Mapping colonial bushfires through historical newspapers – Fiannuala Morgan

We caught up with Dr Fiannuala (Finn) Morgan to talk about her fascinating project, Historic Fires Near Me – an experimental visualisation of nineteenth century bushfire reporting which forms part of her ongoing research reconstructing Australian bushfire records from 1850 to 1900. The idea was sparked by the devastation of the Black Summer bushfires in 2019–20 and the troubling media narratives that downplay the growing severity of bushfires in Australia. Finn wanted to dig deeper into the historical record to see what the past could tell us.

As a librarian by trade with a passion for literature, Finn turned to Trove’s 1 enormous archive of newspapers and journals. To extract and synthesise this data, she taught herself to code, using a mix of techniques – from simple searches to more advanced methods like Named Entity Recognition 2. The result is a browser-based tool that reveals where fires occurred and how often, across the colonies, over a fifty-year span.

In the following conversation with SLV LAB’s Innovation Lead, Sotirios Alpanis, Finn shares what inspired her research and how anyone can learn to wrangle messy cultural data as she did. She also spoke about the distinct fire histories of different Australian colonies and the ways fire was used as a tool of colonisation in Victoria.

Transcript

0:00:02 - 0:01:09
Sotirios Alpanis
Picture the summer of 2019. Smoke hanging low over towns and cities. Whole regions of the east coast burning for months. The Black Summer bushfires pushed Australia into a crisis that felt new in its scale and fury. Even in the middle of the devastation, some claimed it was nothing out of the ordinary, that The 19th century had fires just as bad, and that history proved this wasn't a climate emergency.Yet when Fiannuala Morgan heard that, she wasn't convinced, and unlike most of us, she had the tools and the instinct to check the record.I'm Sotirios Alpanis, Innovation Lead at State Library Victoria and you're listening to an SLV LAB podcast, part of its conversation series where we discuss emerging technology, digital experimentation and library features with artists, technologists and workers in the cultural sector.0:01:09 - 0:01:35
Fiannuala Morgan
The significance of this dataset is the fact that we don't have standardised, consolidated file records preceding the 1930s, really. It's rather recent. So, building up this longer record of fire since invasion is really important right now, as fire events are getting worse, as the effects of climate change are becoming much more apparent. It's just a necessity.0:01:35 - 0:02:58
Sotirios Alpanis
Doctor Fiannuala Morgan, or Finn, lectures in Publishing and Communications at the University of Melbourne. She's a former librarian and archivist who taught herself to code so she could work directly with Australia's huge digital collections. Her recent project use basic natural language processing, especially Named Entity Recognition, to mine digitised newspapers in Trove and rebuild a record of bushfires in Victoria from 1850 to 1900.The scale of that task is enormous. Trove holds millions of pages, and historical reporting is messy and written across shifting colonial landscapes. Finn's work shows how straightforward computational methods can uncover patterns we've never been able to see. Her research highlights the fires that shaped Victoria's 19th century, how they changed over time and how colonists use fire as a tool of expansion.Most of all, it gives us a clearer view of the past so we can understand the stakes of the present. The following conversation was recorded on the unceded lands of Wurundjeri Woi-Wurrung people of the Kulin nation. We acknowledge the traditional lands of the Victorian Aboriginal clans and their cultural practices and knowledge systems. And we'd also like to acknowledge and pay our respects to any First Nations listeners joining us for this conversation.0:02:59 - 0:03:45
Fiannuala Morgan
A lot of this work is about reconstructing records of bushfires across the 19th century. This is a period where we don't have a lot of historical material about what took place across this time period. We know that there were some big fire events in Victoria in 1851 and 1898, but we don't know much about what happened in between these years.So this project really was focused on looking to try and fill in what happened in this 50 year period. What other bushfires took place using newspaper reporting – newspaper articles that are held on Trove, so that's digitised newspapers – and extracting information about bushfire reporting from that material.0:03:45 - 0:03:49
Sotirios Alpanis
What was it about this research that drew you into it, and why did it interest you?0:03:49 - 0:06:27
Fiannuala Morgan
