FloodWatch: Combining Wearable Tech + Disaster Alerts

Video thumbnail (Frame 0) Video thumbnail (Frame 1706) Video thumbnail (Frame 2492) Video thumbnail (Frame 5189) Video thumbnail (Frame 5802) Video thumbnail (Frame 6555) Video thumbnail (Frame 7924) Video thumbnail (Frame 9563) Video thumbnail (Frame 11439) Video thumbnail (Frame 12171) Video thumbnail (Frame 13350) Video thumbnail (Frame 15200) Video thumbnail (Frame 18346) Video thumbnail (Frame 19481) Video thumbnail (Frame 21792) Video thumbnail (Frame 24912) Video thumbnail (Frame 27498) Video thumbnail (Frame 28014) Video thumbnail (Frame 40042)
Video in TIB AV-Portal: FloodWatch: Combining Wearable Tech + Disaster Alerts

Formal Metadata

FloodWatch: Combining Wearable Tech + Disaster Alerts
Title of Series
Part Number
Number of Parts
CC Attribution 3.0 Germany:
You are free to use, adapt and copy, distribute and transmit the work or content in adapted or unchanged form for any legal purpose as long as the work is attributed to the author in the manner specified by the author or licensor.
Release Date

Content Metadata

Subject Area
FloodWatch is a prototype application that provides location-based alerts of flooding on Pebble Smartwatches. Building on the open source flood map PetaJakarta.org, FloodWatch aims to provide residents of Jakarta, Indonesia with time critical alerts of monsoon flooding via their wearable devices. PetaJakarta.org is a real-time flood map which integrates emergency services information, social media, citizen journalism and sensor data to provide real-time situational awareness for both residents and government agencies in Jakarta. Existing disaster maps and mobile alerts require the user to interact with their smartphone device to consume and interpret reports. Alternatively, connecting real-time disaster data to wearable devices for the provision of actionable intelligence "at a glance" reduces disruption to user activity, improving time to response. Furthermore, through the use of predefined locational filters the user is able to register for alerts for specific regions of the city providing a semi-automated user-centric interface to incoming reports of disaster. This presentation examines the motivation behind developing wearable tools for disaster response in the world's fastest growing city, and explores the software's underlying open source geospatial technologies.
Building Computer animation Information Interface (computing) Plastikkarte Denial-of-service attack Student's t-test Variance Relief
Rule of inference Touchscreen Mapping Projective plane Adaptive behavior Denial-of-service attack Bit Real-time operating system Denial-of-service attack Ultraviolet photoelectron spectroscopy Infinity Computer animation Term (mathematics) Computer cluster Software Representation (politics) Conditional-access module
Metre Functional (mathematics) Multiplication sign Sheaf (mathematics) Denial-of-service attack Real-time operating system Water vapor Disk read-and-write head Area Twitter Inference Different (Kate Ryan album) Endliche Modelltheorie Traffic reporting Logic gate Condition number Area Information Forcing (mathematics) Neighbourhood (graph theory) Projective plane Denial-of-service attack Cartesian coordinate system System call Type theory Uniform resource locator Digital photography Computer animation Website Self-organization
Area Dependent and independent variables Scaling (geometry) Mapping State of matter Multiplication sign Projective plane Water vapor Real-time operating system Uniform resource locator Different (Kate Ryan album) Website Traffic reporting Condition number Physical system
Point (geometry) Presentation of a group Open source Image resolution Orientation (vector space) Real-time operating system Distance Number Term (mathematics) Different (Kate Ryan album) Diagram Information Computer architecture Physical system User interface Information Mapping Planning Database Denial-of-service attack Line (geometry) Exterior algebra Process (computing) Algebra Personal digital assistant
Cross-platform Software developer Feedback Neighbourhood (graph theory) Denial-of-service attack Bit Plastikkarte Software bug Connected space Web 2.0 Type theory Uniform resource locator Prototype Computer animation Software Computer hardware Traffic reporting Computing platform Physical system
Multiplication sign 1 (number) Real-time operating system Mereology Medical imaging Electronic visual display Endliche Modelltheorie Office suite Physical system Email Channel capacity Mapping Repetition Keyboard shortcut Interior (topology) Menu (computing) Maxima and minima Instance (computer science) Connected space Type theory Process (computing) Uniform resource name Smartphone Moving average Permian Geometry Point (geometry) Statistics Service (economics) Control flow Plastikkarte Event horizon Computer hardware Traffic reporting Condition number Computer architecture Graph (mathematics) Scaling (geometry) Information Denial-of-service attack Volume (thermodynamics) Database System call Subject indexing Uniform resource locator Computer animation Integrated development environment Software File archiver Schmelze <Betrieb> Table (information)
Point (geometry) Emulator Uniform resource locator Uniform resource name Forcing (mathematics) Multiplication sign Dew point Electronic mailing list Denial-of-service attack Right angle Event horizon Emulation
