TLDR, I built the Hive Report Card to let people snoop in on their own writing, or the writing of others. I had a chat with a friend the other day, who is a bit of an English language nerd (and so am I) - and we came up with a list of new features, none of which anyone has asked for.
None the less, here they are - whether they are useful or not, is in the hands of the person who looks at the data, and chooses to understand it.
New Features:
Addition of Topic Clustering.
Topic Clustering, in short, looks for words and keywords that typically appear together. This means that if you write on several different themes, topics, or even in certain styles, the act of clustering should bring these posts together. It is an algorithm derived formula in this instance, looking at how often words appear together. You can do fancy versions of this using machine learning and the like - but this is not one of those implementations.
Here it is, run on my account's entire history:
You can also see this as a distribution and see your engagement by topic. Note my long history with the blockchain we were formerly known as, and you'll see that that was some of my highest engagement. We probably had a larger readership and more comment density back then, anyway.
The topics are unique for each person and are generated on wor density. You can also see it develop and change over time:
You can see how my usage of camera things started to pick up in 2019. Then the COVID lock downs and social isolation stuff happened. Then you can see after the world (or at least, my part of the world opened up) I BOOMED with photographic content. There's also a spike there, that during COVID, I played the Final Fantasy games in sequence, and wrote about each of them. You can also see my time when I used to post about Splinterlands a lot.
It tells me a lot about how my interests, and what I've written about change over time, and I think it is a pretty cool representation.
Topic Clustering can be used for many purposes, and this is not a very sophisticated implementation of it.
New Feature!
Vocabulary Analysis
Ever wonder how many words you know, and how many words you use? I am not going to implement flowery, outlandish and overly verbose phrasing here, but I am cognizant of the fact I am guilty.
If you post in multiple languages, this vocabulary analysis will make it look like you have a massive vocabulary. You don't. You just use words from different languages.
Here you can see how many words you might know - and how many you have used on the blockchain, and how often.
We learn new words pretty quickly when we grow up, and hopefully, we use them on HIVE too. I am still growing up, and I am still introducing words that I have never used before on HIVE into all of my posts, as you can see in this chart. You can probably see where I had big spurts of time when I was reading books, with lots of new words added per month, or I explored territory that was previously an enigma.
Next, there's distribution of words, and the highest frequency. Also, how many characters your words are built from. This is your alphabet soup era. Embrace it.
New Feature!
Active vs Passive Voice, Adverb Usage, Filler Words
Active voice is more powerful than passive voice. It makes for more interesting writing that flows better, and is more engaging. It gets to the point. Passive vice has less impact.
Based on these, you can see how your engagement manifests based on the way you write:
With the addition of these new features, there is a little bit more processing time required by the tool, but to my knowledge, this sort of tool, able to analyse writing on a broad scale, quickly, for an individual user does not exist on any other platform.
The best thing? You can do whatever you want with the output, and it isn't locked behind a paywall.
Awesome! Is this vocabulary analysis tailored to english? I have the impression that with my mostly German posts this doesn´t make too much sense.
Yeah, it only works with English - because the stop words are defined in English "this", "that", "is", "the" , etc
As the tool does not have language detection (not likely something for me to implement) it would make less sense in other languages. I don't know any other languages, but I know (from the Simpsons) that "die" would be a German stop word, "The Bart, The" ;)
Cool stuff. We can get some insight into how we write that could be useful. I would expect my themes to have changed over the years.
!BEER
Important to also note that we've all probably written on more than ten topics / themes, but this (really fast) version of topic clustering which is (not as accurate) shoehorns stuff into 10 groups.
I had a quick look and will dive into it more later. Thanks again!
No worries mate, hopefully it is a useful addition. I keep saying no more new features for this tool as its pretty complete and robust, yet here I am each week with new ideas, creating a monster :D
I really like the latest additions ( being a word nerd myself and speaking more than a handful of languages ). In fact, this makes it tempting for me to use a lot more words outside of English ( my second language ).
Keep up the great work! 💪✨
Thank you! I keep saying I can't think of new things to put in this tool, yet here I am a week later. :P
👍👍