I just moved to Canberra to begin a PhD on colonial Australian fiction, serialised fiction published in Australian newspapers. And at the same time that that happened, Black Summer took place. And Canberra was completely enveloped in smoke and it was very confronting. I just started working in the National Library and we had to close for a day, which was the first time that had happened in a long time because of the amount of smoke.And at this particular point in time, there was some commentary about the nature of those fires. They were evidently extremely bad. The fires had been taking place across the continent since about winter, just about, it felt like. It was obviously an unprecedented kind of event in at least colonial Australian history. And yet at the same time, there was a kind of a climate of commentary that denied the severity of this crisis.And some of that commentary would draw on a longer colonial archival context to say that "the severity of what was taking place wasn't that bad, because Australia had always had bad fires, and in fact, if we look back to the 19th century, we see a lot of really bad fires on a massive scale. So this can't be something new, it can't be the product of climate change.It's just the Australian environment more broadly". And that felt wrong. And scientifically we know that's wrong. But within that context, and especially working at the National Library, it felt like there should be a better way being able to speak back to these kinds of anecdotal invocations of the past to discredit the severity of the crisis we found ourselves in.And I started looking at newspaper reporting, looking at 19th century newspaper reporting on bushfires, those that specifically you could find through Trove, which is digitised newspapers. So, you could search for the word bushfire and find tens of thousands of newspaper articles about bushfires from the 19th century. And based on finding that very large archive, began this work of extracting information from those articles to put them into a standardised kind of data set to build up this record of fire between, especially 1850 to 1900, with the intention of saying that, what took place during the Black Summer was unprecedented by building up this record of the past.0:06:27 - 0:06:37
Sotirios Alpanis
It's interesting that you switched, or at least were initially interested in fiction as serialised in newspapers, and actually you ended up looking at de fictionalising the historical narrative.0:06:37 - 0:07:30
Fiannuala Morgan
Yeah, that's a great point. And in fact, it did become the focus of my PhD. And very quickly it changed kind of trajectory. But I did remain interested in the serialised fiction, especially stories about bushfires, because they were also published in newspapers, often alongside the reporting of the bushfires that were taking place as well. So you had these two different genres of bushfires in the same kind of medium, which I was really interested in, in terms of thinking about cultural attitudes towards fire and how this changed across that 50 year period.So this research really was about constructing that dataset to fill in this gap in the historical record between 1850 to 1900, but then also looking at these other kinds of representations and fictional narratives about bushfires that were also published alongside that reporting as well.0:07:30 - 0:07:38
Sotirios Alpanis
I'm conscious, as a relatively new person to Australia, I actually don't know that much about Black Summer, so I was wondering if you could introduce a little bit.0:07:39 - 0:08:08
Fiannuala Morgan
So 2019 through to 2020, a large portion of the eastern seaboard of Australia was basically on fire in terms of the scale of those fires. They were huge, they were ferocious, and their impact led to a huge amount of deaths and impact on the ecology and the environment more broadly. They were a firestorm, something that we haven't seen in colonial history.0:08:08 - 0:08:14
Sotirios Alpanis
And you mentioned Trove. How useful is it having all of that historical data in one place?0:08:14 - 0:11:13
Fiannuala Morgan
Trove is kind of it's an incredible resource for anyone who's interested in cultural history or history more broadly, but especially for Australian historians. It's kind of, not to overuse this word unprecedented, but it's a really unique archive. Firstly, because of the coverage of the material that's in Trove, the digitised newspapers itself. Something like 80% of historical Australian newspapers have actually been digitised and uploaded.So the coverage is really expansive. There's a huge amount of material in there. But then secondly, the way that you can interact with Trove and search these historic newspaper articles is a really wonderful, intuitive kind of interface. So you can search for keywords to find articles about specific kinds of topics or subject matter. If you're doing family history, it's a really incredible resource.You can look up ancestors names, specific places, and draw up a huge wealth of material about that particular subject matter. So for historians and for myself doing Australian cultural history, it's an incredible resource for being able to look into these periods of Australian history that haven't been historicised as substantially and start to fill in some kind of gaps there.Obviously, there's limitations to this material, and with my research itself, I talk about reconstructing a record of fire in the 19th century, but nominally it's it's only that nominally, really. This is a data set of place names, which I've extracted from newspaper articles that mention bushfires and in many instances, those place names do relate to a historic fire that took place, but in other instances they don't.And I can talk about that in more detail. But 19th century journalism is a strange thing. There's different