 I know we're getting towards the end of the afternoon now, so thank you all for your presence here still. Yeah, so I'm going to talk a little bit about, I guess, the mixture of data science and sustainability, so nicely let in, obviously, by a lot of the previous speakers and what you've heard earlier today. So essentially, look, how do you bring scientific data in that you can structure in such a way as to be able to interrogate it and run scientific models on that. So I'm just going to talk about a couple of use cases on that and how we approach this. So one of these projects, GLOPAC, it's called, if I mentioned, so essentially this is looking at sustainable packaging and how can we increase the use of sustainable packaging in society. So they're an early stage TRL technology right in its levels right now. So there's an EU project to help them come along the development cycle and bring them into use. So a nice example here of some wheat, and it's actually just straw from the wheat, which is essentially a byproduct and a PLA, which is a bio-based polymer, and how they can be integrated to produce some food packaging. So bringing these together is looking at the whole supply chain along this. So who are the packaging producers, even the polymer producers, right along to the food business operators and then to the consumers. These have to be tactily nice and they have to be able to perform. So how do you understand and analyze that so that you can keep food safe and keep costs right and integrate the circular economy into this. So essentially our role in this is harvesting those data sets and enabling kind of a safe secure environment for competitive companies to be able to put that data in one place and then for us to be able to run the models on that. So essentially we host those data sets and the models. So the food business operators look at their requirements from a quality and safety perspective being too primary along with the various other aesthetics as well and then on the other side of the equation you have all of these polymer producers, packaging producers and they have different performance characteristics and these are non-linear relations. So you need models in the center of this when essentially when a company inputs their requirements into this. So we've designed a decision support system then that can look at this. So kind of getting under the hood a little bit in terms of how this works. So you have the declined layer who essentially is the user and they input the parameters and this is where the data will come back out so they can visualize as well. The structured query layer then to be able to interrogate ask the intelligent questions in the right way. What does this translate in terms of the data sets which are over on the opposite side and then your multiple models then required in order to analyze and interrogate that data depending on the specific query. So bringing all that together in one platform enables the food business operators essentially to be able to input that data. We can automate some of these things like for example the prices is an API essentially in automated input for that. So obviously it's a key parameter and that's a highly variable parameter. So being able to integrate you know allow a user to create a particular scenario that can determine how do they want the particular performance of their product and key scientific parameters in that as well. So you know very simply not too many fancy graphics are required but essentially get back a ranked list of how these packaging materials perform given all the particular scenarios. So sustainability obviously being a key score and that's one of the key factors. So multiple universities involved in this you know developing the various different models. I need to give credit to Maeve as well who's one of the leads in cram on this technology. So for look at one of the particular models in this scenario just drill a little bit deeper behind actual models behind this. So obviously if you're bringing new materials onto the market that are in contact with food you need to understand what's the risk associated with that. So this is actually built on a previous European project called FACET which essentially was looking at additives and other contaminants that are in food stuff but also adhesives and inks and items and actually the packaging materials themselves and how they how do those contaminants essentially move through the food packaging and come into contact with the food. So there's a large complex project that looked at bringing a lot of competitors from various different stages along the food chain or supply chain in the food industry to supply their data and bring that all onto one interface. So we've since taken that model and brought it onto the clouds as obvious benefits that I don't need to explain to this audience and essentially then we the migration model itself is to kind of get into the science here now but two key parameters in this the diffusion and the partitioning. So the diffusion essentially is the rate at which particles move through a piece of material and then the partitioning is essentially the barrier in between that prevents the movement. So essentially you stack all of these up together and the sink medium being in this case is food but it could actually this also works for like wearables for example. So you know people are interested in looking at what migrates through a strap of a watch and the sink medium in this case actually would be would be directly into the human body. So some complicated equations that we've scientists to figure out but some of the key parameters here one of the big values I guess that the FASTA project was that the coefficients here were experimentally derived so like there's a four-year project where experiments are set up to simulate and measure actual physical movement of migrants through materials and so that was all cataloged and you know which becomes quite complex