Analytics of Story Metrics

Thanks for doing this. I don't know how to use a python script myself, but it's interesting to look at the sample result.

It would be interesting:

1) To plot multiple story curves and see how much the curves vary
2) To plot the impact on the curves of things like a) changing scores, b) the acquisition of a red H, and c) falling off the new story hub list.

I wonder too whether it matters much what time of day the stories are published. My observation is that most stories are published around 3 am of the Eastern Time Zone, and maximum views begin around 5 hours after that time and last for about another 7 hours, until around early afternoon for that time zone. The OP's curve seems consistent with this -- there's another uptick in the curve around 30 hours.

It would be interesting to compare score curves, vote curves, and view curves and see whatever inferences could be made from them.
 
Pretty cool, not sure how I could use this but would like to try given I have a current series up to 17 chapters currently.

Brutal One
 
I’m too old and not bothered

Just in case you think I’m being deliberately sarcastic it isn’t intended to be. It’s just that when you say “I’ve written a simple python script” it doesn’t look simple to me. Perhaps it’s my age?

When a new story goes up I check it every couple of hours and you can quite easily, spending just a couple of minutes each time, keep track of how readers are rating your story. Counting your 5,4,3,2 & 1’s is so very simple.

It isn’t complicated to do it no matter what anyone else says.

After it comes off the front page it really doesn’t matter because it’ll only show up in a search and the score won’t vary very much. Not enough to warrant keeping a check on it.

This was mentioned in another thread recently and I’ve no doubt someone will pop up with a link.
 
Hi all...
the data scientist in me wondered how my stats for posting a new story evolve.
Decided to write a simple Python script to track it.
Would be cool if this were part of the Lit dashboard.
Anyway, if you are interested, here you go...

Impressive. I've wished on a number of occasions that there was something like this in the form of a graph like the one you attached. The numbers themselves don't mean a lot to me, but the trends do.

I wish there were a way to implement this that didn't require the user to download the stats at intervals.
 
And once again science and the empirical mentality, so useful in many other areas, crashes and burns in futility up against the impregnable ramparts of the imaginative and chaotic human mind....
 
yes. Agreed. There is lots that could be done. These are very good
suggestions!...

The problem was not the Python scripting or adding more analytics, it is the data collection. This was done by hand.
The Lit team has the actual data, so for them, it would be automatic from their DB.

I should mention, the python script didn't render correctly in my original post....
there needs to tab spacing all throughout (those that know, will see it immediately).
but it is straightforward.

thnx for your interest as well....I guess I am not the only freakkkky...

bye.bye...

There are plenty of data nerds here, so your input is very welcome, even if we don't all share your script-writing skills.

I think it would be a terrific feature for the Site to offer graphic displays to authors of the data concerning their stories. Seems like it would be fairly easy to do from a technical standpoint.
 
Hi all...
the data scientist in me wondered how my stats for posting a new story evolve.
Decided to write a simple Python script to track it.
Would be cool if this were part of the Lit dashboard.
Anyway, if you are interested, here you go...

Nifty. Looks like the forum software is not showing your indents properly, even though when I quote your post I can see that you put them in.

If you really feel inspired, I'd love a script that automatically scrapes the site to update the CSV...
 
Nifty. Looks like the forum software is not showing your indents properly, even though when I quote your post I can see that you put them in.

If you really feel inspired, I'd love a script that automatically scrapes the site to update the CSV...

Me too!
 
There are plenty of data nerds here, so your input is very welcome, even if we don't all share your script-writing skills.

I think it would be a terrific feature for the Site to offer graphic displays to authors of the data concerning their stories. Seems like it would be fairly easy to do from a technical standpoint.

I'd argue it's interesting to non-data nerds, too. I like to be able to see trends and relationships, even if I don't have much interest in the absolute numbers on which they're based. I suppose you could call that a form of data-nerdery, though.

A graphic display would make the numbers meaningful to me in a way that numbers alone do not.
 
I think that it’s too easy to overthink scores. It’s clear than only a small slice of readers vote in the first place. I get the impression that readers vote on how much they liked the story in general, without any major effort to delve deeply into analysis of characterization vs pls, etc. (I could be wrong there, for comments received sometimes do mention those.)

