The Insight Podcast
The Insight Podcast
Ep5. The umpire strikes back
In episode 5, Louise and Gráinne talk to Insight post-doc and international hockey umpire Dr Alison Keogh. They also chat to Ahmed Jouda who interned at Insight during the summer and who happened to have a story about mapping Lesotho. Finally, they catch up with Insight's Citizen Science project Crowd4Access to find out how it's going.
Grainne Faller 0:04
Hello, and welcome to episode five of The Insight Podcast. I'm Grainne Faller,
Louise Holden 0:09
I'm Louise Holden
Grainne Faller 0:10
I can't believe we've got five under the belt Louise, can you?
Louise Holden 0:13
No! It's like we're pros now or something
Grainne Faller 0:15
Hardly. We have a great little show lined up for you today. Later on, we're going to be talking to the people behind Insight's citizen science project Crowd4Access.
Louise Holden 0:29
We're also going to be talking to Ahmed Jouda who did an internship with Insight in UCD, like a lot of people. It's an interesting role to take, and we wanted to hear about his experience. Actually, we found that he had a great story to tell from beginning to end, so we'll look forward to that one.
Grainne Faller 0:41
Yeah, he was a lovely young chap. But first, we're going to talk to Alison Keogh, who's based in Insight at UCD. She works with Brian Caulfield in the Sensing and Actuation group. But Alison also has a secret life as an international hockey umpire. So we decided to ask her about that.
Alison Keogh 0:57
I kind of got into umpiring almost accidentally, probably around 10 years ago, now, my dad used to umpire all my hockey games. And so you know, you figure out kind of what the rules are. And then when you become more moody, to the referees, particularly your father, they hand you back the whistle and say, off you go, it's your turn. So I was very lucky. I mean, I suppose I started at 18 or 19. At that stage, there weren't many kind of young females and probably still aren't in that sort of scene. So it meant that I suppose I was pushed quite quickly, to progress to an international level. I got selected for this umpire development program within Europe, which I suppose kind of brings you on to the scene a little bit earlier, maybe then you were probably intending to be there. And then from there, it's a case of perform your way up, or not, as the case may be. So I've been very lucky. I've got I think, 78 caps now internationally at a senior level, travelled the world with it. And yeah, Germany was first international in over nine months. So it was nice to be back on the pitch and doing that level again.
Grainne Faller 2:05
And what was that like traveling during COVID?
Alison Keogh 2:07
Yeah, it was strange, because yeah, the last time I had been on a plane was was February when effectively nothing was happening. You know, the odd person was wearing a mask, whereas this time was obviously completely different. Obviously, masks are mandatory. First time traveling for such a long time. Because once I arrived at Germany, and then had to get a train on still slow Dorf. So you were wearing it for the whole day, basically. But actually, it was fine. It was normal. Other than the masks, you know, obviously, you're aware of your sanitizing, I suppose you're aware of the amount of things you touch and who you're interacting with. The plane over was empty. You know, so actually, it was lovely. Compared to usual, probably.
Grainne Faller 2:44
You were an umpire during the World Cup for Ireland. So well was that what was that like? It must have been incredible.
Alison Keogh 2:50
Incredible from a number of perspectives. So from an umpiring perspective, once you kind of get on the ladder of international, there's five different grades that you can get to so I had kind of got my international badge in 2013. And then if I'm honest, I hadn't really developed for around three years, I think I got upgraded once. And then kind of around the time I finished my PhD, it just all clicked. And I went from having I think 30 games in my first three years of a badge, to doing 50 games in the last two years alone. And so with that came three or four different upgrades. And I suppose part of that was culminating in, between September and November 2017, I got upgraded twice, did two kind of major events, and then out of nowhere, got selected for the World Cup. As far as I concerned. It wasn't on my radar. So being there was, you know, a huge surprise and a huge privilege. And you know, England hockey and the FIH did a fantastic event with their, their stadium. And their setup was just so professional. I mean, I suppose I didn't expect to be there. But then no one expected the Irish women to do as well. So it was just this kind of surprise after surprise. So I mean, did I ever think I would be standing there watching them in a final? Never, never quite frankly. And so it was really an honor to be there from a range of perspectives to see them do so well. And, and obviously even just to experience the event was fantastic.
