Better Story Organization

[dale coba]

WordCloud !


(a blacklist could be used to exclude common terms)

- Dale Coba

[WilloWisp]

WordCloud !

At a rough guess, I'd say Yearbook Pictures is my most fanservice-y installment of the Virus Alert series. Here's a wordcloud of it.



Now, what kind of meaningful information could you derive from that cloud to better categorize it?

[dale coba]

We're going to need a bigger exclusion list.

All common words, all words below four letters... not hard to set that up
(but did you ever notice how no one around here ever seems to actually be a programmer?)

- Dale Coba

p.s. your cloud is semi-side-waysish, and the colors make my eyes bleed)

[WilloWisp]

We're going to need a bigger exclusion list.

All common words, all words below four letters... not hard to set that up
(but did you ever notice how no one around here ever seems to actually be a programmer?)

- Dale Coba

p.s. your cloud is semi-side-waysish, and the colors make my eyes bleed)

The initial point I was getting at is that pure vocabulary is not sufficient for useful story categorization. It's rare for me to even say "fembot" in Virus Alert, and the sleeper nature of the 'bots themselves means that I couch a lot of things in innuendo. Yearbook Pictures may well be the sexiest, disassembly-est, malfunction-est chapter so far, but there's nothing about a word cloud analysis of it that would suggest so. Beyond 1.00, there's been virtually no usage of even the term "sleeper".

I hadn't noticed the thing about programmers, largely because I am, by profession, a kind of programmer. My area of expertise simply isn't suitable to creating a word cloud generation tool. It seems a moot point, since they exist in the wild, with highly configurable options. I just don't think it would provide a useful basis for story categorization.

Ps: I used wordle.net, and picked purely random settings. The color probably didn't bother me because I'm colorblind.

[dale coba]

Your points are true; I wanted an impartial tagging procedure.

If I had a list, and knew how to write a UNIX pipe command,
I'd take each story and strip out the common words; then use wordle or such on the condensed remains.

- Dale Coba

[WilloWisp]

Your points are true; I wanted an impartial tagging procedure.

If I had a list, and knew how to write a UNIX pipe command,
I'd take each story and strip out the common words; then use wordle or such on the condensed remains.

- Dale Coba

Pretty sure wordle lets you specify words to exclude. There shouldn't be any need to pre-filter.

[dale coba]

I think it would have to be a large list.

If by some strange accident I succeed in teaching myself HTML5,
(but don't hold your breath)
I'll try to customize the wordle API to our purpose.

For the list: take a good story, and rip out all the special words.

- Dale Coba