because the different temperatures the rate of diffusion is different so and different polymers and different mediums so so that data set now exists and is available for users so and in terms in terms of being able to figure out how is your packaging going to perform is it going to keep so adhesives and inks essentially are kind of quite quite a number of toxins in them is it going to keep those away from the food stuff and even if they do come true how much of it comes true so very quickly to click down through you kind of get to see essentially this is the assimilating a migrant moving through the different layers and how you might visually very simply visualize it as it moves through and essentially comes towards steady state where you see the drop-off in quantity of that migrant at each different level right through so okay now now we understand what has moved to the packaging and has come into contact with the food stuff then we need to understand okay what's how are humans exposed to this so so there's you know different use cases for our for for populations and there's different ways in which you can consider them and how those demographics are broken down as mentioned so essentially there's data sets that exist on what people eat and what chemicals are in those foods of so essentially we they're very large datasets but we bring those together to understand what the exposure is to the population so as I mentioned data sets and the chemical concentrations so you end up then with a distribution of exposure so very briefly to understand the the consumption habits and we focus on one individual just as an example we look at what they actually eat and usually on a 24 hour recall and so we haven't we have an understanding of the frequency of what the eat various products the amount that they eat and then the different cons chemical concentration so you can imagine that this all when you start to stack up all of this the volume of data really explodes because you're looking at so many different chemical contaminants at different percentages and then you start to layer in more and more people into it and what they consume so I'm running out of time here rapidly so I'm trying to get through this quickly so so once you have a population then you can start to rank them in terms of who are the most likely to be exposed to the maximum amount of a particular chemical or group of chemicals however you want to classify it so yeah so if you have enough like many thousands then you can start to graph this out or so normal distribution nice easy easy one to understand but essentially you can see where where are the entities that you need to be concerned about like you know at the moment you're on 99 percentile so anything so essentially what we're looking at here in this scenario is that there's only 1% of the population that would exceed the maximum dose if that's what in this case was deemed as the hazardous amount of a chemical so essentially that's how you can start right from understanding the chemical concentrations right through the people who are exposed to that and how they how they come into contact with that so yeah so making data available and competitors making data available in in a kind of in a secure and confidential environment is essentially has been the enabler of this you know when you go right back to the structure of the polymers and there's so many different variances there the structure of the packaging material and how that results right out in into a your migration model so couple of takeaways I guess in terms of bringing bringing complex data together being able to scientifically classified in such a way that is it's not just like sentiment it's like you have to get right into understanding you know exactly the right units in the right quantities and in the right way to be able to execute those calculations so you get accurate results out so so that's and then enabling the right people to view that in the right way or in being able to anonymize data so competitors can't reverse engineer back out the information is essentially really key enablers in helping companies to bring unusual and different materials into the market like like the sustainable packaging so that's it thank you thank you Brendan and I have a time for like a very quick question if you if you keep the answer quick and we know that I find that fascinating because I feel like especially in this day and age we're much more aware of how sustainable our packaging is you know we're going to the shop with little analogy we're coming home and most of the stuff is just going straight in the bin you know you're just unwrapping and I'm putting it in the bin so this is really really important technology and research that needs to be done I'm just wondering what is the time scale for something like this at the con from the conception of the new product packaging to the actual like scaling out to market is is it going to take so I'm a lot a lot of these kind of products are rapidly coming on to the market is already being used so a lot of it is about I guess designing to meet the minimum set of requirements you know a lot of packaging right now is way over engineered but people are used to that and it's safe and it's cheap so it's easier to do that so if you want to be able to if people are comfortable we come back to the minimum requirement well then actually this isn't that far away but one other tiny comment as well is that actually the whole recycling industry is in a bit of a disaster at the moment so you know there really has to be a significant focus put on reducing and reusing that you know back to that old thing we all learned as kids so you know that that's that whole industry is you know is essentially collapsing around us right now so yeah big focus needs to go into that and you know for us all to have a better society yeah well there's hope right you're doing the work there okay well thank friends and again thank you very much