Natural History Museum of Denmark: An open source natural history collection

Unknown Speaker 09:35
Okay, I guess we'll start, and other people could join as they go along. Hi everybody, my name is Andy. And we're here to talk about the Natural History Museum of Denmark and their new open source, natural history collection platform today. And I'll just give you a bit of an idea about how it's gonna it's gonna work. And roughly speaking, we're just going to do some introductions into the we are and tell you about my project, and then generally discuss the project hopefully with your input, and from the questions in the chat if that's possible, or in case there are no questions, we do have backup questions, but yours, take precedence. And so please do ask us anything you like, and pop them in the chat. And as if you could just skip the slides forward. Thanks. So, I'm Andy, Andy Cummins I work for a company in Brighton in the UK called Cogapp. We're a digital agency and we'd be more museums, galleries, archives, libraries, etc for a really long time, maybe 3035 years from now, maybe 30 years. Yeah, I've worked on them I had a technical background, and I was a project director on this project so kind of steering the project and dipping in and out here and there, sort of oversight from the Cogapp side, and Neil.

Unknown Speaker 10:56
Hi everyone, I'm Neil Hawkins, I'm a senior developer at Cogapp. And I was a technical lead on this project with the Natural History Museum of Denmark. So that involved getting involved in the beginning, proposal stage and conceiving of the plans to build the build the end product, and working closely with Anders and Andy Yeah. And it's,

Unknown Speaker 11:24
yeah hello everyone my name is Anders, I am Head of Design and digital at the Natural History Museum. I've been working on digital projects, both in the galleries and the collections and I've been kind of steering the project on the museum site, asked me for a long time, really wanted to create these open this open digital portal. And I think I'm just gonna quickly run, just give you an idea of what is the National History Museum and what are our collections and one of the key things I wish I could do is just to, to bring you then and see the collections as they are, they are immense, we are, we have 14 million objects in our collections. And we are also spread around different sites and now you see some really neat photos here but in reality is a really, really large, large collection divided into MANY, MANY, MANY, MANY smaller divisions, and only a few, few tiny, tiny things, actually in reality display, whilst the 14 million objects signed in the behinds all collection is one of those that goes almost. It's one of the oldest activity I think I believe there are a few national history museums in Italy that are older than ours but it's 400 years. And if it contains to have anything for me to write, to be using whale skeletons. We, it's sort of like an archive of everything we have for on the planet, things that are older than the nose own solar system, but there's also a what, what we really proud of is like we have these time series so we can see how what happened since the 16th century up to today and we can basically see how we affect our planet, or whatever we want to investigate. Now, all this, ideally, is sitting in our collection management system, and we are using something called specify, we decided on that system in 15, in reality, we got it more kind of moved in, in, in 20 I would say. The thing about having a division about having open collections is that we would really like to show all the things we have all the meteorites all the butterflies etc but the reality of our museum is actually that we don't really, we, we can't really share that easily. As you see here is just a quote from one we interviewed, like, it's just, who knows who and then you can call up. It's also like, Yeah, somebody asked what's in the collection they will have to email me that something well collection managers could a graph and say it the larger picture around in the United States and EU at the moment we also the all natural history museums are now trying to come together and create one big infrastructure that should be collecting all the collections, into one system. And that's also been a driver for us to basically say okay we need to find a way how to invert our databases and share everything that we have in behind the thick walls. So this is like an initiative that is running at the moment is basically to compare with to say it in one sentence one world, one collection. So it's a whole natural heritage that we had. So what to do. We started off planning a bit and to look at how could we invert our specify how, what, what could we do in the first thing we just quickly did was to look a little bit about what's out there so we did a kind of a landscape analysis of Natural History Museum's online collection portals and others.

Unknown Speaker 15:10
And there were some, a few learnings of it when we met them if you could see qualitatively a bit out then you could say that we learned that MANY of these portals they actually are built under some assumptions that people know what they are looking for that is that if you are really an expert in something you and you want to know about, you want to find do they have this Diptera insect or something you could type that in and find that under a bit so collection basis, which couldn't kind of come in with an open question or, like, just route along in MANY natural history collections at that time I have to say this person 19 We've made mainly build sort of beta deposits and ListViews table structures there are very few it wasn't that visual, and we did some group testing of what would people like to have, have a such a portal and we knew we heard from, from from laymen or non experts that too just to see the to get a visual image of our collections and see what's on the shelf was interesting to be taken by the hand, accessibility to be welcomed feel welcome, not just be met by scientific terms, was, was important. And also that you can share it on social media and or like from our experts and that's something internal at a museum. People want to have it as a freely fast tool, I want to see what we got. Can I can I share this with my colleagues, or my other curators, I wanted to have specific needs of the disciplines geology has different needs and psychology, it now gets a bit into our technical specs but and then also that it's fast and shorten the time it takes to just use our portal. So with this, with this we simply dreamt or wanted to simply see could we build some visual in a dynamic collections portal for everyone basically both to support our machine but also in a broader sense to support our, our national audience as well as, as international audiences. And I can say also that through the MTN community we also knew about the work of Cogapp, and this is how our collaboration, basically started. Now over to you,

