previously awarded star. On the website of TheFork, no public data was available for the restaurants Eden, Marrees and Graphite. That means that these restaurants simply were not mentioned in the reviews, so we could not give them a score. Based on the reviews, we certainly did not expect a Michelin star being awarded to the restaurants Versaen and Zout & Citroen. For Versaen we didn’t expect a star because the length of their reviews was considerably shorter than the data we saw for other restaurants with a Michelin star.

The added value of textual data
With this project, we have tried to show the added value of textual data. Tekst is everywhere, but it is often overlooked in analysis, because it isn’t number-driven data.
Uit onze ervaring van opdrachten bij onze klanten weten dat tekstuele data waardevol is. In dit specifieke geval blijkt het oordeel van de Michelin expert minder goed overeen te komen met de meerderheidsstem van de restaurantbezoekers. De geheimzinnigheid rondom de beoordeling van restaurants voor de toekenning van een Michelinster blijft voorlopig voortbestaan.

More inspiration about (texual) data
In case you are excited about the things you can do with (textual) data, we have a numer of other interesting articles on offer:

Want to know more about how we conducted this analysis? Then please check our website and other channels. We will shortly publish our extensive article to explain our approach to this text analysis.

Want to read how we came to our prediction (prediction model based on text analysis) and our top 10? Read the article: Who will receive a Michelin star?

A big thank you to Anya Tonne, Dion Parameswaram, Nikki van Gerwen, Terri Seuntjens, Wouter van Gils and Lotte van Bakel (Underlined).



Do you want to know more about this subject? Please contact Jurriaan Nagelkerke using the details below

Jurriaan Nagelkerke, Principal Consultant

+31 6 29 62 00 11


Latest news

In 2022, we are going to have a great year of celebration! Many weddings expected

21 February 2022

Plan your calendar free and make sure you have plenty of party clothes in your closet,... read more

Can you escape our “Power BI Escaperoom”?

7 April 2021

In these boring lockdown-times we are all desperately looking for ways to still interact with our... read more

Building a book recommender from scratch

1 April 2021

Almost every day we go online we encounter recommender systems; if you are listening to your... read more

Subscribe to our newsletter

Never miss anything in the field of advanced analytics, data science and its application within organizations!