Gráinne Faller 3:29
That was such a fantastic experience. I remember actually watching the final with my with my son. And then my daughter walked into the room and she was like, "Ooh, girls!" and she sat down to watch. She was just so excited. It was fabulous. Yeah. So how timewise I mean, that sounds like you're operating at a very high level in that perspective. How do you manage to balance that with, you know, your postdoctoral research now and all of that.
Alison Keogh 4:48
I couldn't do it without the support of Brian. I mean, being being perfectly honest, he's been understanding from day one. And actually, when I started working in Insight, it was just after my PhD. You know we had a kind of a chat about work and opportunities and that kind of thing. He happened to have kind of a little bit of work that needed doing, and as we were kind of sitting down and saying, "Yeah, this sounds great." I was like, I actually have to go to New Zealand for three weeks. And you know, I mean, that was literally day one. So it was the start of it. But since then, he's been fantastic, I suppose, doing a PhD, you become very self directed in your work. And you get used to just having to do maybe slightly different hours, you know, at different times in your work. And so I've just transferred that into hockey, and vice versa. Last year, there was a huge schedule, basically, between February and June, I was out of the country every week for about 16 weeks. And it might have just been a quick trip over to Holland, and then straight back, but it was still out of the country. But I just became very good at learning to work anywhere, at any time. And then, you know, I have the understanding with Brian, that, you know, once the work is being done, that it doesn't matter where in the world I am. And the work that I'm doing, I suppose has been suited to that, in terms of the type of data collection or the type of studies I've been doing recently. So it you know, a lot of it is luck and a lot of it is a huge amount of support from a lot of people. You know, I obviously train an awful lot as well on top of that. So it's just it's figuring out where you can fit what you need to do in. Skills that I think most of us have developed during a PhD, but you just end up having to hone a little bit better.
Louise Holden 6:34
So Alison, can you talk to us about your current research?
Alison Keogh 6:36
Yes. So, since starting in the Insight center in UCD, I've been in the personal sensing group with Brian Caulfield, and I suppose I found myself kind of mainly focusing on the idea of usability and the human factor side of wearable technology applied to healthcare. So I've done kind of a range of different studies, some of them are linking with industry partners, such as Novartis, where, you know, there might be a couple of new sensors that are on their horizon that they want to test out, you know, is it worth us going further with these for our clinical trials, so you know, we'll do a friendly trial with with healthy adults and test them out for a week or so. Currently, the project I'm on is called Mobilise D, which is an IMI funded, multi European site, huge project that is looking to develop digital biomarkers of real world walking. But again, I suppose, when we have cohorts of Parkinson's, MS, and COPD, or a range of illnesses, basically, we want to make sure that what we're asking these people to wear is comfortable, that they're willing to have their movement on their health care tracked by researchers and healthcare professionals, what they think about technology, you know, all these kind of range of things, it's really important for us to know, because I suppose it paints a picture as to where we can go with this, what's feasible, what's realistic, and it brings the patient voice into it. So it means then, you know, I'm kind of looking not only at the individuals, but also a little bit wider. So we're asking the researchers opinions, we're doing a couple of literature reviews on it. So there's a different kind of number of different studies that are involved in that particular project. So that's kind of taking up most of my time at the moment.
Grainne Faller 8:21
In terms of this sort of research. Is this it for your career? Is this what you enjoy? I know from from your bio, that you practice as a physiotherapist. So you have a lot of experience in a lot of different kind of areas. I'm just interested in this research that you're doing at the moment is, is a it's a kind of a cliche question. But where are Where do you see yourself in 5 or10 years?
