We're sorry but this page doesn't work properly without JavaScript enabled. Please enable it to continue.
Feedback

Open Activity Trackers for Research

00:00

Formal Metadata

Title
Open Activity Trackers for Research
Title of Series
Number of Parts
49
Author
License
CC Attribution 4.0 International:
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.
Identifiers
Publisher
Release Date
Language

Content Metadata

Subject Area
Genre
Abstract
Activity trackers are used in a wide range of research, from movement science to psychiatry, from criminology to rehabilitation science. In this talk, we will outline the requirements for activity trackers to be used as research devices. Legal and design aspects will be highlighted as well as the view from data science and field research. We will show how an open source alternative to the currently used devices is a ( is the only?) way forward. Activity trackers, such as fitbits are widely used amongst researchers to collect physiological and behavioral data in lab settings as well as in natural settings. In this talk, we will focus on data collection in natural settings and first outline the requirements a research device has to fulfill. More specifically, we will highlight the data privacy and protection aspects, as well as aspects of data collection that allows not only social scientists to use the data but also data scientists to learn from the collected data. In the second half of the talk, we will first present the platform we developed to collect data with a Samsung customer grade fitness tracker and then suggest an open source and open hardware solution to overcome the obstacles that researchers currently face. The first design decisions to create a community lead platform are presented and the design process highlighted. In the last part of the talk, the audience is invited to contribute with feedback in a structured manner, to make the next steps towards a modular and open Fitnesstracker for research purposes. The work on the Samsung platform is done in cooperation with Peter Bosch, Lieuwe Rooijakkers and Frederick van der Meulen (all CS students at Leiden University). The work on the process design has been done together with Assia Kraan (Hogeschool Amsterdam), Ricarda Proppert (Leiden University) and Klodiana-Daphne Tona (Leiden University and Medical Center).
23
Thumbnail
20:06
25
Thumbnail
38:54
27
31
Thumbnail
50:36
32
Point cloudOpen setSign (mathematics)1 (number)Open setProjective planeMultilaterationData miningPerfect groupUniverse (mathematics)Envelope (mathematics)LoginStudent's t-testComputer animation
Computer hardwareSoftwareComputer configurationOpen setInformation technology consultingComputer fontComputer hardwareTrailSlide ruleComputer configurationOcean currentComputer animation
Open setPoint cloudGroup actionInformation privacyVariable (mathematics)Set (mathematics)BitArchaeological field surveyObservational studyQuicksortInheritance (object-oriented programming)SpiralPhysical systemCASE <Informatik>Prisoner's dilemmaFaculty (division)State observerShared memoryUniverse (mathematics)Flow separationRight angleLevel (video gaming)InformationTouch typingDecimalClosed setOrder (biology)1 (number)Physical lawHeegaard splittingStudent's t-testInternet forumNP-hardDrop (liquid)PlastikkarteGame theoryRegulator geneProjective planeBarrelled spaceMereologyEntire functionWeightNatural numberProcess (computing)Wechselseitige InformationComputer animation
Point cloudOpen setVideo trackingBit rateBlogMoving averageComputer configurationSoftwareComputer hardwareGradientCuboidMereologyLevel (video gaming)Projective planeMultiplication signInformationMassLimit (category theory)WeightHash functionNetwork topologyLoginExecution unitSystem callMoment (mathematics)Latent heatGroup actionSet (mathematics)Real numberGoodness of fitComplex (psychology)Semiconductor memoryTask (computing)Student's t-testBit rateTrailIntegrated development environmentQuicksortPrisoner's dilemmaWechselseitige InformationVideo gameUniform resource locatorElectronic mailing listComputer programmingOrder (biology)Software frameworkSheaf (mathematics)Process (computing)State of matterRange (statistics)Computing platformCondition numberSelf-organizationNoise (electronics)Subject indexingObservational studyFlow separationUniverse (mathematics)Variable (mathematics)GradientFeedbackGame controllerPoint (geometry)Open sourceSoftware development kitMobile WebPattern languageComputer animation
SoftwareComputer configurationComputer hardwareOpen setSoftware development kitTemporal logicPoint cloudScalabilityGroup actionMobile WebFeasibility studyParameter (computer programming)Level (video gaming)Real numberGradientComputing platformSoftware development kitLevel (video gaming)Process (computing)Computing platformTotal S.A.Multiplication signPublic key certificateMathematicsObject (grammar)Connectivity (graph theory)Pattern languageWeb 2.0FrequencyMobile appGradientInformationLimit (category theory)AuthorizationCartesian coordinate systemMaxima and minimaGroup actionOnline helpVariable (mathematics)Software developerReal numberMetric systemFood energyObservational studyApproximationCASE <Informatik>Axiom of choiceVideo gameBit rateMeasurementSubsetScalabilityEstimator2 (number)Computer configurationRight angleChaos (cosmogony)Incidence algebraBarrelled spaceWeightCuboidRouter (computing)State of matterComputerDigital electronicsOffice suiteRational numberDreizehnMonster groupEqualiser (mathematics)Student's t-testDisk read-and-write headSensitivity analysisFitness functionVotingDecision theoryComputer animation