Unknown Speaker 17:46
Andy. Cool, thanks Anders and yeah we're gonna say that to Matt and he's on a plane, maybe I can't remember where we were going, but it was to MTN, and he noticed that I was writing my talk on the plane which maybe is not the best thing to be doing but, and he said, we started talking then and, you know, a few years later there was the opportunity to pitch for some work. And with when something like that comes into Cogapp We usually take a look at the work and we hold it up to these principles that are on the screen, and we take these principles quite seriously. And I'm just going to kind of map, our, our, sort of planning and goals against these principles, and so on the next slide you'll see I've drawn out a couple of couple of things here. So you've got better for the people who use our work and better for our clients. And this really means that when we do our work we're trying to make sure that the audiences get what they need, as Andrew said, both scientific and researchers in this case, and the general public, but also our clients, so we want to make sure that our clients lives are made easier, or at least minimize the difficulty of this extra work that goes into putting this kind of thing online and with the numbers that Anders is talking about, you know, 14 million items, that's a lot. That's hard work, and they're all really, really different to one another, and the language is very difficult, and the demand knowledge is extensive. So to get the opportunity to get under the skin and that was really exciting for us. So, you can see here we wanted to build an accessible bilingual destination for both the public and, and scientists, but also present the museum as it really, really is, you know it's a modern research based facility for want of a better word, and we want to show that you can dig into this stuff in a nice, clear, clean way but maintain that professionalism of the scientific research, and it's you know it's in the title of this talk, and leading our industry forward with our work we tried to do that in whatever way we can through our work so we share our work, ideally, whether it's open source, like it is in this case which was really exciting for us, but also using standards, you know, we're big fans of triple AF and this entire collection that is online so far is all accessible by a triple AF, and you can download manifest, you can download images and and take advantage of all of that, which then digs into cross institutional research, and so some other natural history, collecting institutions are starting to support triple AF. This means that you can put specimens from these distinct institutions side by side online for the first time and really compare and contrast those collections. And we also wanted to, I mean, this came out in our discovery work. And we discovered something called debuff that Anders knew all about because they're his neighbors in Denmark, that this is a really science based on science oriented biodiversity. Information Facility, it's the global biodiversity and Information Facility really driven it scientists and but we saw an opportunity to push and doses institutions information through this platform and then draw it out the other side to take advantage of that work that's already gone in there. And so if the scientific community, and then take advantage of that work, and present it in a more user friendly way, say, for people who might visit the museum, you know, school children, people like that. And the last one is, you know just be an HMDs mission is to, you know, further research and and all in all ways in our sort of natural world and, you know Cogapp is very in tune with that kind of thing. So, as part of our mission to try and enable the changes in that we need to make a sustainable world, and this one in our own way. Obviously ticked that box so we're really excited to work in that way. Next is the development, and Neil Do you want to talk a bit like that.