Alison Keogh 8:43
Yeah. And it's funny because when we were in undergrad and physio we were asked to actually do a project, where will you be in five years, and I had effectively ticked off everything in two and a half years, I was miserable. And if you told me, you'll be doing a PhD, I'd have laughed in your face. But as it happens, you're in your mid 20s, you grow so much, and you change so much as a person that what you think you want a 22 and what you actually want a 27 or 28 are different. So where do I see myself in five years, I don't actually tend to go that far in advance. But I suppose in terms of is if they said I, coming from a clinical background, I suppose have a passion for have doing research that is clinically relevant. And and that takes the patient into consideration. So I suppose that's one of the reasons that I've ended up kind of looking at the human side of things. I think the mix of skills that I now have, would probably, you know, if I so desire, lend myself into industry, and I suppose that is an option that is in my head. But for now, right now I'm very happy with where I am, for a multitude of reasons. I enjoy the research. I enjoy the environment, you know, so that fits really well with my hockey. I enjoy teaching or the academia side of things. I do a couple of modules in undergrad courses that I enjoy being part of. So I count myself lucky that I have the option of going into industry because of my skill set. Bosch even if I was to do that, I would want to make sure it's in some sort of clinically relevant area that is clickable and translatable into suppose real healthcare.
Grainne Faller 10:23
Alison Keogh, thank you so much for joining us today.
Louise Holden 10:29
Next up, we're going to talk to Ahmed Judah who did an internship with Insight at UCD.
Ahmed Jouda 10:51
So basically, at the end of second year of my computer science undergrad degree, I knew that I wanted to complete an internship in the summer, to apply my skills in the real world. So after contacting my network, 10s of people on LinkedIn and in person, I met Dr. Aonghus Lawlor from Insight, and he basically told me about this project. So the my main goal when I was speaking to him was that I didn't just want to code like I didn't want to go into pure software engineering, I wanted to also apply the technology and the skills into the real world and like actually make decisions. So he suggested this internship for me on a project called - so basically an app called Strava, which is a fitness app, gave their user data to Insight, and basically wanted us to do some cool projects with it. So one of the cool projects that he suggested I could work on was to maybe visualize training sessions for marathon runners to achieve their marathon goal time. And that's what I ended up working on.
Grainne Faller 12:05
And what was that experience like?
Ahmed Jouda 12:07
So definitely, with COVID, and like working from home, I think it could have been more immersive and more beneficial if it was in person. But I still had my weekly meetings with my mentor, Dr Aonghus, and I managed to complete the app. So it was it was very unique experience working from home, I'd say it was a bit tougher, not having not being surrounded with you know, the atmosphere of people working, and like that motivate motivates you. So I'd say the hardest part was staying motivated every day, throughout the 10 weeks, especially when you get stuck on a problem and like, you don't have immediate contact, I guess that was probably the hardest part of it.
Grainne Faller 12:52
And in terms of the benefits of doing an internship like that, at this stage, you're in third year aren't you? What did you feel during the internship and brought to that? Like, what were the benefits for you?
Ahmed Jouda 13:04
So there's a few things probably the most significant one is it's made me sure that I do want to go into the data science stream in third year. And so I've come to I've come to UCD knowing that I am more leaning towards data science, but you know, you can't really know if you haven't been in the field working or like don't have like a proper project on it. So doing the project, in machine learning and data science with Insight, definitely assured me that I'm making the right decision, and I'm doing something that I like, that will be the first thing I'll say. And then the second thing is the technical aspect. So obviously, I've done all my modules have scored well in them. But to be honest, you can always score well and still not be that good. You know, because you know, there's a certain way you can beat exams, and I think many students have mastered that. But when you work in a on a project in a research center, such as Insight, you kind of have to know what you're doing, or you have to learn the technical tools and actually apply them. So it definitely enhanced my technical aspects as well.
Louise Holden 14:17
I just wanted to ask you quickly, what was the time commitment like for you over the summer?
