Goodreads Librarians Group

message 1: by evan (new)

Oct 03, 2013 10:08AM

Hi everyone,
I wanted to let people know that we are going to start updating data for books currently labeled as 'Unknown Book xxxxx'. This will be an automated process - while we've done a lot of testing, there are bound to be some mistakes.

In the first round, we will be focusing primarily on fixing titles and authors. We will unfortunately not have publisher, format or language information. We also know that some books will be under an ASIN when they should have an ISBN instead. Subsequent passes will hopefully update these missing fields.

The first pass will consist of ~25,000 books.

Please let me know of any questions or concerns!

Thanks,
- Evan

reply | flag

message 2: by Banjomike (last edited Oct 03, 2013 11:38AM) (new)

Oct 03, 2013 11:38AM

Any chance of a list of books that are changed? Not 25000 of course but something for librarians to look through and check. We have experience of spotting book boo-boos.

Is this going to be another weekend job?

reply | flag

message 3: by rivka, Former Moderator (new)

Oct 03, 2013 11:40AM

Mod

A sample has already been reviewed.

reply | flag

message 4: by Banjomike (new)

Oct 03, 2013 11:49AM

rivka wrote: "A sample has already been reviewed."

I meant after the full run. If Evan is correct in his suggestion that "there are bound to be some mistakes" then I for one would be willing to browse through the resulting books and see if I can spot anything obvious.

reply | flag

message 5: by MissJessie (new)

Oct 03, 2013 12:03PM

Rivka, samples being viewed isn't much comfort given the many times in the past amazing problems have arisen with massive changes.

It would be good to be able to review the changes without just stumbling along.

reply | flag

message 6: by Monique (new)

Oct 03, 2013 12:11PM

I have to agree with Banjomike here. If you (meaning Goodreads) have reason to think it might not go 100% smoothly, then why not release a partial list to Librarians after for a spot check. We might not catch anything but if we do, we at least might be able to find patterns in what went awry making the next round easier and less prone to issues.

reply | flag

message 7: by evan (new)

Oct 03, 2013 03:14PM

Hi!
I'd be happy to share links to the updated books, and definitely appreciate any feedback you have! I'm not sure it's feasible to post all 25,000 links, but I'll start by posting a random sampling of 100.

I will be starting the update in just a few minutes, and will post the links once the update is done.

reply | flag

message 8: by evan (new)

Oct 04, 2013 08:07AM

reply | flag

message 9: by evan (new)

Oct 04, 2013 08:07AM

reply | flag

message 10: by MissJessie (new)

Oct 04, 2013 08:15AM

Thanks Evan for responding to concerns. It's nice :)

reply | flag

message 11: by Monique (new)

Oct 04, 2013 08:35AM

I did a rapid spot check of about 12 books. Two had ASINs that are no longer on Amazon. That obviously doesn't make the book invalid, just makes it hard to confirm.

I also noticed that this does not automatically combine works. So, someone will need to go back and see which ones need combining and do that.

Those are just my initial thoughts.

reply | flag

message 12: by Banjomike (new)

Oct 04, 2013 09:29AM

Just a few notes:

https://www.goodreads.com/book/show/1...
shows on Amazon as having two authors
http://www.amazon.com/Puppy-Tricks-St...
the second author is a dog (must avoid making joke!) but would you expect both authors to be added or do you have a dog detector.

https://www.goodreads.com/book/show/1...
has an illegal author name "Dr. Blaise T. Ryan". If the import/fixer is just going to bring in authors then this is the sort of thing that needs to be fixed manually.

https://www.goodreads.com/book/show/1...
is still listed with an unknown author but the Amazon page
http://www.amazon.co.uk/Walking-God-Q...
does have "Harvest House Publishers" listed (and another bunch of real people inside)

https://www.goodreads.com/book/show/9...
this first author is still unknown but Amazon.co.uk lists Teri Fritz in that place
http://www.amazon.co.uk/Would-Like-Fr...

Back later

reply | flag

message 13: by Banjomike (new)

Oct 04, 2013 10:00AM

https://www.goodreads.com/book/show/8...
still listed as unknown author but Amazon has Steve Russell for that ASIN
http://www.amazon.co.uk/BLOOD-OF-INNO...
or on amazon.com
http://www.amazon.com/BLOOD-OF-INNOCE...