genres of articles you can't really anticipate or expect. And then on top of that, journalism changes a huge amount across that 50 year period. The technological changes that are, introduced with the railway and the telegraph means that some of the reporting from earlier in the data set, there's a kind of time difference between a place that's mentioned that might have been a site of fire, and then actual reporting that, you know, is written about that place.So, it's a bit fuzzy, the relationship between the place and the event, of course, towards the end of the century, that kind of collapses, that relationship between the event and the reporting on it. So you've got more kind of up to date, I suppose, reporting, but it means that the dataset itself is a record of fire in a quite a provisional kind of sense.0:11:13 - 0:11:58
Sotirios Alpanis
It's funny, actually, one of my first jobs in the GLAM sector was working on the British Newspaper Archive, which was digitisation project in the UK, based at the British Library. And, yeah, we looked at Trove as a real inspiration. So it's definitely had an influence globally as well. Working on that, one of the things that really stood out was the sheer variety of stories and life that was represented in newspapers.And as you say, you see these historical trends in the UK context. For example, you see a big difference when stamp duty was removed for newspapers and newspapers actually grew physically and paper, became cheaper to produce. It's just really fascinating. I guess in those early newspapers you start to see a change in the way that stories are marketed to people and the kinds of stories that are written.0:11:59 - 0:13:25
Fiannuala Morgan
Yeah, totally. And it's exciting and it's challenging sometimes doing research into the past, especially the 19th century, because our own perceptions of what journalism is and what you would find in a newspaper really don't track with the reality of what you find in these accounts from the past. So the serialised fiction is one thing. I mean, most of where people would read fiction in Australia, where that came from was newspapers itself.There was a huge amount of serialised fiction published in Australian newspapers. There's another project called the To Be Continued Database, which has been led by Professor Katherine Bode at the Australian National University. That has looked at basically bringing all of those stories back together into one centralised database from the material in Trove. And she's put together about 50,000 serialised novels, novellas, stories that were published in Australian newspapers up until 1955.So, it's not just journalism in newspapers, it's all other accounts of colonial life and culture as well, which makes them this really rich resource for thinking about different cultural attitudes and perspectives in the past, as well.0:13:25 - 0:13:31
Sotirios Alpanis
So you touched on it a little previously, but I wonder if you could talk us through some of the methods and techniques you employed.0:13:31 - 0:17:12
Fiannuala Morgan
My background more generally is as someone who did a conventional kind of humanities Bachelor of Arts degree and then has also worked in the library sector for over a decade or so, mostly as a reference librarian, but became increasingly interested in digital collections as well as these kind of database aggregators like Trove, and wanted to start developing some basic digital skills to be able to access and work with these collectionsThat I knew were so interesting but didn't really know how to get into them in any kind of significant sense, apart from clicking through various PDFs or something like that. So in brief, the methods that I use, actually pretty simplistic because they do come from a perspective of just kind of cobbling together some basic skills, not from a data science or computer science kind of trained perspective at all, but more kind of open source algorithms and just kind of, I suppose, heuristic, messy, adaptive ways of working with this kind of messy text as well that is developed from, you know, interacting with this material for so long.So, in brief, the way that this was done was firstly just using the really powerful kind of affordances that Trove offers in terms of getting into these collections, which is just a keyword search, which sounds pretty basic and not cool, but is actually really powerful, especially if you've got, research question about a historical phenomenon that is very culturally or nationally or geographically specific, like bushfire.And like that terminology actually is. So just searching for that word bushfire and beginning from that, I think I got about 120,000 articles between 1850 and 1900, which is too many to do anything with, kind of from just like reading them or transcribing them myself. But then reading through, a lot of them started to develop a sense that there is a way to automate extracting the information about the reporting from them, basically.And the way I did this was by identifying that very often within a kind of word count proximity of that term, bushfire, if there was a place name that was mentioned very often, that place name would relate to that bushfire as a historical event. So based on that kind of finding, I used what is a really old natural language processing algorithm