Labelf Blog

Discover the latest product updates, announcements, and articles from the Labelf team.
June 30, 2022
Use cases

Stockholm's Most Disappointing Tourist Attractions

Summer is here, the sun is shining and you’re ready to go sightseeing in your flip flops, your Hawaii shirt and with your unnecessarily large camera (your mobile phone is probably better at taking photos btw). 

Most articles like this help you find out where to go in order to get the best snaps to post on your socials, which will then hopefully increase your status so much that you get invited to that Tupperware party that you missed out on last year. This article however, will help you find out which tourist attractions in Stockholm are the most disappointing ones so that you can avoid them!

Here are the five most disappointing tourist attractions according to Tripadvisor reviewers:

To find this out, we downloaded all the reviews of Stockholm tourist attractions from the Tripadvisor website (turned out to be about 4000).  We then uploaded that data into the Labelf Platform and started labeling the reviews one by one with either “Happy” or “Disappointed”:

As you go, the magical Elf within Labelf will quickly get an idea of what a disappointed review looks like and what a happy review looks like. Labelf will then start giving you suggestions on what labels to put on a bunch of reviews, and let you correct these suggestions. This makes the process faster and easier:

The best part of all of this is that we don’t even have to put labels on all of the 4000 reviews in order to find out what’s being said in them. Instead, we can let Labelf do the rest of the work when we feel confident that it is doing a good job! If you look at the top of the screen in the above image, you can see that the accuracy of Labelf on this task is 94%, which means that Labelf will correctly guess if a reviewer is expressing disappointment or not in 94 out of 100 cases(!). At this point, we had labeled about 400 out of 4000 reviews, and since 94% is a really good accuracy we decided to go on to export the data. We exported the 400 reviews with the labels we had put on them manually, but we also got the remaining 3600 reviews with labels set by Labelf!

You can even see if a review was labeled by a human (user) or by Labelf! 

You can then work some excel magic in order to assemble the results that we showed in the first picture, and thus find out that Skansen is the most disappointing tourist attraction in Stockholm! 

To be fair with Skansen though, a good chunk of the disappointed reviews had more to do with rain, not being a kid or Covid restrictions, so don’t take it too hard if you are reading this Lasse Berghagen!

P.S.

I want to give credit to my colleague Rikard who did the actual labeling :)

Per Näslund

CTO & Co-Founder @ Labelf AI

Change Cookie Settings