 And we move on to Alessandra de Molde from the University of Antwerp, which is a member of the REIT Co-op. Okay, good afternoon. I am Alessandra de Molde, and today I will talk about creating a workflow to trace consumer values in London auction ads. Let's delve right in. Very shortly put, my research is about why people decided to buy what they did. And for this I look at both words in auction ads. These auction ads were placed after debt, bankruptcy, or moving abroad, and basically listed the possessions of the person that left died or went bankrupt. And, for example, we could find a test bet, as you can see on the left, I think, for you guys, which might be described as having beautiful chins, cotton hangings, or a sideboard, like you can see on the right, which might be described as being from Mahogany, having an inlaid top, or having a beautifully carved feet. I look at 18th century London, because this is a time and place where new and fast-paced consumption patterns emerged. I work inductively, I let the sources speak for themselves, so to say, and I work on a large scale. To get this large scale, I created, I gathered a lot of data manually, and I also cleaned it manually. I have over 5,000 pages of printed auction ads, which I supplemented with other printed sources, such as philosophical treatises, furniture manuals, which are basically interior guidebooks according to one design guru, and auction catalogs. I also use handwritten account books as well. The printed material I ran through Transcribers, I trained a model for which I achieved a 15% character error rate on the training set and a 1.89% character error rate on the validation set. My methodology is too prompt. First, I do a distant reading in the form of digital text analysis, neighbor analysis, amongst other things, and then I do a close reading. I do keyword search and other things for contextualization. Now, after performing one of these digital text analysis, I found about 500 bores being offered for sale. Classical bores, bore cases, bores of prints, collection of bowls, valuable stock of modern bowls, and these appear consistently throughout my research period. So, I opened up Transcribers again. I did a simple search with word preview and found that bowls were actually books being misread. We can see here books of prints, collections of books, valuable stock of modern books. Now, in the first instance, I can see why this has happened. The other ones are a bit more of a mystery, but this example shows that next to spelling variations, it's actually a transcription variation also can have major implications for corpus analysis, both distant and close. Unfortunately, not all of these conundrums can be solved by a simple search. For example, my second example will show this. Here, I created a scatter plot which shows the semantic shift of beautiful between 1760 and 1780. What we would expect here is beautiful to shift away from words related to making an object, such as handmade or finished, to the objects themselves as being described as beautiful. We can read about beautiful diamonds, wardrobes, and butts. Now, I think we all know that sex is a popular marketing strategy nowadays, but I find it a bit odd that this would be the case in the 18th century, so I delved a little deeper and I found two more plausible explanations. On the one hand, it turns out that butts was a popular add-on for locations, such as Lambert butts here, where the auctioneer, Mr. William, held his viewings. On the other hand, it is also a unit of volume or a large cask used to store liquids, typically varying in capacity from 108 to 140 gallons. These casks would probably show up in the possessions of a brewer or a wholesale wine merchant that went abroad bankrupt or died. Thank you for your attention and I look forward to any remarks or questions after this session or at another point at this conference. Perfect. Thank you very much for your presentation.