I think "panda" is a program you use to grab and queue up hits without having to use Amazon's horrible interface? It's a pretty slow trickle because most of the retrieval tasks are coming back negative.
Sorry, no. panda stands for PreviewAndAccept So, you take a link like this one for instance https://www.mturk.com/mturk/preview?groupId=3QTV91O5K78OAUDOBLE3597ASNRR69 You put andaccept after the word preview and before the ? Then when the hit pops up you would automatically accept it. People run it with other extensions or scripts to make it repeat over and over again at a predetermined time frequencey i.e., every 5 seconds etc.
https://www.mturk.com/mturk/previewandaccept?groupId=3AIRCNF3DA1OKH6A5BZQRTB5Q7PMQ4 This is the panda for the extract. The initial pay is .02. I just did one with 9 tabs and got a .45 bonus, so that would be a .05 bonus per tab. The hit itself doesn't state the bonus structure. I assume that it will always be .05 per tab? It would be nice to have the bonus structure in the hit, so these are a little less "scary".
This is the second extract I have ran into that has this. They have multiple years of donors listed. Do you only want the more recent? Like this one has 2012-2013 donor list and a 2013-2014 donor list. Do you only want the 2013-2014 lot?
Batch pulled. Ran my analysis and costs have increased fourfold. Retrieval rates down to 2% from a steady 10%. Stay tuned. If I don't come back, it's because they killed me.
I've found a couple by adding search terms to the google search like "nonprofit" or the state (all the ones I did today were Cali), but for most I've had to put N/A with a "Too generic a search phrase" in the comments section.
Expect a batch today - I'll keep you updated. I'm implementing the following suggestions: 1. adding state to the info you get about the org 2. explaining the bonus structure in the instructions 3. adding that you should take the most recent year 4. increasing pay for homepage and categorization (in the hopes that it stabilizes retrieval rates)
I also added a link in the retrieval task that will use google to search inside a domain, i.e. donors site:redcross.org Testing in sandbox now.
Also, if you don't have the qual, I'm opening it up on a limited basis. PM me your Worker ID (preferably put it in the subject of your PM with no other information, but I can manage if you don't) and I can add you to the qual later today.
Looks like it's live. Should get lots of donor lists today because the batch is all New York. Anyway, happy turking! I'll be in and out today, checking on the batch, checking the forums, etc. But if I don't see you, happy weekend!
OCMP34, There's somebody scamming donor list retrieval. Extractions are coming up with no donor list link. I'm skipping those for now, there seems to be over 300 extraction jobs up with no donor list apparently.
I'm still getting an error on the extractions? Is there a different panda for them than what I have..maybe?