Unknown Speaker 22:16
I will, I can find the mute button and say the technical development really started with approaching what we're working with. So, when we approach any project at Cogapp with a new set of collections data, particularly in this instances, it's the sort of first properly scientific set of collections data that we'd work with or certainly that I had worked with. We really like to dig into things and get under the skin of it. And part of that exploration is to find out what's in use, what data is in use, how it's in use, where it's stored, and then work out from there. How do we get it out. What do we need to do to make it suitable for public consumption and publish it and all that wonderful that sort of thing. So, the aim really is to be able to have a system in place in, at the end which can gather that data from various sources. Ideally, autonomously, then process it and publish it. So, the Natural History Museum of Denmark. There's two main types of data that we're working with this. The actual collection information so that's the specimen records, the objects, which we refer to them as, as Andrew said they're stored using specifi seven which is a collection management system in use. And that's a biological collections data management platform specialist, data management platform which is open source and available on GitHub. If anybody's interested, records from specify are exported in what's known as Darwin Core Archive format. This is a flexible but fairly simple way of compiling biodiversity data into a neat package. Initially this sounded really scary I have to admit, and the archive fourth, it turned out the archive file format is actually quite nice and easy to work with, has been very well designed as a standard for sharing biodiversity information which is complex, but the way of sharing it, it's actually quite straightforward and simple. Now just to pull out one of the main challenges of the So, and as touched on the ambitions for what the museum wanted to present to two people. I think one of the main challenges we identified was the, how we were going to use common names for specimens and for species. So within the museum context the scientific name for something is widely used and understood, but the common name or the vernacular term for something is probably what a lay person, or your eye could use to search something so easy ways to slice and dice and query the, the objects using non expert terms was, was the main challenge identified there. The other, so that's the collection, data, Metadata side of things, and then images are the other big part of this, presenting this connection. images are attached to turns out to records and specify this stored on a separate media server at the museum, various formats they come in and sizes ranging from small to incredibly large. There was obviously challenges there. In terms of processing them in terms of implications of storage costs and where sort of things. But the outcome really was okay, make them available for free play of standards, then we can use them in a consistent way on the application. Have all the things that are triple F gives us fast rich deep Zoom delivery of the images. That's our Anderson horse right nearly at the end, and making the object data and images accessible to people, and if not trivial, easy very easy for people to use and share. So here's a technical overview or architectural overview of the of the end result. Now, in terms of collections data, you can see here on the left we've got to specify seven bubble or circle. And as and as Andy mentioned, we decided to, rather than building a whole new pipeline for processing specimen data.

Unknown Speaker 26:53
We actually worked with GPF and their existing data infrastructure to Pro two processed information from the specified collection system. Now gbf have a data infrastructure that ingests information from specify on an automated basis, processes it interprets the records. So just to clarify the records or occurrence records, which are instances of when a species has been spotted at a particular place and time. So these records are parsed and interpreted by gbf who do certain things like checking the taxonomic terms attached to them, reconciling them against its own backbone and international checklists of species information, lots of clever things, and a very robust infrastructure they have built for, for doing that. So that was a decision we took early, which was actually a change from the original plan but I think with the way we pivoted the project there was of massive benefit to the project and also other goals of the museum in terms of sharing data with with gbf and potentially wider ambitions of being able to share information with things like discos and as mentioned. So we're using the GPS API's and after it's been processed to access object level information on the, on the website. So that's all available under Creative Commons licenses from their API's and that provides our object level data, and search capabilities. Now there are instances where the GPF API's don't provide quite what we need for the application. So we did build some of our own smaller data structures around this. So you can see in the middle we have this data store circled with vernacular search and image Metadata. So there we have a species level set of data, which is specific to the Natural History Museum. So one of the things that we found we couldn't do gbf was get species information at the institution, granularity so we have some custom work there to build some separate things which drive, drive, things like vernacular search so being able to search for the common name of the species rather than its scientific name, the images we have triple F image pipelines, so the data source again is specify, we have a scheduled job which harvests objects from specify pulls out references to images and then puts those images into the start of the pipeline which then kicks off a whole bunch of serverless infrastructure which processes the images and the end result is an image server which looks at a, an image bucket which has been processed, and those are just served straight to straight to the application. One other thing I'll just quickly touch on here is that we have some really cool maps on the website. And these use the G Biff Maps API, which is a really fantastic API for for data with specimen data with geographic information. And we also making use of tiles from the stamen design, which really neatly fit in with the design of the website so I think that's one of my favorite parts of it. That's not on this diagram. I think that's enough technical. Thanks never lose any more people.

Unknown Speaker 30:50
I love people can follow up with you, I'm sure. And so others you want to give us a quick spin at the site.

Unknown Speaker 30:56
Yeah. I would like to Yes, so here you see actually the site has not gone well, we have a silent launch so you actually have in here a sneak peek of of the collections portal we just took a few slides but I think we should also do a little live demo was it just in a second. But this is the front page, and, and a, well, it's we it's built so we you can see, we have 14 billion objects but we certainly don't have 14 billion objects in the portal yet, but there is a little counter, and you can see that we have reached during the last few months of this project, we've actually reached managed to get 353,000 legacy data, pushed in there. And it's as a museum person here, I'm pretty happy about this because before we were doing digital illustrations, sort of ad hoc, and when people had time but now there's even that kind of a motivation you can just, you can go on your, your portal site and you can see what if you register something yesterday. Theoretically, the counter would then keep going so when as an internal culture thing here, this is really nice. You can search you can see here you can go speak flatline search, we'll try that in a second year, just to give you another. Here's some images for the from the mobile version, this what you're looking at as a giant squid. And there's deep Zoom functions with this, and you can see there's a design over here where you can either download the photo anchor it fullscreen it soon, etc. Here we are into the advanced search, and you can see here, you don't need to know that giant squid is technically called aka to theist ducks, the, but you could actually just write giant squid. And then there's an autocomplete and you will see the suggestions coming up. But let's say