This one shows "Jason Vey" as primary author with "Gary A. Shilling" as secondary
https://www.goodreads.com/book/show/9...
Primary author should be "A. Gary Shilling" (different name) and I cannot see any sign of "Jason Vey" in connection with the book. Jason Vey on Goodreads does not appear to be into financial books
https://www.goodreads.com/author/show...
Looks like an Unknown Author has been combined with Jason Vey. Yuk.

Back later

reply | flag

message 14: by Susie (new)

Oct 04, 2013 11:45AM

I have just noticed something strange with one of the author profiles - this guy shows "A Bushman." on the author profile, yet if you click on the book listed there it still shows as "Unknown Author".

Another thing of note is that the period appears to have appeared in the wrong place...

reply | flag

message 15: by Banjomike (new)

Oct 04, 2013 12:15PM

Another one with an illegal author name
https://www.goodreads.com/book/show/1...
If your script could scan for a list of not-approved sequences, Dr, LLD, etc then you could post a list for fixing.

Still listed as unknown author https://www.goodreads.com/book/show/1...
but Amazon has "Colin Haskin"
http://www.amazon.com/Ferret-Girl-ebo...

this one has 2 author listed "Isumi" and "Ted Smith"
https://www.goodreads.com/book/show/1...
neither Amazon, nor anywhere else, has a mention of Isumi and the proper author is "T.A. Smith" as is shown on the cover.

oops, also wrong on Amazon
https://www.goodreads.com/book/show/1...

Back later

reply | flag

message 16: by Susie (new)

Oct 04, 2013 01:02PM

Here's another one that hasn't gone through completely, because there's no author name (has put it under the "." profile):
https://www.goodreads.com/book/show/1...

reply | flag

message 17: by Banjomike (new)

Oct 04, 2013 01:37PM

more duff author names matching Amazon
https://www.goodreads.com/book/show/1...
http://www.amazon.co.uk/Nils-Holgerse...
logically, any author name which includes a plus sign is due for review.

still shows as unknown
https://www.goodreads.com/book/show/8...
amazon has "Tammy Valentine"
http://www.amazon.com/Fresh-To-Your-D...

this one would need the author combining with the proper version
https://www.goodreads.com/book/show/8...
https://www.goodreads.com/author/show...

this naughty book is still showing unknown
https://www.goodreads.com/book/show/1...
Amazon has it listed as by "Blake Cross" (the 'look inside' is NSFW!!)
http://www.amazon.com/Beauty-Undresse...

Back later

reply | flag

message 18: by Banjomike (last edited Oct 04, 2013 03:03PM) (new)

Oct 04, 2013 03:01PM

still shows as unknown author
https://www.goodreads.com/book/show/1...
but the logs show "amazon_sable updated the book Dividing Worlds by Jan Ögren " so the logs know who the book is by.
Amazon agree,
http://www.amazon.com/Dividing-Worlds...

We have an author "Jan Ögren MFT"
https://www.goodreads.com/author/show...
with an edition of the same title but when I tried to remove the MFT from his name I got the prompt to merge author profiles with the "Unknown Author 30" that shows on the book above. Something appears to be a bit confused.

EDIT: if I go to the Unknown Author 30 page there is no sign of the book

reply | flag

message 19: by evan (new)

Oct 04, 2013 06:03PM

Hi guys,
Thanks for all the feedback! It looks like the biggest problem is with the author names - either not updating it, bad characters sneaking in, or getting combined with a different author incorrectly.

We're going to see if we can address some of these concerns with the script. For example, this run has uncovered that some books are storing multiple authors in a non-standard format (at least when one of the authors is a dog!). We may also be looking into using a different data source.

I don't have an ETA for when the second pass of the script will happen, but hopefully it will be soon. I'll update this thread once I know more.

reply | flag

message 20: by Deon (last edited Oct 05, 2013 01:38PM) (new)

Oct 05, 2013 01:38PM

publisher comes up weird
https://www.goodreads.com/book/show/1...

!ruby/hash:ActionController::Parameters text: Harvest House Publishers language: ''

reply | flag

message 21: by Kia (new)

Oct 05, 2013 01:51PM

#20: thanks, it's Harvest House Publishers, checked it at amazon

reply | flag

message 22: by Banjomike (new)

Oct 05, 2013 01:58PM

Kia wrote: "#20: thanks, it's Harvest House Publishers, checked it at amazon"

I've reverted the changes you just made. This thread is for debugging developer problems. Not for editing.

reply | flag

message 23: by Banjomike (new)

Oct 05, 2013 02:07PM

This one looks like it is an automatic NAB (music CD)
https://www.goodreads.com/book/show/1...

illegal author name
https://www.goodreads.com/book/show/1...

reply | flag

message 24: by Banjomike (new)

Oct 06, 2013 03:56AM

I've just noticed that most of the bugs reported in this thread have been fixed. Have the devs finished with that info? The bugs were being left for diagnostic purposes.

reply | flag

message 25: by Abcdarian (new)

Oct 06, 2013 06:58AM

Banjomike wrote: "... The bugs were being left for diagnostic purposes."

