 So, first of all, I guess I'll say straight up front that I feel a little bit intimidated to be talking to all of you because I'm not a technical specialist, I guess you would call me a subject matter expert. I'm an expert in my domain, by no means an expert in how to do any of this stuff. So, I can see the value in it and I guess that's what drives me forward. So, that's the first thing I'll say. I apologise if it's not technical enough for you, but perhaps you'll see how it is that regular people can engage with vocabularies and use them to actually produce something of value for the department. So, I'm a user researcher, a quantitative user researcher, and came into the user researcher team and they said, oh, can you please just give us some insights from our spreadsheets and they gave me a whole bunch of spreadsheets and none of them were all, they all measured different things and talked about the same thing in different ways and there was no standard. I did used to work in metadata management and so, of course, the very first thing I did was I thought, oh, I think I need to make some standards apply here. And then I need to find a way to actually get all that, so they use qualitative data, so they're talking about videos and photos and handwritten notes and I need to somehow make a way of making this discoverable and searchable. So, the best way for me to do that is, of course, to take my unstructured data and make it structured and by unstructured data I really do mean incredibly unstructured, not just data that isn't in an architecture, I mean stuff sitting around in space. So, the problem that we're trying to solve, that I'm trying to solve right now, is that I've got a lot of research artefacts. I need to make them discoverable. We are working in an environment where people need us to be able to provide them with some insights very quickly. So, being able to actually search something is crucial. The research data itself that we create is not reusable or able to be re-analyzed because it's all qualitative data and essentially the analysis is sitting inside the person's brain. I have a particular passion for ontologies and so I can see that if I can actually look at the links between the things in people's brains then I can perhaps help them to understand how it is that they're actually doing things in a much more ordered fashion. So, this is just something I sent out to the people I was trying to convince that we needed to make a library. So, we decided to make a research library. We decided to order and structure the research. That's a very difficult thing for the qualitative researchers to do because what they're doing is a creative process. So, mostly they're concerned that I will stump on their creativity with my order and whatnot. So, we're trying to make it very usable and I'm doing a lot of user research and making it user-centred but at the same time trying to meet with certain standards and whatnot. So, the stuff that goes into the library, research notes, audio, video, photos, some demographic data which we pull out a lot of data all the time and re-analyze it from data.gov.au or from the ABS and then once it's being used for a particular purpose it's then discarded. It would be nice if we could actually store that somewhere. We've got completed artifacts of research. So, obviously the reports that are produced and also we've got some raw data. So, we're creating data collection tools based on the vocabularies that we're creating to enable the researchers to input their observations in a way that's actually standardized. So, for example, they might note that a person speaks a particular language and in the past they would have just made some kind of note in whatever way they wanted to make it. They might have written down Italian or they might have written down Fiorentino or something like that. But the way that we're producing forms for them so that they can type in IT and it automatically brings up the ABS Australian standard language classification data. So, that just makes it easy for them to meet standards. Then the complexities of creating a library of data about people is that we've had difficulty doing it because we can't purchase, there's no library platform that we're allowed to buy. Department of Human Services doesn't support purchasing of any library platforms that exist in the cloud. Where we could use it in the cloud, the cloud needs to be in Australia. It's very difficult to even get that. There are obviously legislative requirements around records management but in addition on top of that we've got the National Human and Medical Research Council's guidelines that we also need to adhere to. So, there's ethics considerations and obviously every single person has a different level of consent that they complete. When they do a consent form they'll tell us what they're consenting to and so each individual person that we're tracking may have a different level of consent. So, that creates complexity. Essentially creating discoverability requires a well thought out and robust structure to support it and that's why I'm sitting here today talking to you now. So, what we're doing is, this is a list for me to tick off essentially. So, we're working on a draft taxonomy. So, how are we describing, what are we observing? We've actually done this as a whole of government project. So, user research is done across many different government departments and so we have 16 different agencies that we're working with who are doing user research and creating research and qualitative data that they need to store and none of them are really sure how to do it. So, we're working out a taxonomy and that taxonomy needs to incorporate all the different things that the different departments are doing. We're then having each of the members go through and do a qualitative assessment of how have we gone with describing this thing. So, if we're describing, for example, a service, if I've described it in my way, are you happy with, does it meet your needs as well? And