called Named Entity Recognition, which was developed way back in the 1990s.It's a pretty rudimentary kind of algorithm that will identify proper nouns that most often like people, places, or organisations in unstructured text and just will then automatically kind of pull that out. So I found that if I used that algorithm within a specific word count of that term, bushfire, I could kind of start to build up these snippets of reporting about a bushfire that took place at a specific point, a specific place.By doing that. Again, part of being able to do this was the fact that Trove also allows you to download text from that platform, and that was really important. So using their API, which is a way of being able to kind of access this material and download it on bulk, to my computer was also important in being able to do that as well.0:17:12 - 0:17:25
Sotirios Alpanis
Now that you've employed Named Entity Recognition, I wonder if that gives you a new lens through which to see modern AI and things like that. It's definitely a form of literacy where it can help you understand how the modern digital technologies work.0:17:25 - 0:19:55
Fiannuala Morgan
Along with something like keyword search, which does give you a way into a collection, something like Named Entity Recognition, especially if you're interested in names of people or organisations or especially place names, is another really powerful way of also opening up a collection, especially if you've just got a research question, which is perhaps about place, or it's a place based question.More generally. You're able to extract and identify a lot of information that could be really relevant. Of course, I should say that there are other difficulties beyond just extracting this geographic information. Of course, in my case was the fact that once you've done that, you need to assign geographic coordinates to those place names, which as it turned out, was a really challenging process and in part because of the nature of the reporting on bushfire reporting in the 19th century.I mean, these articles would go from the hyper specific and granular because they, you know, often newspapers that would come from a smallish town. So they would be talking about people's specific properties down to their name, up to the level of the colony, talking about fires reported across, you know, the Victorian colony or something like that.So that massive change in scale in the data that I had also made it really difficult then to think about, well, how do I map this? So how do I think about attaching geographic references to these place names? And the solution that I came up with was probably not the most optimal solution or the best one, but it works for the nature of the data that I have, which was finding a historical gazetteer, as well, the gazetteer of historical Australian place names, because Australian place names have also changed across time.And that was something that became very evident as well. With this historic reporting, changes in how Ballarat was spelled. Changes in, you know, when Port Fairy was renamed as Port Fairy, all of these kinds of changes that also needed to be accommodated. So I came up with my own kind of messy approach to assigning geographic coordinates based on my own knowledge of reading a lot of these newspaper articles and kind of approximating where the coordinates basically.Yeah, should go.0:19:55 - 0:20:10
Sotirios Alpanis
That sounds super familiar. Dealing with historical data often gets a bit messy and maybe solve a problem for 80% of the data you're working with. And then gradually you start to chip away the other 20%, and you come to the point where you're comfortable with some of the errors and things.0:20:10 - 0:22:36
Fiannuala Morgan
Yeah. And I think this is something that, I mean, most of my research generally does work with historic material, especially that that does come from Trove. And it's become quite evident to me that you have to accommodate the messiness of this kind of historic material and the cultural data that you're working with. If you bring these ideas of or expectations of gold standard data, or become obsessed with getting rid of all of the spelling errors or something like that, you're just simply not going to get any work done because there's too many, it's too much kind of labor to address that one self.But also it doesn't necessarily impede the work that you're trying to do. So again, something like a keyword search that allows you to have a couple of, I suppose, characters or spelling mistakes in there and still match to the term. You can kind of take that and adapt to that for your own research as well. So very often these place names that I could identify and extract from the newspaper articles that would have a spelling error in them, I wouldn't have to correct them necessarily.I could still work with them, work with this idea of just, you know, allowing a couple of character substitutions or something like that to still assign a geographic kind of reference to them. And of course, also identifying that, yeah, not all of the place names in the dataset that I put together do relate to a historic fire that's necessarily taken place.It's just, I suppose, something that I've accepted as a part of this dataset and that I communicate, you know, if I'm working with climate scientists or fire scientists and have kind of little measures in there to qualify my certainty or hesitancy around some of the information that I've generated. And rather than aiming for this kind of perfection more, being very attentive and understanding and aware of the methods that you've used and how they've produced the dataset that you've constructed, that's kind of more important.I think, than aiming for some kind of gold standard measure that just stops you from doing anything.0:22:36 - 0:22:42