Possibly that should have been made clear to us non-techies in the thread title, or at least near the top. I was tempted to start "helping" until I read all the way down to post 22. :-)

reply | flag

message 26: by Banjomike (new)

Oct 06, 2013 07:09AM

Abcdarian wrote: "Possibly that should have been made clear to us non-techies in the thread title, or at least near the top. I was tempted to start "helping" until I read all the way down to post 22. :-) "

I would have thought post 1 makes it fairly clear what is going on. I expect the problem is that many people seem to read the end of a thread first. We'll blame it on the staff for not making it clearer.

reply | flag

message 27: by Susie (new)

Oct 06, 2013 11:58AM

I went back and only corrected one of mine that hadn't been changed yet after Evan posted his update. If I post anything I notice in future, I'll work on that basis, assuming that things have been noticed once a new update is posted.

reply | flag

message 28: by Abcdarian (new)

Oct 06, 2013 02:44PM

Banjomike wrote: "I would have thought post 1 makes it fairly clear what is going on..."

Um, not to clueless me anyway.

Abcdarian <---has some bugs of her own to work out

reply | flag

message 29: by evan (new)

Oct 10, 2013 02:37PM

Sorry for being missing in the comments everybody. I've been tracking the bugs being reported, so it *should* be ok to fix the issues on Goodreads if desired.

reply | flag

message 30: by Melbourne (new)

Nov 01, 2013 02:45AM

A picture of Ehud Sprinzak is available on my profile page. To go on this page https://www.goodreads.com/author/show....
I have some bio stuff for him as well if needed. Let me know in a reply to this comment. Thanks in advance!

reply | flag

message 31: by Koenraad (new)

Nov 01, 2013 04:39AM

Koenraad (koenraadkelemen) | 6989 comments

Melbourne wrote: "A picture of Ehud Sprinzak is available on my profile page. To go on this page https://www.goodreads.com/author/show....
I have some bio stuff for him as well if needed. Let me kn..."

In order to be able to use an author's photo it needs to be one that is free to use. This means that either:

- You own the rights to the image (usually meaning that you created the image yourself).
- You can prove that the copyright holder has licensed the image under a free license.
- You have explicit permission from the author or the copyright holder of the photo to use it here.

Does the photo on your profile meet with one of these conditions?

The bio stuff is always welcome, just post it in this group (preferably in a new thread) and a libriarian will add it for you. :)

reply | flag

message 32: by Moloch (new)

Nov 04, 2013 01:03AM

The Amazon imports have often some mistakes: the publisher name is often spelled ALL CAPS, and the language is wrong (example: https://www.goodreads.com/book/show/1... had "MONDADORI" instead of "Mondadori", language English instead of Italian, and the usual details in the title like "Italian Edition" and the imprint): can it be fixed?

reply | flag

message 33: by evan (new)

Nov 04, 2013 12:13PM

Hi Moloch,
We're experimenting right now with what kinds of data we can programmatically fix. The upper-case is something we can partially fix, but the extra words in the title and the language will unfortunately be harder.

reply | flag

message 34: by Paul (new)

Nov 09, 2013 03:55PM

The author's name on this book is a typo, it is not John Eidsmore it is John Eidsmoe.

BOOK: https://www.goodreads.com/book/show/6...

CORRECT AUTHOR: https://www.goodreads.com/author/show...

reply | flag

message 35: by Moloch (new)

Nov 09, 2013 04:02PM

Another curious thing about this Amazon imports is that they seem to assign an ISBN number to Kindle editions, but then you open the edit page and you see an ASIN number (I don't know if I'm being clear)

reply | flag

message 36: by Andréa (new)

Nov 13, 2013 12:01PM

Moloch wrote: "Another curious thing about this Amazon imports is that they seem to assign an ISBN number to Kindle editions, but then you open the edit page and you see an ASIN number (I don't know if I'm being ..."