Point (geometry) Group action Game controller Pay television Service (economics) Multiplication sign Direction (geometry) Real-time operating system Student's t-test Shape (magazine) Graph coloring Different (Kate Ryan album) Semiconductor memory Term (mathematics) Representation (politics) Energy level Circle Traffic reporting Form (programming) User interface Area Dependent and independent variables Mapping Information Polygon Moment (mathematics) Denial-of-service attack Integrated development environment Telecommunication Video game Configuration space 5 (number) Resultant Window
Logical constant Building Open source Code Multiplication sign Direction (geometry) Source code 3 (number) Water vapor Black box Mass Stack (abstract data type) Meeting/Interview Internetworking Term (mathematics) Computer hardware Software framework Data conversion Office suite Stability theory Metropolitan area network Forcing (mathematics) Cellular automaton Connected space Integrated development environment Freeware
Satellite Group action Multiplication sign Direction (geometry) 1 (number) Set (mathematics) Mereology Information privacy Logic synthesis Web 2.0 Mathematics Meeting/Interview Core dump Office suite Endliche Modelltheorie Data conversion Series (mathematics) Addressing mode Extension (kinesiology) Physical system Area Predictability Collaborationism Mapping Moment (mathematics) Feedback Data storage device Right angle Reading (process) Row (database) Point (geometry) Functional (mathematics) Service (economics) Open source Image resolution Code Element (mathematics) Power (physics) Twitter Revision control Causality String (computer science) Energy level Traffic reporting Plug-in (computing) User interface Shift operator Scaling (geometry) Validity (statistics) Information Consistency Projective plane Interactive television Mathematical analysis Denial-of-service attack Database Uniform resource locator Computer animation Software Personal digital assistant Information retrieval File archiver Video game Routing
all right the 2nd speaker Thomas wilderness and he's gonna introduce a quite interesting talk to talk at least for a topic you and it's about a flood watch so putting together wearables for people and disaster relief thanks so much and thanks to all of you for coming i'm I'm really impressed and Singer were up against some light on discussion right now so you have this many people in the room is great so I wanna talk to you about flood watch which was a really cool name of that we came up with this as we submitted false for and them like OK but finished building something and to show but we were really lucky and so we previously I was working in a smart infrastructure facilities University of Wollongong down in Australia and I had a great summer scholarship student and has who came to work with us for the summer to examine and new interfaces to uh to cartographic information during during disasters so
and just to give you a bit background to the project that I've been working on and and continue to work on now and I believe that
that the Jakarta the map Jakarta project which is all about producing a real-time representation of flooding in the city of Jakarta Indonesia and the city of Jakarta Indonesia is a fascinating place to understand in terms of climate adaptation because it is a ball and the city is now below sea level so you see in them up on the screen to
the the ocean to the north when you stand in the street at the harbor you look up at the fishing fleet and the difference between the head to type difference between the ground and the CD is as high as I am here today the flood wall as either with me so there's still a 20 million people live there and that's the population Australia in 1 megacity and all the water that falls during the monsoon season has to be pumped over the seawall so also the ball and condemn now this year because of El Nino and then learning in that there was no dry seasons of the monsoon has been running for about 10 10 months now and so predicting the weather flooding is it will happen is a function of infrastructure failure so funny happens in a place not because of fluvial conditions but because the wall collapses all the pomp explodes all the electricity supply for the gates and turned off so developing a probabilistic model of which neighborhoods will flood and in Jakarta is is near impossible because you're trying to probabilistic determine which 1 meter section of concrete wall that will collapse any given time so what we did is we turn aside tennis for inference had and so we crowdsource the locations of flooding through confirm reports in real time so we ask people on twitter we ask people on a government application call clue which is built by Google and we ask people on uh citizen journalism website called passing matter run by an organization called that don't we ask all of those people to tell us when it's floating in the neighborhood and so these are all the confirmed reports of people telling us is flooding this year so far since devotion to the project started in December and so we collect these in real time and this has been really successful we're running this for a couple of years now and I'm an idiot gives people an idea in real time for where conditions of are and so the
thing that we we've been doing this you have to talk about this a force for the last year is then movement piping that data into the control room of the Emergency Management Agency and now they take those information those tweets in real time and then they can delineate the area that the neighborhood those tweets are coming from and look at the photos and they say OK the flute is this high or its waist-high all high and so then we get a publicly available not that looks like this