Ahmed Jouda 14:23
So I would say it was full time. So it started it was 10 weeks approximately It started in early June and in early September, maybe end of August, actually. And I'll say within the week, it will be I say I put in at least three to five hours a day. So it was it was it was a solid. It was a solid full time job I would say.
Grainne Faller 14:46
You have ab interesting background yourself Ahmed. You're involved in all sorts of social entrepreneurship, social innovation. Can you tell us a bit about that?
Ahmed Jouda 14:53
I moved to Ireland five years ago, and it was then that I was first introduced to social entrepreneurship. So in 2016, I joined my school in Dublin after living in Kerry, and in that school, they were running a young social innovators project, which is young social innovators is basically a competition for schools, where projects compete to make social basically, projects that have to make a social impact in their own community. And my school had a particularly huge project. And one lunchtime, I was invited by a friend to attend one of their workshops, or so called, I think they were called mapping meetings. So I out of boredom, I decided to go that week. And I ended up being sucked into the core team of Global Citizens Mapping the Future. So Global Citizens Mapping the Future is basically a project that makes maps for countries in the developing world that can't afford Google Maps. So many people don't know this about Google Maps charges huge amounts of money for mapping the country. So for example, a country the size of Ireland would cost over 150 million euros just to map. So many countries wouldn't be able to pay that amount. And what we do is we use an open source website called OpenStreetMap. And we basically get students, and like people of the general public, people of the country itself to map their country. Basically, there's a satellite image and you just plot, so plot, say that's a tree. So you just put that that's a tree, you plot a river, you plot a road. And then there's obviously the senior mappers, who approve everything before they go on to the website. So that's the core of the project. And we managed to get. So we were working closely with the country of Lesotho, which our school had a sister school in the country there. And we sent transition year students to Lesotho, every year for the past 10 years. Unfortunately, it might not happen this year with COVID. But it will resume after hopefully. So what we do is we basically got Lesotho to become the best mapped country in Africa. And that was about two years ago. And once we finished that we started... So we realized that we have a huge amount of data from the maps that we created. So what we did was we started diversifying and opening up new side projects. So one of them was an emergency services, web app. So basically, there's a concept called the Golden Hour. And it says basically to in order to save most lives, you need to be able to reach emergency services within a maximum of an hour anywhere in the country. So what we did is we build a web app with a button on it that basically users can just press the button and then emergency services will be sent the coordinates of that user, and then they will reach them and making sure that they can reach them within the golden hour. So that was one of the side projects, there are other side projects on soil erosion and starting startups in the country, recommending certain types of business and so on, in order to increase the chances of a successful venture. So that's, that's the main idea behind global citizens. We won Young Social Innovators in 2016, won an award in 2017, as well. And then we went to present the project, the social enterprise word scope in the Ukraine, the stage or its scope, and we managed to win the first prize, we won the World Cup for Ireland in 2017. And from there we met with like the governments of Lesotho we went to Dáil Eireann and like, it was it was quite fun.
Grainne Faller 18:57
That's fantastic. And was that the beginning of your journey in data science and what you're doing now? Is that why you're doing what you're doing
Ahmed Jouda 19:04
I guess, subconsciously, like back then I had no idea what I wanted to do, but subconsciously, like I started, like, I loved maths all my life, all my life. And then data was something that I learned probably through this project at first. And I think I just never knew how to put everything together. And in the end of sixth year, I was just like, I think this is what I want to do. And it just all came together. And then I decided to do data science. So at the time, I had no idea. But subconsciously it was all built to go towards this.
Louise Holden 19:34
Ahmed you mentioned he moved to Ireland five years ago. Where did you return from?
Ahmed Jouda 19:37
So before Ireland, I was in the UAE, UAE in Dubai for seven years. Before that I was born in Iraq, for so those six years before that I'm originally from Palestine, but I never lived in Palestine. So yeah, that's that's the life journey. Yeah.
Louise Holden 19:56
So that's where the global perspective comes from.
Ahmed Jouda 20:00
Yeah.
