Last Tuesday evening, the time had finally come: over 100 people accepted our invitation to a talk and drinks in ETH Zurich’s main building.
The auditorium gradually began to fill up from five o’clock. As the volunteers started socialising, they eagerly swapped ideas and talked shop. Shortly before 5:30, the room fell silent and the suspense mounted. Then, at 5:30 sharp, the event officially kicked off with my talk, which I will now summarise here.
Our 2016 went very differently to planned!
On 18 January Adi Kälin reported on our project in NZZ, which received a huge response in the media, including a report on the news the very same evening. Far more importantly, however, our inbox was inundated with email after email. To cut a long story short, we could never have dreamt of so much dedication! So we wanted to thank you for your relentless devotion with this drinks reception.
The statistics make impressive reading (status 1 November):
- 8,087 pointers received
- 5,751 images improved
- 661 volunteers (90% men)
Apart from the gender gap (where are all the women?), the fact that there is a solid core is also interesting. That’s our Top Ten (currently eleven people!), who work continuously and have sent 56 per cent of all the information we’ve received! That’s 4,581 pointers as of 1 November! As our crowd has broad interests, it includes specialists for mountains, steamers, industrial buildings, power stations, railways or simply local knowledge, for instance.
What do we do with your emails?
First of all, the core message: we receive excellent information from you! Keep it up! We don’t receive any abusive emails; we don’t receive any comments along the lines of, “Oh, that’s a nice picture.” No, all the pointers are knowledgeable. More often than not, you even back them up with evidence.
We then check the plausibility of your information. We can’t do any more than that. This isn’t a problem in itself, though, as another volunteer will soon correct the information if someone is “wide of the mark”. We then enter your information in our so-called metadata fields: the title field or the description. Sometimes, we can even improve datings. We often adjust the subject headings. And eventually every identified or improved image from the category “Do you know more?” is moved into the category “You knew more! Thank you!” We iron out any typos from your original comment and copy it as is into the comments field. So you see, it involves a lot of elbow grease!
Incidentally, the comments field can be searched in full text on E-Pics Bildarchiv Online; we’ve reconfigured it like this at your request. On that note, generally speaking we keep receiving great suggestions from you! Before the NZZ article, there was no workflow, no refined category system, no automatic reply email etc. In the weeks that followed the article, however, we gradually set this up, also thanks to some valuable pointers from you.
A lot of donkeywork! We can get even better!
A small suggestion with a major impact: structuring your emails differently makes our work a lot easier. What most of them looked like until now: this is the Fraumünster in Zurich, best regards, J. Bloggs.
From now on, please write your emails as follows: Joe Bloggs: Fraumünster in Zurich.
(NB: the appeal didn’t fall on deaf ears: the first emails following this pattern already arrived at eleven minutes past nine that very same evening!)
Deborah Kyburz: the crowd in auditorium D 7.1 in ETH-Bibliothek’s main building, Nicole Graf giving a talk at the front, 15.11.2016 (KOM_000465, http://doi.org/10.3932/ethz-a-000986551)
Our daily service: Ø 3 hours/day
Apart from adjusting and improving our internal workflows and fields on the database, we have also developed a daily stint. Four of the six members of staff at the Image Archive are responsible for one day each: Mondays and Tuesdays it’s Julia Fink, Wednesday is my day, Heike Hartmann takes Thursdays, and Roland Lüthi sorts out the emails on Fridays. Barbara Giezendanner helps in the background with queries on digital copies, such as if images are back to front on the database. Marie-Rose Bröchin, the sixth member of the team, is our specialist for postcards and old views. Incidentally, she is currently preparing an interesting bundle of over 3,000 historical railway pictures from all of the world, which are continuously being uploaded (many of them untitled).
On average, three hours a day are spent processing the emails. Usually, the emails are dealt with the same or the following day. It then takes another day to upload the new data onto the image database.
We still have quite a few man-years ahead of us
Here’s a statistical thought experiment (with a wink) on the work carried out: on the basis that 0.4 hours are spent on each improved picture, this works out at 2,300 hours for the first 5,751 images. That’s more than a man-year. Projected onto the remaining 354,249 images on the image database, we have 141,700 hours or 70.85 man-years ahead of us ;-)!
As of 1 November 2016, it has taken us 615 man-hours to process the emails. In other words, we need an average of 4.5 minutes to check the plausibility of the information and adjust the metadata.
An online survey on your motivation and socio-demographic background is scheduled in December/January. As we are also playing something of a pioneering role with our joint project (by which I certainly don’t mean that we’re the only ones conducting crowdsourcing, however!), information on crowdsourcing is called for at many conventions. As a sociologist, it’s a nice opportunity for me to discover more about you, your media use etc. and perhaps even get to the bottom of the gender gap :-).
Mini video series on crowdsourcing
From next year, ETH-Bibliothek will produce its own videos on selected topics – including crowdsourcing, of course. The entire event was also videoed. We will publish the video here as soon as a short version has been cut – one of eight episodes in the miniseries on crowdsourcing.
Deborah Kyburz: he final preparations are being made, video recordings of the entire event, 15.11.2016 (KOM_000466, http://doi.org/10.3932/ethz-a-000986549)
The entire afternoon in the Collections and Archives reading room at ETH-Bibliothek was filmed in the background. Six episodes were shot with six volunteers. In the videos, the crowd activists explain how they approach working with the images.
At the end of the talk, there was an “awards ceremony”: the Top Ten were allowed to pick one volume each from our picture books before posing for a group photo. Incidentally, nine of the eleven Top Ten were present:
Maximiliane Okonnek: group photo of the Top Ten with Nicole Graf, 15.11.2016 (KOM_000470, http://doi.org/10.3932/ethz-a-000986553)
Drinks reception and plenty to talk about
After the thirty-minute talk, it was time for a drink. The crowd picked up from where they had left off in the lecture theatre and were soon chatting away again. The guests also enjoyed talking to the Image Archive staff. Adi Kälin, the NZZ journalist who had really got the ball rolling in the first place, also spotted some familiar faces in the crowd: “I bumped into some journalist colleagues, a staff member from the Historic Building Conservation organisation, former governor Hartmuth Attenhofer and former SP Cantonal Councillor Ueli Keller.” Kälin’s report on the event is available here. The day after the news of Switzerlands’ Italian-language television and broadcasting service (Telegiornale of RSI) reported on our project as well.
ETH Zurich’s web and digital media team (Maximiliane Okonnek, Deborah Kyburz and Kathrin Reith), and Multimedia Services from ETH Zurich’s IT Services (Nathalie Schmidig and Roland Lanz) were responsible for the video production, media coverage and assistance with the launch of this blog. I would like to take this opportunity to thank you all and the Image Archive team again wholeheartedly for your dedicated work throughout the year and especially during the crowdsourcing events!
Complete image information
Kyburz, Deborah: drinks reception with the crowd in ETH Zurich’s main building, against the backdrop of the Polyball 2016, 15.11.2016 (KOM_000470, http://doi.org/10.3932/ethz-a-000986553)
DOI Link: https://doi.org/10.35016/ethz-cs-1651-en