where we have these individual reports but we also then have the real-time heights of water inside people's houses of the current so conditions across the whole city so red being the highest 150 all I down to 10 to this comedy that's 70 cm which is around about need and as the Informatics Section pushes those things out to the Jakarta government different responses are triggered for different areas depending on the height of the water so this has been pretty good and so even
that playing around with the state in Africa is running this project in developing this system and and so 1 of the things that we were very cautious of all countries over the for the beginning was that we should really the status or the Republic but how is the public going to consume it so it's all very well having a nice API and pushing the data to some systems so that you can download in your GIS but how do we give this data back to the community and see what really hard to build Oracle multi-stable not a website that allows you to on your mobile phone user gets into the location just like Google and you can see all the reports in there but if you look at the same on the desktop thing you see scale overview and so if you're a government user institution you might wanna see the whole city 1 time if you're an individual user then you see and you see that the map as displayed at your location with reports around and so on we have this
amazing architecture which uh this is what happens when you work with architects and designers the giver excitement started lines and points and tried to explain how your your system works and this and forcing the resolution hasn't come out so well but what we're saying here is all the sources of information coming in from the crowd of people on the ground being sucked in basic too big process database having a user interface to the government and then we pushed out to our API and their API has been been really well used and so we have a number of downstream uses the US government Australian government the World Bank all-consuming information for the for the for most the trimers unwisely but we're in this presentation at least we're really concerned ourselves not with how as GIS users we can get our data but actually giving back to the community in working with people who are on the ground and so information is really critical we seems nice use cases of people using on not to say all exploiting it's coming so I should put the kids from school early or I should change my around on the way home all I should I should leave it to you if a from traveling I should make alternative plans that's
great I but 1 of things require conscious of is that I'm I'm a job and so I think about everything in maps but not everybody in the world things so everything in maps and so presenting a flood map to someone who we saw the diagram before the lines on the problem that too complicated now we've got you many good these greatly is information that we've gone enough in real time and astrophysical really were excited about that and we're we're kind of giving ourselves a pat on the back but then I wonder what the user experiences on the ground of having to open your mobile phone and say OK any these orientate myself to the rest of the city which may not be may not be used to doing that and then we need to understand all these different layers of information are telling me and what's that criticality uh obviously red tends to be in danger but how dangerous and what does that mean compared to an individual point coming from 1 person and so 1 of the things that we've been thinking about is uh the evolutionary mobile technology now it's including wearable technologies and wearable devices so can we can we use some of these wearable devices so instead of having to get out your smart phone and and look at a map and interpret that information we could push alerts directly to users in real time that are geographically related to them but they're probably not thinking about in terms of a map that just thinking about in terms of proximity and so we created a
bit of software and sees that created the 1st prototype last summer called for watch and and so much is rarely simple takes uh uh a pebble smartwatch to run any any pebble with is 1 type of knowledge and run the pedals knowledge and it calls are API and then it is a filter to say OK give me all of the reports the within 5 km of The Watchers location which is determined from the user's phone it displays an analyst that's it that's all you and so for us this is very new and many
of you click on 1 of the reports unless you can open and you can read the text this is text that someone has tweeted out to say there is a flight at this location and so then we're pushing us through a whole other web of connectivity in our system pushing that directly to a different user this probably very nearby within 5 kilometers but just going on OK there's there's fighting over there in in that neighborhood and so we're quite excited about this but I'm also very interested and happy to talk with you laughter about feedback about what other people think because this is really for us a prototype this is the 1st and pieces of whether we ran this targeted to 1 piece of hardware platform or the development has been very much looking at trying to be mobile cross-platform and so which kind of interested about whether this is the right way to go or not but it's an interesting idea anomalous and so