Unknown Speaker 33:07
that the key thing there and it is that that that information isn't in your collection management system, the giant squid information that that happens as part of the publication to this platform through eg Biff, so we're kind of enriching the material that the institution holds through gbf and through some work that Neil and other colleagues have done to enable this to happen so if you went into specify the museum's collection management system and search for a giant squid. Probably not, nothing would come up right and maybe notes or something.

Unknown Speaker 33:44
Yeah, nothing. Yeah, taking all our data is basically just a scientific name, a year, and occurrence, etc. So to get to this point, all the data needs to be enriched. But let's try it and I'm going to stop sharing here and just see if we can get. See, shear the screen here. I hope you all see this, and this is the again the front page. Yes, we feature some of our collections already here on the first page is kind of like take a peek, you can basically just with an informal tone, let people come in, and then if we, if we start to write something it could be. it could be a butterfly, basically, then we can see where we get off. Sorry for that. I took a picture so we can try to get a gala theory is a is a is a very known expedition, for instance, that we have had here we get some old currents at the moment we don't have MANY, MANY, MANY images of all things, but we could show so let's see what we got of images. Then you get into this mysterious world of deep sea creatures that were collected during the Galaxy Expedition. If we press one of them. This, I have no, this is just one example. Then, we're able to walk, we'll show that in multiple images downloaders. We can, there's also labels and other images. You can share it there's a triple if manifest for the researchers, there's also the real data basically you can harvest that out. And, and have all the data, but we also show here, the full object, detailed here, which is basically the, the hardcore data that is stored in the, in the database. Now MANY of these are also made so you can basically hover out and link the data so you can basically browse around the in the database. We did some initial testing as I just mentioned in the in the before we started off the project and one of the researchers even said like, well, even though I'm really interested in. In, especially in something special. I'm always interested in the things that are just, just around it. So this serves that need also. Now if we just go back again and go to. So, to the map that Neil was just mentioning which is a really nice feature, everybody can relate to maps, and, and our museum and and me, has never really seen what is our geographical coverage here. And with this, you can see four out of the 86,000 occurrences that we that we now have geographically geo located. You can Zoom in on this nice map, which is stamen design, and we can basically start to see what we got from let's say Faroe Islands. Well we got 72 objects from Faroe Islands, we can see if there are any, if they are photographed or if they're just the data. And we can browse them out here. So that's basically how it works, three lovely methods that suits the design

Unknown Speaker 37:42
might have seen on the, on the little panel that came out, we don't have it for everything but where we have the information we always using the binocular terms to keep that accessibility level as high as possible.

Unknown Speaker 37:56
We can't there's also more filters built into the portal you can actually search for country, or we could also say, what, What did the, what came to the museum, between 16 162 to 18 132. Now on the map here is not that much because it's not geo referenced but if we if we did the data list view, It would be a lot of more specimens that came in.

Unknown Speaker 38:32
We can try. Why is it is doing that I guess like from the open source perspective, because we're using these kind of other platforms like DX specifi or keepeth anyone who has so gbf is a big deal in this world. And so if you have specimen or occurrence information in dbf, then the work that we've done together will give you a big head start on making something that looks like this. And there are a lot of institutions in there, a lot of museums which hopefully once we go live in a few weeks or whenever it is, we can send them you know some information in case they're interested in making use of it. But also, there are other kinds of collecting institutions in there, and universities and so forth, and research and research universities.

Unknown Speaker 39:34
And maybe if I just go back, you can see another feature that has been built in here is that several of the, like there's a little check on each object that already kind of leads the way, because MANY of the MANY MANY organisms on the planet. Some don't even have a trivial name, but if you didn't, if you didn't at least tell people, well, we are within the mollusk, we are within the snails now. That's, that's where we're going so we already kind of telling there so this is hard coded into, and not a part of the data, either that we pull out here

Unknown Speaker 40:20
is kind of a hierarchy of how can we help someone, orient themselves within this certain material by saying, Well, here's the strange organism. Somebody probably won't know what it really is classed as but they might know the one above or it might not even have a class at that level and it goes one above to something that we think somebody would recognize and be able to then make more informed decision around where they go next. That's, That's great, isn't it, Neal roughly Yeah,

Unknown Speaker 40:49
exactly. And while this is a very API driven site. It's not completely sort of free range as it were, there are, you know, there's a, there's a very sort of basic content management ability, which essentially is a few files which you use to, to tell the website which collections, should it pull in, which specimen should it feature, which term should it use on on leave for particular labels in terms and so again from the open source angle that's very adaptable, if you have datasets in, in GB for example, it's a case of changing the identifiers for those datasets.