Sotirios Alpanis
You mentioned that you presented your work to climate scientists. How do different audiences react to your research?0:22:42 - 0:24:15
Fiannuala Morgan
I mean, I've gotten some really good responses from different fire scientists who are really interested in the material because of course they're aware that this is, a period where we're missing a huge... we're missing this information that's really important in terms of climate modelling, fire modelling, and also just understanding, how the environment has changed since invasion and colonisation more broadly.So they understand the nature of the source, I suppose, that this material is constructed from and that it is, you know, provisional and contingent. I suppose some other audiences perhaps outside of the sciences, sometimes might get more caught up on the potential errors in the dataset, that it is a bit messy and that perhaps should aspire to greater precision or something like that.But really, the significance of of this dataset is the fact that we don't have records from this period of time, and in fact, we don't have standardised consolidated fire records preceding kind of the 1930s, really this that's rather recent, the actual kind of records that we do have. So building up this longer record of fires since invasion is really important right now as fire events are getting worse, as the effects of climate change are becoming much more apparent.It's just a necessity.0:24:15 - 0:24:27
Sotirios Alpanis
The other thing being able to analyse at that scale, where you're taking in all this historical data, it allows you to show a pattern without messy data, or the data that you can't quite work with? You still get a scale that allows you to see the overall pattern.0:24:28 - 0:27:07
Fiannuala Morgan
Yes, the work that I've done with this dataset, it's about 110,000 entries of historic fires between 1850 to 1900. I've really only looked in detail at the colony of Victoria so far. That's something that became really apparent really quickly through constructing this data set was the realisation that, well, we don't have a national fire history because already we're talking about pre-federatio. But also it became really clear that each colony had a really distinct fire history.It had really different legislation. It had different kinds of industries, and kind of cultural attitudes that meant that bushfires that happened there a really different, across these different colonies in this period. So I've really only looked as far as that of Victoria in substantial detail. But certainly looking at that data across that 50-year period, there are pretty evident kinds of trends.And I mean, I suppose the most obvious one that was clear from the outset was that bushfires were really consistently reported across this 50 year period. It kind of has been suggested, perhaps, that they were underreported because bushfires became so common or something like that, but certainly they remained completely newsworthy across this entire 50 year period. And certainly reading it en masse, it's very evident that fires became worse across this 50 year period in Victoria, worse in terms of, scale and of course, worse in terms of fatalities.As the colony also expanded in population as well. And what was confusing, I suppose, about seeing this was also kind of realising that the colony in Victoria, never really made any concerted effort to contain or repress or control bushfire, also, across this period of time. They were the first colony to introduce fire legislation in 1854, but they really didn't amend or update that legislation until there was a massive bushfire at the end of the century, and a Royal Commission that then prompted some reform to kind of address this.So really for me, yeah, there was a question of why. Why is there seemingly so much complacency, when bushfires evidently so bad and devastating? Which led me to look at how fire was important to colonists as well in terms of, you know, industry and everything like that.0:27:07 - 0:27:14
Sotirios Alpanis
Okay, interesting. So the point in that being that fire was essentially used by the colonists to help control the expansion of the colony.0:27:14 - 0:28:57
Fiannuala Morgan
