Hacker News Reader

victorianoion Show HN: Analyzing top HN posts with language models

and it also looks like most topics are constant over the years - https://drive.google.com/file/d/1ilYn9cnEZwiH1FioUtU9ummhmvn...

victorianoion Show HN: Analyzing top HN posts with language models

BTW there is an implicit recency bias in the dataset, since 2017 the number of top 3K post became more frequently and the avg score is larger year after year as the community in HN grows:

- Number of top 3K per month of publishing - https://drive.google.com/file/d/1beAPP9ijruMUs5DN5wOVsBArvxP...

- Avg score of top 3K per month of publishing - https://drive.google.com/file/d/10nSIgH1a6DN6XrDU2DyMJTCgsIg...

victorianoion Show HN: Analyzing top HN posts with language models

Nice idea and analysis! I reproduced it as well with https://graphext.com and got similar clusters https://drive.google.com/file/d/1-kXsKezu2_S07rQn-0bjbHuUXHE...

victorianoion BERTopic: The Future of Topic Modeling

You are totally right! We are actually making the pricing public next week. We will keep the free version plus starting at $30/month for individual accounts and $150 for team accounts ;)

victorianoion BERTopic: The Future of Topic Modeling

take a look at Graphext ( https://www.graphext.com ) it automatically creates the clustering embeddings using BERT for you + great visualization libraries to interpret the clusters :D it took us 5 years to build the product

victorianoion [dead]

How Trump's supporters perceive some of the new (like Universal Basic Income) and old ideas (Stimulus packages) than a billionaire that could had run for president as a Republican defends after the COVID-19 crisis.

victorianoion [dead]

Just in a few weeks, 200K people started following the @RateMySkypeRoom or the Bookcase Credibility projects on Twitter. Every few minutes a new screenshot of someone on TV speaking from home gets analyzed. Claude Taylor, a veteran of presidential campaigns who worked for the Clinton White House, is behind it.

Since they have already rated almost 1000 Zoom and Skype rooms I thought that would be big enough to turn all these screenshots and reviews into an structured dataset to find some clear patterns on what makes people look good or bad on videocalls.

After getting all the tweets with our scraper Tractor, I uploaded the dataset to Graphext and filter out only the original tweets containing a rating between 0 and 10. Then I clustered all the ratings using our simmilarity algorithm for short texts, which uses word2vec.