that the architecture of the system in set so that the pedals knowledge is an interesting piece of hardware and for us because it's waterproof lots of the models they produce are waterproof which obviously when you're working flood conditions is quite important it's relatively affordable and I'd say there it's important to realize that cell phone penetration in Indonesia is 150 % so more than half of population have 2 phones many people have a smart phone call from the statistics on people using that we know the iPhone usage is quite high the so there is obviously that there is an issue of it's not going to be affordable for everybody not try develop this in roll out for the old Jakarta people that will be able to afford this type of technology but we know that we do have a use audience who is interested in this have technology and is able to purchase and and we feel that this is an interesting way to go over again I'm kind of do we bind ourselves to 1 piece of hardware and talk about that more in a 2nd is the way this works is you have a table provides you with sandbox JavaScript environment on the mobile phones and so the phone has this this JavaScript amount where you can say listen for data from a specific API and when that data is received filter and organize it and push it to the what and so we use this really nice
i've logical to have s by the guys books and this is chemical unavailable system this year we're actually we're executing turn on the phone in the sandbox JavaScript environment to do that location filtering in real time so we thought originally I was thinking well we could we should really push all of our processing back to the database which is what we do with everything else so we have an Amazon Web Services instance in Singapore which runs out of service and we find that works really well for us even during peak conditions we push all of our uh geometry processing down the database and it doesn't break a sweat you want since it scales and this post is when you go index and amateur columns works really really well and so then we start off playing with ideas will OK sees say I'm at this location and give me all the points that are on this occasion but then 1 of the issues you have this connection time because in Indonesia and into comedy often struggle with the 4 G even for gene network never drops out of research capacity and prayer time everyone stands outside the mosque teletext their friends in the network just like in the office and I were just the bond with systems the current cannot send an e-mail during that time it's just just peaks and troughs and so then the size of the graph our data we see the actually because we only show the most recent reports in the last hour on the map the volumes of information that we're that we're pushing out as you quite low so within an hour maybe we have uh maxim during a really peak flood event a couple hundred reports tends quite localized people at texting and fairly real time so we're able to and get rid of the reports images archive and we don't display on the map real-time from say 6 hours ago that would be too confusing so data volume is quite low so what we do here is we have to grapple reports in the whole city In this works really efficiently it's quite small data volume and then we just used to have processes a very rare quickly to say OK just filter them out the ones that around uh modifications of the thing that's kind of an interesting as we start to see some of these more geometry processing and stuff going on on the part of Clive devices as opposed to traditional doing this in the database so is an example so this is
actually at the end of last week was thinking about making my way to bond here force for GVC localized flooding event people on and messaging cinemas is about something and that inundation that's
causing flooding to be on the road and I love this use of m 0 geez inside which officers from inside the pictures of Mars chemical and so then and we see on the right hand side uh the uh wee firebrand emulator the Canon SDK provides an emulator we uh spilled the firm's location to be just outside of where this cluster finds is and we can see those points are then represented on the watch just by listing and ordered by time showing most recent 1st 100 times 10
to 12 going so few things to think about for the future so the other thing I have really got my head around yet is we have this so we have this fantastic results not just of opinions of the people of the points of the reports confirm reports of learning we also have these problems that the government is producing and they're doing that and in near real time so in the control and saying yeah there's some pens and OK we see in the market up and then it starts pushed out my life straight but have another had onions how we take that data and provide a representation to it in this kind of wearable devices in terms of a fuel if you're in the polygon should that like a different user interface experience on the watch the and then a few outside the body and always and it did come you can represent the different levels in some ways I think is probably an interesting design questions and we greatly windows and students working on this is year a thing about working in a very tightly constrained design environment that the pedal watches is very uh as a free small processes of rare a small memory footprint and just about has some colors now just about animation and through blocking so I think but I think there's some interesting things that we could do in terms of alerting people to the fact that they're not only people's telling this what's around nearby also that the government has said OK now we we officially delineated as flooded and are all the actions that that come along with us and most people in the city will understand that once an area areas is delineated exploited then that means the government response of some form is on the way great I think