Unknown Speaker 41:33
And then you also accommodates the language to, obviously, stick into English for us today. You and his team were able to translate the interface language in text, and that's all. That's all in there.

Unknown Speaker 41:51
Yeah, so should we,

Unknown Speaker 41:53
should we go on to questions now. Yeah, I think. So we've got like 12 minutes for questions and there's some in the chat. I scroll back. And so, Andrea, asked about the non experts that you interviewed Anders. Do you mean museum visitors or non academic science enthusiasts or was it a bit of a,

Unknown Speaker 42:15
it was, it was a group of well known scientists, basically we invited people in who didn't have a scientific background. And then we did what we did actually was that we we placed a lot of objects around in the room we group them a little bit so. So the idea was basically that they were walking around in a website, but they didn't know it, and then they could pick up the objects that were on the. They were grouped on different tables. And then we were asked why they would pick this object or what they, What would they like to see etc. And so that was how, how that was done.

Unknown Speaker 43:00
Cool. And next on to does specify act as your collection management system as well. And he kind of dies the collection management system is

Unknown Speaker 43:12
USBs.

Unknown Speaker 43:14
That's right.

Unknown Speaker 43:15
Yeah, so that's effectively you know like TMS for an art museum, or whatever, and it's also open source or. That's right. And there's a consortium behind it, And it's geared towards natural history collections. And like Neil says it has the ability to publish the GPF, which is what we took advantage of, and

Unknown Speaker 43:40
good question from Francis. And this. Yeah, do you have plans to supplement permanently, your museum data with data from GVF so talking about vernacular terms everything specifically which it seems like handling monocular terms has been quite uncontroversial with your colleagues, which has been great MSA, But maybe putting them in the collection system with would start a fight, potentially.

Unknown Speaker 44:12
At the moment, if I understand correctly, I mean at the moment we, we will continue using the, the backbone from GB for the vernacular terms. I'm aware that it's not all perfect, but it is really a giant step as forward as we can get. I know that in other other the other portals that special vernacular databases are being used, but I believe this resources the TV services are really good, and so we're gonna stick with that.

Unknown Speaker 44:48
Yeah so Francis where you say, we would it be added instead I think it's actually, this is all new to Cogapp as we learned as we went through that. It's quite the scientific terms are scientific as in they are what they are, and that's what they are, the vernacular terms are less so, so you can have. What was the example was Robin, somebody mentioned at one point, Robin in the UK is different to Robin somewhere else, I think it might have been in Canada or the US maybe I can't remember, but so literally the word, the vernacular terms are almost contextual in some ways and. And that same idea, the same deal for for Denmark, there was a whole list of vernacular terms that are specific to Denmark. So if you did supplement the scientific language, you would have to supplement it knowing that it will be misconstrued depending who was reading it, because you would have to have a vernacular term for what people think, a call something, and a vernacular term for what people say yes, that's why the controversial thing, We were happy about because that could be a whole political Trixie thing.

Unknown Speaker 46:02
Cool. I think we might have misread Francis question, but.

Unknown Speaker 46:12
So I think there was one more, but I love

Unknown Speaker 46:15
this. Thanks, Adrian. That's a nice comment.

Unknown Speaker 46:20
Yeah. Yeah and Thanks Andrea to the sort of endangered status was another benefit of our saying well let's embrace jeebus they've been working on this year's tech, the good work they're doing, which is to kind of rationalize a lot of this stuff across international borders, and then we can focus on taking the good parts of that, that we think would be useful for the general public, like, you know, that maps in the endangered status, whilst, leaving some of the other stuff. And in gbf where scientists know where, you know that's where they live a lot of the time, and so they can go there and look at all of that additional stuff that we maybe don't foreground as much in on the portal that we've built together.

Unknown Speaker 47:10
Yeah, they're very efficient, achieve if I'd say, every time we said all mapable pull this data in from, you know, some institution or the endangered list. It's, you know, we'd sent an email. Thomas a GPA, oh yeah we just actually we just done that so I think you can get it through here. Okay.