Yes, fire in terms of how useful it was for agricultural and pastoral purposes, in terms of its relationship to the various Land Acts as well. From the 1860s, when land was opened up to selectors, freeholds that had to be converted into agricultural or pastoral kind of purposes in a certain amount of time. In terms of opening up areas of the colony that were impassable and really difficult to access, especially Gippsland, Fire was a really powerful way of basically clearing land for colonial expansion.And evidently it had a real utility in terms of allowing the colony Victoria to expand and develop across the 19th century. So much so that, even the massive fire that takes place at the end of the century in 1898, the Royal Commission does talk about the kind of causes of that fire being attributed in part to selectors basically burning off and trying to clear land, especially, in Gippsland, because some of these, forests that, they wanted to use for pastoral kinds of purposes – I mean, really difficult to clear these massive trees – and fire was a really powerful way, basically to be able to do that.So in terms of thinking about fire, I think as a kind of tool of colonisation, thinking about the lack of legislative reform as a way of basically safekeeping or, ensuring that the colony can kind of continue to expand.0:28:57 - 0:29:05
Sotirios Alpanis
So you mentioned that your background is in humanities. I was wondering how you tackled the transition to using code, and what were some of the lessons you learned?0:29:05 - 0:30:28
Fiannuala Morgan
I'm not like a programmer by any means, nor will ever be that. The kind of skills and competencies that I've built up. Kind of reflecting on this now after quite a few years, are actually pretty, pretty basic skills. I think that we just don't do necessarily a great job of teaching some of these approaches or ways of thinking in ways that especially bridge things like the humanities or information studies and, say, computer science or data science.So my background was knowing that we had all of these digital collections, especially like on Trove, massive, interesting digital collections that I wanted to find a way to work with. And because of how large they are, had to be some kind of computational approach. But the skills that I've developed a really just about kind of data wrangling and being able to pull out information from digital collections, newspaper articles, Hansard, any kind of digitised collection and wrangle it into a form where it's kind of standardised.And then I can construct a little dataset, whether that's kind of 2000 entries or 120,000 entries, to kind of analyse, to answer this question about cultural history more broadly. And those approaches mostly come from natural language processing kind of softwares and some data science techniques.0:30:28 - 0:30:37
Sotirios Alpanis
I wonder, did you undertake some of the historical research initially? And then it dawned on you the scale was too big and you had to employ some sort of automation?0:30:37 - 0:33:16
Fiannuala Morgan
I suppose I knew there were some kind of historians and people who were doing that kind of research, where they would begin with a large corpus and then have kind of a a question that they would ask of that large corpus. The point being that you could never do this kind of research just by reading this material yourself.You had to use these kinds of digital techniques. And I was interested in the methods that these kinds of people were using, approaches where you could kind of do a thematic analysis of, you know, a lot of historic documents to identify themes, you know, to answer these questions about the past, but invariably, sometimes a lot of these people did have a computer science background, which I don't.And I failed maths in high school. But I did do, some computer science subjects across my PhD. I did a lot of those online "learn to program with Python or JavaScript" kind of courses. Of course, I tried to learn Python and JavaScript at the same time that yeah, in retrospect was yeah, not helpful. Ambitious. Yeah. Well, yeah.Ill advised. And found that I wasn't really getting anywhere. Things really changed for me when, across COVID, Stanford University opened up the introduction to computer science subject as a public service initiative, and they allowed people from all across the world to enrol, thousands of people. And I was able to enrol in this subject and within six weeks found that I had kind of mastered the rudimentary basics of programming with Python, which is discussed as quite an accessible kind of language.And to me, that just speaks to that programming or certain digital methods are not pedagogically, they're not taught very well. And you can be taught these kinds of basics if it's a really good subject. And I'm an advocate of using these kinds of tools that kind of means you don't have to learn to program to do these kinds of, this analysis, you know, asking big questions about thousands of documents or big corpa or something like that.My issue with that is that very often then you don't have oversight of what your methods actually are, or kind of what the tool is doing to produce something. It's produce something, but you don't really understand completely the mechanisms by which that's happened.0:33:16 - 0:33:18
Sotirios Alpanis
What sort of tools are you referring to there?0:33:18 - 0:35:03
Fiannuala Morgan