that is there in a 2nd thing I am also really interested in is through the pedal STK between the primal and the phone that the communication between the 2 we cannot see access all of the information about the geo-location provided by the mobile phone including direction and the accelerometer so tend all by moving through the city just like with Google traffic not really bothered if there's a traffic jam behind you but you are below this is 1 ahead of you so I think a really exciting prospect that say of all instead of just showing the reports in this circle 5 kilometers start a shape that and so we're still doing stuff geometrically but were also then looking ahead and saying if we're traveling a certain speed and really wanna before forecasting out to sea water pores ahead of us in that direction 5 10 15 kilometers away as opposed to just showing the things I'm mean arrival in a minute you know this is so how much time is available to someone who is in a vehicle to be elected to the fact that funding is ahead also on say this is the moment our flood watches just passive so it's just you would open and it's pulling the reports in a nearby this year and have released their and their their timeline API which allows you to push alerts to the what and so you you subscribe to their API new push them data and they push it to all users and so we could start to do a really nice thing say OK so a new polygon has been designated push out of all of the users in Jakarta but then filtered out by location and show it to it's in North Dakota so only show to people who are in in North Dakota time or even you can have a subscription service where you could say In the configuration for that is say and and I want to always be alerted about this area because that's where the school is and that's where my kids on so any time you regardless of where I in the city I was once again let's from from that region and so that's kind of interesting how we build that mapping inside 1st to do the filtering in our back and not quite sure of and then add color
challenges and we have now a better development environment than when we 1st started a few years ago which is great now of this team is very well and we have a fairly stable internet connection although still can can be trying times unlikely that the office has not yet flooded as in the water hasn't quite come in and so it is pretty well but and for me as a massive force G of we build all this stuff that we have talked about all and been built with free and open source stack and that was the real winner for us because we went and had a conversation the government and other NGOs on the ground and of course the 1st question often is how much does it cost and so we will as well as a constant free it's open source you can take the if you want take the source code ammonia cells if the different government somewhere else in the world or in Indonesia wants to use it and they can do and we make the data open as well on all of the data is is archived in his as is is actually recorded the National Library of Australia and so all everything is is open but now was time to work with someone you working with with a piece of hardware here and it's not open so a black box it's closed source is not open source but I kind of feel that this is a an interesting direction for us to move in but what do we do what's the challenge for us as a community in terms of all of these devices are now coming out the coming onto the market we want to be using our free and open source that otherwise it is going to be overrun in a deluge of of Google and Apple smartwatch so how do I build start to build open source stuff but in this proprietary framework and I don't have an answer for that year the code for that is open source but it still runs within this black box environment we can't see inside that which is said to me is is a real shame so I'm gonna there
and say thanks very much for listening and that was interesting that few the 1st Christians had this is really cool and I think it follows the USA and I had my smartphone and
I have this set and that the 1st thing that I would really want to do it is to be able to get from my house to my grandmother's has returned to the civil and to have routing information say what's the what's the safest route your what's the safest chance that the really neat I agree I think that sounds like an also idea and and all the parts of that now you know we've got OpenStreetMap so it's there is no national map of of Indonesia really and so the World Bank funded the matter OpenStreetMap team to make the de-facto map which is OpenStreetMap so every government agency uses openstreetmap which is amazing so that's what evidence based on this and so that all it's all in the united data is in that the the rows in the direction we could do it and I think that that might be beyond what we can do as a small team but then I think we've got the a we got API functionality now we can push things simple things and that would be really interested in something of people building on top of what what we've done now be great the way so what is the kind of culture of flight privacy in Jakarta so the question 1 other the issue of like I feel like if this is in in the US people would be like a want the knowing where I'm being at all times you know I don't want is apt to know my location necessarily but on is is the same as the other issues about that yes to to an extent but not in the not perceived in the same way as it is in the west I think that some