I'll show you in a moment what my daggy spreadsheet looks like in relation to that and obviously they're making additions or changes or suggestions. So, we're doing that in an ontological way so that we can, because in recognition of the fact that it's most likely that I'm going to need to have, hello? Hello? Hello Brigitte, we can hear you fine? Oh, someone was trying to talk I think, anyway. So, we're also at the same time checking existing vocabularies. So, we've been using obviously the Research Vocabularies Australia last night. For example, I was watching TV and at the same time going through Research Vocabularies Australia and trying to find things that had already been defined. That's a long process that we're currently sitting on around about 1200 individuals that we're putting into the taxonomy. Once we've done that, we're then going to look at existing properties in ontologies. Well, if we've already, you know, if there's a URI already in existence for something that's already been described and is out there in the world, we would like to obviously reuse that rather than use our own, where it might be broader or narrower than what we're looking at, where we're noting that and obviously putting into my decades spreadsheet. You know, we've chosen not to use that one because it doesn't quite meet the right definition. Where something's not necessarily open or where something's not necessarily defined with a URI, but may have a data source that we would like to connect to. For example, a lot of the ABS data obviously has there's a large bank of, you know, they have their data spine that they're creating. It makes sense for us to connect to that rather than perhaps connecting to something else that's out there in the world that might meet our needs. But if the ABS one has an existing data set behind it, we can consider what we're defining and work out whether or not we could use the ABS one rather than using the one that's out there in the world in which case we might create our own new URI for a term. Obviously, we need to have a look and see once we've worked out what already exists in the world, then we need to look at what needs new definitions or if we actually make an entirely new vocabulary. And that's what we're hoping that we might be able to put that up onto that research vocabulary study database so that others can use it as well. So we do, as I said, we've got 16 different agencies, but there's more than just 16 different agencies that are doing user research at the moment. So once we've done that, that's sort of when we get to the end of what I know what to do. And so then I call Nick and go, okay, I think I've got my list now. And I think we've worked out what we're going to do and we've worked out what the definitions are. Can you help us with actually going and doing the thing and making it into an ontology? So that's when I'll be calling Nick. So in terms of working with other government agencies, we can't ever put our libraries together at the moment because legislation doesn't allow that. There is the capacity to create consent forms that where a person has said, yes, you can share across government, in which case we could actually create a whole of government user research library. There's two benefits to that. One is we do work together with other departments. So Department of Human Services, for example, are always working with other government departments on different projects. Because user research is part of the digital service standard. Everyone needs to do user research when they're creating something that is a digital platform now. If we can actually interrogate our libraries and know that your library is the same as my library, when we're looking at things that we've defined, then that will help us to pull out insights much quicker. The other thing we can do is if we can actually create a whole of government user research library, then we'll have enough data that we could actually have a look and see whether or not the patterns that our current researchers are applying are actually represented in the data. At the moment, people who are doing the work, the researchers, are anthropologists and ethnographies. They all have a lot of experience in the field, but they're still relying on what they intuitively see as the patterns. It would be great, I think, to actually be able to use a large dataset to actually look at it and say, well, who are these people really and what are the things that are really connected? That's a really big part of the work that I would like to get to in the end. User research relies very heavily on two different things. One is personas and the other is archetypes. A persona is saying, here's an example of a person who is of that particular cohort that you're interested in, and we're saying that that's representative of that particular cohort, but we don't really know, for sure, because we haven't ever got the data together to actually look at that. The other thing we use are archetypes. Now, if I wanted to use myself as an example of an archetype, I'm a 42-year-old woman and I have two children. Facebook uses an ontology to represent, you know, work out what ads it's going to display to me and what things it thinks I might be interested in. And, you know, quite frankly, the things that Facebook presents me with are not always things that I'm actually interested in, and I suspect that's because the patterns that they're using aren't always representative of me. So, I think it would be great if we could actually look at the archetypes that we're using in government. So, for example, we use an archetype of or a cohort, for example, of an Indigenous person and we might represent them in a certain way and I think it might not actually be how things really are in real life. So, I'm really excited and interested to see what the library can do to help us understand people better. So, this work needs to be done in DHS and it's something that needs to be done in a whole, a broader whole of government. So, it's just a piece