I've noticed the same problem. Additionally, if you search for a book's ASIN, it will bring up the Kindle edition, yet when you look at that edition's page, all that is displayed is the ISBN. (If you go through to the edit page, the ASIN is there, but if you click between ASIN and ISBN both wind up disappearing.)

I'm not opposed to an edition having both an ASIN and an ISBN, as the ISBN is often found inside the book but the ASIN is how readers can find the book at Amazon (or in their own collection), but if it is possible to assign both, the edition page should display both. If the edition page can't display both, or editions shouldn't have both, then whatever is causing the Amazon importer to include both should be fixed to include only one or the other.

reply | flag

message 37: by Andréa (new)

Nov 13, 2013 12:06PM

Some Amazon books (not sure if it's the same batch as this thread is about, but it just started this month, so seems connected) are importing author's names incorrectly -- not just the titles and all caps mentioned above, but "last name, first name" format. As a result, those imported Kindle editions aren't always winding up combined with the rest of the editions of those books, and, in some cases, people are manually adding Kindle editions because they don't realize the Kindle edition was already imported, just under the name "Doe, John" instead of "John Doe."

Unfortunately, I fixed all the ones I've seen so far before I thought to post in here, so I don't have any examples to include.

reply | flag

message 38: by Esme (new)

Nov 15, 2013 02:22AM

Andréa wrote: "Some Amazon books (not sure if it's the same batch as this thread is about, but it just started this month, so seems connected) are importing author's names incorrectly -- not just the titles and all caps mentioned above, but "last name, first name" format."

Here is an example:

https://www.goodreads.com/book/show/1...

reply | flag

message 39: by rivka, Former Moderator (new)

Nov 15, 2013 07:12AM

Mod

Are there any with more recent import dates? (That one's is 10/12.) Because I believe the issue that was causing last, first imports has been fixed, although the effect is not retroactive.

reply | flag

message 40: by Esme (new)

Nov 15, 2013 10:04AM

I don't know. I found just two books with author's last name at first and it was by accident. I edited the other one and it is with the same import date:

https://www.goodreads.com/book/show/1...

reply | flag

message 41: by rivka, Former Moderator (new)

Nov 15, 2013 10:24AM

Mod

Yeah, any with that date (or very close, maybe up to 2 days later) is from the batch before we fixed that issue. Feel free to fix them.

reply | flag

message 42: by Banjomike (new)

Nov 15, 2013 10:27AM

rivka wrote: "Are there any with more recent import dates? (That one's is 10/12.) Because I believe the issue that was causing last, first imports has been fixed, although the effect is not retroactive."

This one has two authors with commas,
https://www.goodreads.com/book/show/1...

The second one was added a month ago.
amazon_kcw updated the book München Manhattan #1 (German Edition) by Vollmann, Vanessa
additional author added: Reiter, Annette
(flag)
Oct 12, 2013 09:38AM (#57641576)

reply | flag

message 43: by rivka, Former Moderator (last edited Nov 15, 2013 10:44AM) (new)

Nov 15, 2013 10:44AM

Mod

That's the same date. I wasn't clear, sorry; I meant 10/12/13.

reply | flag

message 44: by Moloch (new)

Nov 17, 2013 03:29AM

https://www.goodreads.com/book/show/1...

This was another case (now corrected by me) of a Kindle edition with an ISBN that turns into an ASIN when you go to the edit page. Also the title had the usual "(Italian Edition)", language was wrong and the description had stuff that it's not allowed in Goodreads (reviews, critical praise, text in all caps, etc)

reply | flag

message 45: by Z-squared (new)

Nov 17, 2013 08:13PM

I don't know if this is a related issue or not, but it is something I have noticed recently. onix ingram likes to import "audio" format books instead of "audiobook", e.g., https://www.goodreads.com/book/edits/...

Is 'onix ingram' amazon? Anyways, seems like an easy script to write, just changing all 'audio' to 'audiobook', methinks.

reply | flag

message 46: by rivka, Former Moderator (new)

Nov 17, 2013 10:13PM

Mod

No, Onix Ingram is another source altogether.

reply | flag

message 47: by Moloch (last edited Nov 19, 2013 03:04PM) (new)

Nov 19, 2013 02:51PM

Do you need other examples of these imports with errors? This was one: https://www.goodreads.com/book/show/1...

and this https://www.goodreads.com/book/show/1...

reply | flag

message 48: by Z-squared (new)

Nov 20, 2013 10:05PM