I mean, if you're using a topic modelling tool or something, like topic modelling as an algorithm that kind of generates themes which based on a large corpus, but really they're just kind of lists of words that you will think, oh, that's kind of, you know, thematically that's about this and that's about this, but it's a way of doing thematic analysis.But you can get these kinds of tools that I suppose, package up that kind of approach in a way. So it's just kind of, you know, a program that you don't have to necessarily program yet yourself. But again, my issue with some of those is that having those rudimentary programming abilities, which really is just understanding data structures and different methods for working kind of with data, that's important because it's about building up a literacy, about what's possible for working with digital collections.More broadly, if you're using these other kinds of tools or even, you know, ChatGPT or something, really, you're not really in control, I suppose, of the analysis that you're doing, it's being directed by something else. So on one level, it's kind of the black box nature, of not having control over your methods. And what's being produced, but secondarily, kind of having that basic programming literacy.I think it's important, especially in kind of the library sciences and working in archives and libraries, because you start to get an idea of what's possible for working with these digital collections as well, and especially even thinking about, researchers who might be interested in using a specific collection, understanding then what can be done with it just from having this kind of basic literacy around.Yeah, programing.0:35:03 - 0:35:18
Sotirios Alpanis
Yeah. It's funny you mentioned being inspired by the researchers. A big part of that in my own journey, seeing them publish their code openly. And once you have a bit of literacy, you start to understand the techniques they've employed and how you might be able to apply them in your own research.0:35:19 - 0:36:15
Fiannuala Morgan
Totally. Being inspired by, I mean, obviously Tim Sherratt's work was really important for when I started out because he would show you what was possible, especially with these different collections on Trove, and also give you little tools as you're building up those skills of being able to get into those collections. But certainly a lot of the work that I've done is built from being inspired by what other people have done with other kinds of cultural collections or visualisations and then thinking how can I adapt this for the research question that I have?How does this relate to the documents that I have? What can I change here to produce something? I suppose?0:36:15 - 0:36:18
Sotirios Alpanis
So I was wondering what's next for this project and for your research.0:36:19 - 0:38:47
Fiannuala Morgan
Now that sort of constructed this nominal data set of bushfires, newspaper reporting is really about expanding this record of fire across the 19th century, specifically collections that haven't been digitised, which are the fire brigade records from the 19th century. So fire brigades, regional brigades that kept detailed accounts of bushfires in these fire books that some brigades have retained and preserved, and specifically, I'm interested in trying to find these fire books from the very many fire brigades across Victoria to start filling in this more official account of bushfire across this period of time as well.So beginning with the Castlemaine Fire Brigade, which began in 1854. So it's one of the oldest on the continent. But they have records of fire dating back to the 1870s in these fire books, which are really exciting records because they're standardised. So they say where the fire took place, when it took place, how many people responded? Was the property insured?Was it destroyed? How much water was used? When was the alarm sounded? All this kind of information that's repeated across all of these entries well off into the 20th century. Just an absolutely unprecedented, incredible account that tells you about community history. It tells you about changes in the environment. It tells you about population expansion. It tells you all of these kinds of things.So I'm getting those digitised so that that information can be preserved because they're just being maintained and looked after by the brigade themselves at the moment, which is incredible that they've survived this long, which is just also a testament to how important and significant brigades are as well to the communities that they protect and serve. But all it takes is one fire or flood for all of that history to kind of disappear.So I'm really interested in identifying who else has these fire books, which other brigades have them, and making sure that they can be preserved for community history and for the history of the brigades themselves, but also contributing to this record of fire that can be used for for climate science.0:38:47 - 0:38:54
Sotirios Alpanis
Yeah. And one imagines that the new dataset will be really interesting to compare to the one that you've already compiled from the newspaper records.0:38:54 - 0:39:25
Fiannuala Morgan
Yes. I mean, it's a totally different kind of record. Yeah. A fire, again, and I kind of have started to do that in a, in just a kind of experimental sense, looking at some of the fires that are in the fire books and then the reporting of those fires that also occurs, in, in the newspapers as well, which is, yeah, really fascinating to start to build up that relationship between these different reports and records.0:39:25 - 0:39:32
Sotirios Alpanis
It sounds like a fascinating next step in your research. Well, thanks, Finn. It's been really great chatting to you and I wish you luck with your future research.0:39:32 - 0:39:33
Fiannuala Morgan
Thank you. Thanks so much.