of work that we're working on very slightly. None of us have any money to do this work. We're just doing it in a spare time because we all feel that it's worthwhile. So, the way we're doing it, this is my daggy whiteboard where we started and so the thing about this project that's quite different for for us is that what we've discovered is we've got two different domains. So, we've got the domain of government and the services that we use and so obviously, a gift is something that we're using a lot to describe that but we're also, we have this domain of the user or the person and what we're interested in is how they interact with us and so the things that are of interest there, you can see a lot of them are already represented in ABS data. So, income distribution quite clearly has a standard that we can apply to that but other things such as biases are something that's not really mapped very well certainly in government. There is the behavioural bias codex. So, we've used that one for example. We've loaded that into our architecture. So, this is obviously very much from a Department of Human Services perspective but when we're looking at how people interact with us, we're finding that their environment and their social and their personal aspects of their life are really core to how they're interacting with us. So, we're having to define those things. So, the challenge for that of course is, as I said before, is trying not to destroy the creative process and so, as I said before, we've been loading, we've been creating data collection tools for researchers and talking to a professor at UCLA on Sunday, she tells me that that's quite groundbreaking that hasn't really been done before, which is interesting to me. But anyway, she's started to do it for her research and she thinks it's a great idea to create just some things that we know researchers are noting and use standards to sort of automatically input the stuff for them. So, we've been talking to ABS, for example, to say how should we word this question and what are the standards that we can sort of insert in there? So, for example, DHS has a different idea of what disability is to the ABS and so, we've worked through that and decided that we will define a payment and one of those payments might be disability support pension, but we'll also define disability and we use the ABS version of disability and that allows me to actually connect the ABS data on disability with DHS data on people who are on a disability support pension and so, that's a fabulous thing to be able to do. So, we're creating a three tiered library essentially so that we can create the environment that meets records management standards. So, we've got and also meets the consent forms. So, we've got a space that's just for the researcher that has all of the identifiable data that we have completely locked down. Then, we've got a space where the vocabulary sits, which is in a space that's accessible by the team or hopefully across government eventually and then we've got the third sort of part of the library, which is where all the big reports go and where we've done some analysis where we put that analysis. I thought this would be more interesting to you guys, so I don't normally show the qualitative research as this, but this is a bit of an example of how we're looking at the modelling. So, we've been, one of the data collection tools we're creating is one on digital literacy because there's no standard, well there's actually no way of measuring digital literacy in Australia at the moment and so, I've pulled information from the UK and from the EU and created a survey for people to take out to the field and actually get the people to fill out and of course, each of those questions has a taxonomy behind it and so, we've modelled that there. I don't know if you can probably don't know how to make it bigger. My mouth is too fast. Sorry, I don't know how to make it bigger, but I can send it to you. We're still working on it clearly. Actually, this is a user domain and I've been working on that for the past couple of weeks, so working on the levelling of this and so, it's quite substantially different from what it was. Had lots of internal questions with myself about whether or not opening a business is a service event because we're thinking about it from a government perspective or not and other sort of little struggles with self, but anyway, just show you where we're up to. So, it really is just a taxonomy at this stage. Once we've actually worked out what the links are between the things then, of course, we can start looking at coding it as an ontology. So, I think that's about all. I do put on there some thanks. There are lots of people who've been helping us, obviously. As I said, I'm not a specialist about any stretch of the imagination. So, we've had DHS Metadata Management and Ontology's team have been going through our DAGI spreadsheets and helping us, which has been fabulous. Our internal data architecture teams have also been helping us to link the internal MI with our external qualitative dataset. Nick has been helping me. Department of Health has been going through and looking at who owns what, which has been very useful. Obviously, some people in the community, I've made a whole of government working group, use research library community and some people there, like there's one individual in particular who just is really passionate about it. And so, he sends me emails at all kinds of times of the night thinking about levelling and whatnot. Roy Zett has been giving me some great advice, like don't make an ontology of the entire world. And of course, each of the government departments themselves have been providing us with their DAGI spreadsheets. And they're all feeling very embarrassed about that, but I'm like, it's fine. We've all got DAGI spreadsheets. And if we put them together, then we can, we can sort of work through how, what we're calling things and why, and how we might like to define them. So, yeah, so that's, that's, that's it.