Grainne Faller 20:04
In autumn of last year, Insight put a call out for proposals for a citizen science research project. A project called Crowd4Access won that and we spoke to the team behind it to find out more.
Louise Holden 20:18
We're about to talk to Venkatesh Gurum Munirathnam and Bianca Pereira, who are both working on our inaugural citizen science project for Insight, which we're very excited about as an organization. The project is called Crowd4Access. And I'm going to ask Bianca and Venkatesh to explain this. Bianca, would you like to start?
Bianca Pereira 20:36
Crowd4Access is a citizen science project. And the goal of this project is to for us, together researchers and citizens to discover what how, what are the areas of our city that are accessible? Because many people have experience of moving around the city. And so that was, you know, ah there is a path in my neighborhood or a place that I always go that is better or is worse. And that's a way for us to actually collect information about the whole city, before I go to that place? Is that place accessible, and accessible to whom? So that's the the main idea of the project.
Grainne Faller 21:13
And how did the idea for it come about?
Bianca Pereira 21:16
Well, first, I have interest on the area of equality, diversity and inclusion. And I always have this question in my mind, like how can we use technology to improve a quality diversity, inclusion, or also embed those values into technology. So at some point, last year, one of our colleagues, he created like a panel online, so that people can discuss it, one of the topics was about disabilities. And there were two people with disabilities, one who uses a wheelchair, and another person using a guide dog. And they were explaining how hard it is to move around the city. Sometimes the person wants to go to the footpath, but the footpath is too narrow, or there are no ramps. So someone needs to be walking in there go in the middle of the street with a wheelchair. So then my question was, how can technology support that? Support like for us to get information? What are the places that are accessible or not? Then there was this opportunity for the Insight call for funding for a citizen science project. So I said, well, there is something you can do there, because it's a data problem. We actually don't have the information even to start a discussion about disability. And then I contacted Venkatesh because Venkatesh like knows about image processing and say, how can you collect information through images and also by adding informatio n on the map. So that's why I talked with Venkatesh and he said, I really like the idea. That's really nice. And it's my area of research. So let's work on that. And we almost didn't send the proposal. I mean, the proposals not so good. But no, let's try let's try and then we got the grant at the end. So that was really good.
Louise Holden 22:57
And Vekatesh, you you are looking at the imagery side of things, and that's where the citizen science comes in, isn't it because the imagery is actually being collected by citizens?
Venkatesh Munirathnam 23:06
Yes. So, now, when we initially started like, when we talk about the image processing or data analytics, we require a huge amount of information to process it or to understand what kind of information we can extract from the images. Now, when we started like we just took one already existing data set so which is like mapillary wherein they already have captured from the dash cams. When we started observing all the images and the type of data what we what is existing there. So, there is a lot of difficulties. So that is with respect to the vehicles not with respect to the pedestrians or the person who is walking on the footpath. Then we thought, Okay, this is not going to work out completely. Then we started okay capturing ourselves. Okay, few of the images from from the, from the cameras using phones. And then we started uploading, and then we started looking or analyzing the images, like with respect to the pedestrians, or the wheelchair user or like when you walk on the footpath, how exactly you look. So that's where we started. And now we have captured around 20, 19,000 images, right? Thank you. Yep, now we have captured 19,000. And we have annotated, around 11,000 images, approximately
Grainne Faller 24:20
What stage are you at now? You to collect collected 19,000 images? How are those collected? And what how is this actually proceeding?
