submission accepting prozess

[Lugul banda]

i somehow read in one of many messages, that peter sayed to accept the newer uploads first instead the older ones, to fasten the process,

because alot of older uploads are old cuz they may have issues.. sounds logik.. but...

but that process also includes if somthing happens for the acceptance prozess, some files flip under and just getting old because only the newest ones get accepted.

i have such an candidate https://api.safecast.org/en-US/bgeigie_imports/36578

uploaded weeks ago, and was not checked, 3 uploadchecks and approved  other data  i dont know why it remains... 

on the funny part the older bevore and the newer after got accepted but this one remains like it was somehow forgotten...

i did 2 big measurements after that, and allways the new uploads got accepted, but the old one remain still unchecked unseen unnoticed ,
so it seems for me,

because i got no info why it remains? or whats wrong on it?..

actualy i watch it getting aging even older, all new measurements from last week got allready accepted and approved, and this one still remains...

i understand that catsch up with the new may speed something up,

but wouldnt it be better to push problemmatic uploads,
to a certain folder,
just to make it easyer to see those wich are new, by blend out the problemematic ones from the new fresh uploads?

to a place were they can be checked, from more expirianced members, if they have time,

i know from my uploads that some need special attention,  i know if i look on the api page we have alot very very old datapackets from years ago, in relation to that, my uploade is a child of age...
but those old are getting more and more, and i dont see a progress that the problematic ones get prozessed anyway, only if u get up ur stick and hammer the Drum.

for the moderators, 3 pots.. one for the good one for the bad (special attention ones) and one for those wich lack on Datadetails like City hight direction and more... (still dont know why there isnt an atomatic reminder, to the registred email, hey u forgot to fill in the details on *follow the link* so we can put them on the map - remainder automatism..

-----------------

and that brings me to a point, 

Safecast reclaims to be a System of people helping eachother, 

Knowing systems like Boinc, were u can upload data, and let it run process by other people to check for errors problem solutions, 

simple spoken using the power of the people to solve the Problem.

Safecast is a great thing, but here in the prozessing acceptance of the data , on this problem it lacks exactly what it claims, the power of the people,

well i dont know what secret background Projects iam shure some may have working on that Problem are running,

but just imagine wouldnt it be a bad idea to use exactly the power of the people to speed up the acceptance process?,

an simpel to use app that checks the normal problemeatic features, dose time size of the packed, the good in this folder the bad in this folder or if accepped and prozessed from many as a good one , pushed to the upload folder for the final last manual to prevent hacked wrong data flows automaticly in view?

i work with projekts to find asterioids in space, or seti@home, were the same file is prozessed by many and the checked data is rolled out, like good ones in this folder bad ones in this folder after xxx prozessed views give a gren or a red light....

there is Humanity out there, hundret millions of Servertime to free use, why not use it for the projekt special if its standing on the same foodsteps, the power of the people...?
i actively followed Safecast for quite a while now, that i think some problems are just cuz the lack of prozessing options or people and time to check...

iam very convinced if Safecast uses the actualy existing power of the people, 

like it does for bgeige nano, 

it could be very promising to make safecast even more great, 

by using more power of the people, 

wich is there only waiting to get used...

and i am very shure there are alot people out there wich would happy to help
one some of the boinc apps u can earn milestones, for prozessed data, or u get a nice visualisation, about the data u prozess, or a reminder if u found something special, dont foget humans are a bit of collectors of things, on things they like much is learn something new, and catsch items on the progress , thats how the game industry works.
nicly packed u have a tool wich can get rid about alot problems, if used wisely...

just think about it, and if it was allready disscusted, forget my words and maybe get me a notice why that certain uplods is getting older by day ;-)

best wishes

and ps sorry for my bad english :/

[Sean @ Safecast]

Hi-

To clarify, unless a log has been rejected, it’s waiting for approval. We’re a largely volunteer org and have hundreds of logs coming in every day. We try to get to them as fast as possible and are working on some scripting software to help speed the process up, but in the meantime if you’ve uploaded something and it hasn’t been approved yet, and you haven’t received a rejection notice, this does not mean something is wrong with it that no one is telling you, it means that we haven’t gotten to it yet in order to approve it. Please be patient. Thank you.

-s

Sean Bonner

Co-Founder, Director of Global Operations

Safecast.org | pgp key

[Mat Schaffer]

There are a few issues (including some you've commented on 😉) that will help establish folders/buckets for more focused moderation.

A few examples:

- https://github.com/Safecast/safecastapi/issues/456, isolating drives that need clarification

- https://github.com/Safecast/safecastapi/issues/131, improving sorting

- https://github.com/Safecast/safecastapi/issues/431, reminding people about unfinished drives

The moderation flow is something that has finally reached the point where we need to iron it out and speed things up.

This is great because it means we have lots of data coming in and the system is stable enough that we can try to tackle these sorts of problems. I couldn't have said the same a few years ago.

We've had people pursuing automated systems (e.g., AWS ML) but the basic options require a large corpus of both accepted and rejected data. We have lots of accepted data, but not enough rejected data to build a reliable model.

Crowdsourcing is another interesting idea, but systems like AWS mechanical turk are hard to tune. Tasks need to be easy to pick up and the results need to be clear enough to qualify automatically. This can be especially tricky for specialized questions like "is this air/radiation reading anomalous".

My favorite idea is one from Ray - https://github.com/Safecast/safecastapi/issues/379 but I don't think the database for api.safecast.org can't support it today. The ingest system (ingest.safecast.org) would have a better shot at it, but coming up with a way to do data signing and querying across many millions/billions of points quickly is also going to be a big challenge.

Of course if you or anyone reading is interested in prototyping any of the above ideas, we always welcome it.

Our full data set is on https://blog.safecast.org/downloads/ for anyone to use and we're always happy to consider additional data feeds if a particular format or delivery method is required. Direct contributions to https://github.com/safecast/safecastapi to improve these workflows are also very welcome.

All good thoughts as usual. I think we've captured most of them in a variety of github issues and we'll keep doing the best we can with the time we're able to volunteer. 👍

-Mat