the the use of the fact that we have a very strong collaboration both of the government but also community groups on the ground gives a set at a level of trust that we've managed to to acquire by working with those people and come and I think that the and the open source was it's been a really strong tool in that discussion where we said actually we don't store your location and actually we have all of the the user so that we we use so we can't go back to see who said what so the thing is a tricky 1 because actually that all the information is available on Twitter and is available in the Twitter archive consistent with what what we do with and so we're trying we try as much as possible not to build a database of who said what when and so we separated the uses and names from the uh tweet try to least as much as possible it's not always possible and because you want that you want to reply to users and say thank you for your report is a good example of the trade for us of doing this using all who that people the appliances facts and so and the I think it's a it's an ongoing conversation something that we we try to evaluate as we go along but and we we often say well you know the codes that 7 so as you can see already in with the data is open at the end of each floods and monsoon season we publish the data online so people can see it and that that's anonymous and but you could just go and search for that string on it anyway and see you know and so it's interesting ones that I 2nd so related questions like we run into any other issues where but I like a lack of information about flooding area or maybe possibly like a mistake in in the in the reporting has led as any issues yeah and they're they're having a doesn't have been any and mistakes 1st say from the control room but what they're 1 of the things that we have seen is that when so marking areas flooded and then if that team um is really really busy with something else happens then that area can often be left footage and it is no time opposite experiments human-controlled at all times and so that you can sometimes get overlooked elicit if there's a shift change officers manned 24 hours and so this is actually a that's a human interface elements of the of how we design the software for the control room maybe we need is at a time and it pops up saying all did you work is this area still flooded but that could also be really annoying right if like if it if it was against causes a little busy doing these things that online we start with public and no no and mistakes as such but that then people started saying well this area has been voted for 24 hours now in such a clear that that's probably just because of that little areas has been overlooked when the new shift is coming and then re-checked everything of thank you for your valuable information for the benefit indigenous so whole teams developed series given genomic analysis synthesis system on the web so I dealing with the approval was used for for 1 of these core the situation like this so can be applied for full our systems that in the future that teach assimilation systems so if I do just to check I understand you correctly or the question you mean if you got a prediction of where things will happen in the future like the people of West either the time to ours real-time sorry what the time to lead to a always select the current to 2 hours so like a prediction of things in the future of his life I think it could be and as a as a so said in at anatomy the use case in Jakarta because they're that coming of prediction on on actor 1 I have ever seen and they're just on the previous talk little a lot using satellite imagery which we great but the streets of the streets and that which is less than the resolution of the currently available supplementary so I think yes it could it could do exactly what he's saying if we had a more well understood to be predictive like and send out those warnings and the moment the weather service issue the recall morning for rainfall and heavy rainfall and also if the gages upstream a very high if those 2 things occurred and then so that's that's you know alerts on the scale of half a sitting you know 40 million people the following with use of supply misery and and the power learning especially later great time talk after some session using and high end sensor readings and talk about of audiences and the interaction between the work you're done and uh where you been doing and the work on in a safe and q just plug in version has been use a lot of developed originally and it is in the news in Indonesia because of an also that they're looking at modeling in a predictive way enchanted to assess the potential impacts and how they consider potential for kind of feedback and interaction between the 2 projects yeah thanks so uh in a safe consumes our data as the validation model now so we are so we are the definitive of where funding is and consumers in an archive it suggested to do damage assessment predictions and also to check against uh models not necessarily of where floating uh predictions of Wesseling will be the predictions of what resources are required to just kind of went to see if they and they have a they just consume directly from our API and makes a cluster in so far as I know 1 last quick question of very quickly I was just curious about the price of those what users of the retrieval is that the reason you chose them as those as that of a smart watches yeah we're all academics and a few of us have 1 so that tells you what the price points and the sub 100 US dollars for 1 of these Peter was yeah right so thanks again