After doing some tinkering today, I've ended up making a quick script that means I can generate a wordcloud based on a users forum comment history in just a few minutes, ignoring forums hidden by default (rants/offtopic/premier/etc). Here's mine!
If anyone wants to take a glance at their own history, just ask and I'll get around to generating one for you at some point! Maybe you'll find out something about yourself, or maybe they can just look pretty!
- Last edited on 05-Aug-2019 23:26:01 by Hmm
If the signature is done in the actual signature box, it'll be ignored. In the case of yours where its part of the actual comment, I'll try and eliminate them from the list of words entirely so they don't count.
What is interesting about your signature is the fancy symbols actually broke the script haha, I'll get this done up for you ASAP though.
Edit: Signature no longer kills the script, onto making it look pretty.
Here you are Panda
Edit3: TFW "Closed eyes" made it in there anyway
Edit4: TFW everything bar the fancy script breaking ~'s made it in there anyway.
- Last edited on 05-Aug-2019 17:01:05 by Hmm
Unfortunately it's only looking at what's directly on your profile, it wouldn't be much harder to do the entire forums at once (bar taking forever) but I don't think Jagex would look upon me too kindly given some threads have 100k posts alone. In any case, there isn't much benefit to doing that for most people, adding more data doesn't make the data substantially better vs adding higher quality data. Of course in this situation, there is no standard of quality, but subjectively recent is probably more insightful since over a big enough period of time you might end up being fairly predictable/average.
I'll wait for your signal Jokku, but I'm already anticipating the biggest word
- Last edited on 05-Aug-2019 18:40:58 by Hmm
Go through their profile, grab every comment they've ever made, rank the words by popularity, use the logarithmic scale, show them in pretty colours, fit them to random jpg's.
I actually did that exact kind of thing but taken up a notch for my university dissertation, where I did topic modelling on a very specific kind of law, the idea being the computer could tell using varying levels of maths what was actually being said in these cases over time, and potentially allow for insights on to what things the court considered relevant over a period of many decades. All presented in a nice colourful wordcloud. That actually got presented at a few conferences and apparently was a fairly interesting for the lawyers who'd looked at it, but I have no idea how far they went with it in the end, their problem dealing with text encoding now! ( Fun fact, 0.1% of the backups of the legal cases I were looking at have suffered bit level corruption proving the UK government really needs to step up the game with digital preservation of legal documents)
Anyway money, your results! I can sympathise with the vague network connection issues.