Unknown Speaker 47:33
Yeah. So, in some ways we were lucky, I think, literally next door or in the same building is and this is, this place. but one or the other, like I understand about that. The incentive to get things up on the portal is help people push forward on specify. It's also, every single one of those things that's on the portal is now in gbf, which is also, you know, a kind of a big driver for the museum is under kind of humanity, relationship you want to foster

Unknown Speaker 48:05
the portal actually, by, by, by using GPS, and all the benefits that it gives us. We also closing a really important step for the museum which is that everything that we have in our collections, ultimately needs to go to GPF also to be registered there because it then becomes available for any scientists around them in the world. And so, so now, this, this portal that we built with the advantages from GB actually has closed or done that for so now we export things automatically from our collection management system goes into GB comes back to the portal, even in more enriched way.

Unknown Speaker 48:55
GB flight like it too because you know they're, they, they want this information because the more information that's in there, the richer their data set and then the more that can happen in terms of research so it's a very kind of you know, we put a lot of effort into this work, But it's almost like, because of the way we designed the system and the conversations we had and the choices we made on the way through that work is multiplying in the benefits of where the work where the, where it's been published, and how it's been published with a different audience,

Unknown Speaker 49:25
especially with obviously gbf is a, a prominent source for people to cite and use as a citation and data sources and not sure that's what people would want to continue to do in the academic world. So that's important for those things to be there for that reason as well.

Unknown Speaker 49:46
And Adrian's asking this selected specimens that curated and

Unknown Speaker 49:54
at the moment. Yes. At the moment it's a, we have hard coded that in through the into the design, but we can actually, rather, we can move that around as we go along. You can say from now on we are really going to make a massive digitalization push and produce a lot of more images. The next idea we also going to think about is 3d scans and stuff like that. And all that we, we will now. Once we get nonstandard, we're really excited to do that and we will start to tweak that as we go along,

Unknown Speaker 50:30
you know, because this project was conceived as a almost like a beta v1 project, you know this is very much. Get the building blocks in place so that Anders and his colleagues can work in work in the areas they know you know use their expertise and specify G birth and digitization, we created effectively a pipeline for them to then drop this stuff into which it comes out the other end looking great for, and for everyone. Lily and HMD kind of branding and everything. And it's taken a stab at gbf on the way through to serve broader and scientific communities. And

Unknown Speaker 51:13
yeah, so two important things, one on one on that. Related to that is, from the museum's point of view and from this side of things at the staff there is nothing really additional on them. So, the publication pre specifies the single point at which data comes out, and then it's handled by different systems from there so people don't have to learn new systems or get their hands dirty doing things to move files around in certain ways so that's, that's been a really nice outcome. And the other one I've just completely forgotten and that was this, when understood the hard coded species so the. It's sort of hard coded in that it's also, it's, it's basically been turned into a very lightweight CMS and that Anders can change a file on GitHub, and then click a button and it will automatically redeploy the website so it's like semi hard coded I guess but not completely, permanent.

Unknown Speaker 52:16
Um, is there any one more question here, I think, probably have to do the last one so we've got two minutes left, which parts or parts of the model, we share with our open source Cogapp refinery services, and the open collections web app, or andorre. So, basically, not can not call that refinery services, which basically handles the image conversion, and so on, on mass. So, and as in his team could drop in 300,000 images in there, and they would process them all relatively quickly. And that doesn't usually happen, but they do have a bit of time, and the whole of the rest of it is, is open source, and the way that we incorporated Cogapp refining services in there is in such a way that it's kind of a module. If you had, if you already had a triple A FME Server with your images in there, you would just use that one. And if you didn't have that, you could set up your own doctor called Local that refinery services, or some other triple if infrastructure and. But yeah, I mean, on the last slide I think. And, which we'll put up on shed. Scared is. We'll have links to those things and I'll just say this, in case we forget and it's but it hasn't launched yet, it is, you can look at it all online, but we're not really shouting about it, because we don't want to ruin, and this is PR, PR launch, which is hopefully next week maybe the week after. But yeah, you can play around with anyway. Okay. Yeah,

Unknown Speaker 54:02
we're gonna get kicked off today and politely.

Unknown Speaker 54:07
Thanks everyone for coming. Really appreciate the questions were great. And also it's really nice to hear comments that, you know, praise the work we've done, took us a while, so it's really good to hear. Thank you. Thanks everyone.