Bianca Pereira 24:27
Well we started the doing the mapping workshops. So there are two stages for the mapping workshops. First is to collect the images of the the location, especially Galway, Dublin, and Cork, but now that everything's online, we are not actually restricted to those locations. So people who are in villages or other towns around the country, they can also participate if they wish to. And the second one is, based on those images and then local knowledge, actually adding those things into a map. Okay, so then we started with a pilot group because we want to know if our learning materials are good enough if they are easy to understand if other people which are not the two of us can actually do that. So we were around eight people and eight people, right collected 19,900 pictures, okay. And then now we are organizing more workshops for today we have our second how to add information in the map, and then all those images were really like people with phones, walking in 10 minutes, let's say you can take 200 images to really say, oh, how is the whole path from given area to given area? And as part of the workshops, I think something that is important as well to cite is how to keep people's privacy. The person who has taken taken a picture, say, don't take a picture starting front of your house, for example? And how do they keep the privacy and security of other people, for example, don't take a picture where kids are playing every day. So those are the types of things that we talk in the workshops as well, because those things are very, very important for us. So yeah, but a very small group can actually cover a long distance and take a lot of pictures. So we have around 26 kilometers covered already. So if we can have more participation, then then it'll be like, our goal is really to have at least like one of those three seats fully covered. And that's what we are, we are aiming for.
Louise Holden 26:21
And can I ask is the objective of this, do I understand correctly is that at the end of it in the way that for example, we would refer to a sort of an app if you want to find out the quickest way to drive somewhere or cycle somewhere or walk somewhere? Now somebody will be able to find out what's the quickest way to get there in a wheelchair or get there with a guide dog? Is that correct?
Bianca Pereira 26:40
That's one of the things. The goal of the project itself is to collect the data because we don't have information. And this data, one of the premises is that we want this to be open like it's created and volunteered by the community and it needs to be provided open for the communities to use. So it will be provided and can be used to create applications like for some navigations for people using long canes, or wheelchairs. As you said, it could be used by with discussions with local governments who say well what are the neighborhoods that need more accessibility and for whom for which type of mobility? It could be used by commercial organizations if they wish to or non commercial organizations. So the idea of having the images as creative commons and the data itself, open data as well is that anyone can use it for multiple things. Even as part of our workshops already, we have other researchers interested in using the image to identify other elements inside the images. For example, Venkatesh, like we are now creating a prototype there for the recognition of tactile pavement and there are other people interested in cracks on the footpath or maybe like potholes to see what the water accumulates. Because for example, people who have low vision they have a huge problem that because they will not see the water they will just step on it. So if our data can be used by other researchers as well, so then the reach of what you're doing become much bigger and then become really a national effort almost which software that we're really glad that it can be part of
Louise Holden 28:14
And I just wanted to ask you briefly Venkatesh, you obviously work with imagery and image processing and so forth are you finding having such a great data set as it is a satisfying experience? Are you learning much from the process?
Venkatesh Munirathnam 28:27
Yes, computation or image processing is, it's kind of like you can't say you know everything like each each data is new to you and you'll learn a lot. So that's how, like say when I started this particular project, I thought okay, everything will be very simple, okay, you just need to identify a few stuff, but actually when you start it, but the learning process started from the way you capture the data. So when I started looking like okay, this is how it should be that. So, there is a lot of learning which is very much which is required, from my side to understand how exactly the person will will be thinking while crossing a road or navigating through a footpath. So that's how I should be thinking all these footstep to understand the footpath. So that's what one of the learning process that the process is, okay, I have the image. So how it should be like annotated. So, I can just have something okay. There is some something on my image, I just mark it and then I start using an AI technology for understanding what has been labeled or what has been marked on the image. So, I have to understand each and every aspect of the image and also the application. So where exactly my image processing or the contribution application will be used, and how effectively it can be used. So this kind of the learning process, so I have learnt a lot with this project.
Grainne Faller 29:52
That's it for episode five of The Insight Podcast. And that's it from us for this year. We're going to take a short break for two weeks. We will be back again in early January. With more great insights and great interviews from Insight. We'd like to thank all of the people who have agreed to be interviewed for the podcast and we'd like to thank all of the people who have listened and shared and supported and said such lovely things. We've had a blast doing it, and we can't wait to do it again in 2021. But for now, from me and Louise, Happy Christmas and have a great new year.
Louise Holden 30:25
This has been a snoring dog production on behalf of the Insight SFI Research Center for Data Analytics.
Transcribed by https://otter.ai