I myself rarely give any score other than a 5. If it’s a poor story, I generally don’t even finish reading it. Obviously, that’s not universal, but I’ve seen other people say that’s their practice.

The kicker of course are the omnipresent trolls, who routinely and apparently automatically one-bomb stories regardless of how good they are. Much of their bile gets washed away by the sweeps, but they still skew the scores to some degree.
 
I think that it’s too easy to overthink scores. It’s clear than only a small slice of readers vote in the first place. I get the impression that readers vote on how much they liked the story in general, without any major effort to delve deeply into analysis of characterization vs pls, etc. (I could be wrong there, for comments received sometimes do mention those.)

I myself rarely give any score other than a 5. If it’s a poor story, I generally don’t even finish reading it. Obviously, that’s not universal, but I’ve seen other people say that’s their practice.

The kicker of course are the omnipresent trolls, who routinely and apparently automatically one-bomb stories regardless of how good they are. Much of their bile gets washed away by the sweeps, but they still skew the scores to some degree.

All that's true, and I believe your practice is common, but it obviously is not universal because scores do, in fact, vary quite a lot from story to story. And however low the vote:view or read:vote ratio is, 100 votes is enough to provide a meaningful sample size for most stories.

The data at this Site is very "noisy," but there are still messages to be picked up in it, I think.
 
Bramblethorn pointed out the problem (and solution) to the indentation. I'd taken a guess at it and was wrong.

The other problem I had was "ImportError: No module named pandas"

That, of course, was fatal. What is pandas? What python version did you write it for? My system default is 2.7.
 
With regard to comments on the usefulness of any analytics on these numbers:

just wanted to say, that I don't claim it is useful (too much). In fact,
I am not even sure what a "view" really means. Does it mean that somebody
opens your story sees it for a second and closes it?? I would be interested
to know what is the time between pages, do they get to the end.....etc

A view is registered when someone downloads the first page. There's no time lag requirement. I checked that once with a script that opened an inactive story five times. The view count incremented by five.

Isn't this something that every aspiring writer is interested in knowing?

To get right down to the "heat of the matter", I am not sure what
a "score" means. Was this score based upon a solid literary
criteria? Did plot development or character development enter into
the split second decision to press a "star-button" between 1-5?
Don't know. Probably don't care either.

It was just one of things that with a bit of evidence, one tries to
decipher.

I fear that this may reveal far more about the platform, than the
quality of literary erotica.

The score is the average user rating. I keep a spreadsheet that tracks the votes, and the results are reproducible. The question I can't answer with certainty is whether the score cited is rounded to the nearest 1/100th, or whether it's rounded down or up.

The standards are whatever the reader applies, which is highly variable.

Some writers care about the stats. Some don't.
 
Great!....
Look here.
https://www.anaconda.com/products/individual

You want to download this:
https://repo.anaconda.com/archive/Anaconda3-2020.02-Linux-x86_64.sh

when it is downloaded, do the following:
1) execute: "chmod +x Anaconda3-2020.02-Linux-x86_64.sh"
2) execute: "./Anaconda3-2020.02-Linux-x86_64.sh"
3) follow menu options
4) choose to install line in the .bash.sh file.
5) Open new terminal, you should have anaconda3 in your path.
6) optional: create virtual environments.
7) you are good to go with "pip" if needed.

:kiss:

Thanks, but no. Looking at the code, it's something I can do easily in a spreadsheet, and I don't have much call to do it. I've been tracking stats on my stories for 4 1/2 years, so I have a lot of information at hand.
 
A view is registered when someone downloads the first page. There's no time lag requirement. I checked that once with a script that opened an inactive story five times. The view count incremented by five.

The score is the average user rating. I keep a spreadsheet that tracks the votes, and the results are reproducible. The question I can't answer with certainty is whether the score cited is rounded to the nearest 1/100th, or whether it's rounded down or up.

The standards are whatever the reader applies, which is highly variable.

Some writers care about the stats. Some don't.

I can’t understand how any reader can’t understand what the score means and equally can’t understand how a writer who has had a story published can’t understand what score means in the context of submitted stories. I may not understand this python business but I understood how the scoring worked after submitting my first story.

Rounding up or rounding down? Lit works to the thousandth. So 4.754 is 4.75 whereas 4.756 becomes 4.76.
 
Back
Top