BlogMoving averageComputer configurationSoftwareComputer hardwareOpen setStatement (computer science)Information privacyTerm (mathematics)Bit ratePoint cloudRevision controlComputer hardwareRevision controlProjective planeAndroid (robot)Information privacyObservational studyPoint (geometry)Statement (computer science)Standard deviationOpen setComputerGoodness of fitElectronic mailing listPattern languageBlack boxEndliche ModelltheorieMathematicsConsistencyGroup actionBit rateSound effectClosed setCASE <Informatik>Operating systemMobile appView (database)AlgorithmComputer configurationPlanningSocial classFirmwareOffice suiteInformationQuicksortCondition numberFiber bundleProcess (computing)Shared memoryPresentation of a groupTerm (mathematics)MereologyInternet forumService (economics)Multiplication signTrailState of matterPurchasingLoginGoogolChemical equationMenu (computing)EstimatorFood energyProduct (business)WebsiteVariable (mathematics)Line (geometry)Software testingWeightSoftwareTwitterMetropolitan area networkComputer animation
Open setDistanceMoving averagePoint cloudNegative numberAndroid (robot)Independence (probability theory)Shift operatorInformation privacyDecision theoryProduct (business)Computer configurationQuicksortLocal area networkMereologyStudent's t-testComputer configurationProgrammer (hardware)DistanceComputing platformResultantCombinational logicPlastikkarteMobile appPersonal digital assistantGroup actionProjective planeGoodness of fitSoftware engineeringOpen sourceBit rateConstructor (object-oriented programming)Standard deviationMultiplication signIndependence (probability theory)Shift operatorAnalytic continuationMathematicsOpen setProduct (business)LaptopGame controllerVariable (mathematics)Field (computer science)Order (biology)WordMoment (mathematics)Data centerTelecommunicationInformation privacyMusical ensembleSocial classComputer clusterAdaptive behaviorService (economics)Online helpComputer programmingWeightPhysical systemSurgerySpacetimeError messageComputer animation
Open setSoftware developerSampling (statistics)InformationOpen set2 (number)Multiplication signInternetworkingInheritance (object-oriented programming)Projective planeComputer programmingShared memoryComputer hardwareStatement (computer science)Level (video gaming)BitOnline chatCASE <Informatik>Online helpRight angleState of matterPhysical systemGroup actionWorkstation <Musikinstrument>Presentation of a groupAddress spaceVariable (mathematics)EmailRadiusPower (physics)Form (programming)Computing platformOperating systemUniform resource locatorArithmetic meanIdentity managementDrop (liquid)Software engineeringThread (computing)SoftwareUniverse (mathematics)Perfect groupFirmwareFeedbackFiber bundleSlide ruleOperator (mathematics)Mobile appSubgroupDecision theoryClosed setQuicksortComputer animation
Point cloudJSONXMLUML
Transcript: English(auto-generated)
Welcome to the talk, Open Trackers for Open Science with Daniela Gavins. And here, have fun. Okay. Hey, welcome everyone.
And I see people are actually trickling in so perfectly. Perfect that we waited for a few minutes. I'm Daniela Gavins. I work at Leiden University in the Netherlands as a PhD student, where I'm working on a project related to data mining and data science in general and very broad strokes.
I get into my research a little bit later. Today, I'm going to present on Open Trackers for Open Science. What do I mean with trackers? I mean activity trackers. And then for Open Science, we'll look into a few definitions that there are and see what activity trackers,
which impact activity trackers and the use of activity trackers in research as research devices can have on our perception of open science and doing open science
with activity trackers. Where, when does it work and when doesn't it work and what do we need for it to work properly? If everything went well, you can see the second slide now, which is the outline of my talk.
I will start introducing what it means for us to work with activity trackers in behavioral research and in medical research. Then I'm going to give you a very brief overview of what hardware and software options currently exist and what people are using.
And then I will give you solutions to the problems that actually are attached to the current hardware and software options. And then I'm going to ask the big question, what's next? And there, I hope to get everyone involved in thinking along of what are viable options for activity
tracking for research. But let's look at what activity trackers are. When I say activity trackers, I mean those wrist-worn variables or variable devices.
And to see if people are only locked in or actually listening, I want to ask you one question, which is, are you wearing an activity tracker at the moment? And we have a poll option, which is open right now.
So let's see if people are answering. Yes. Yes. Let's see. That was half-half, I guess, for now, what we see. Yeah.
Well, I don't wear any. I don't even wear a watch. But yeah, so this is very much. Can you maybe drop in the public chat what kind of variable you were using? That'd be also super interesting. I'll close the poll. I did not close the poll.
Anyone want to share what kind of variables they're using? Gorman? Anyone with Apple watches? Probably not.
Anyone with a device running asteroid OS? I was hoping to find people that work with, that have a watch with an asteroid OS on it. But all right.
Now that we're a bit awake, let's dive into three personas. I want to share with you how participants in research experience using activity trackers for research. So I'm going to give you three personas,
and then we're going to walk through what they're experiencing when joining such research. So persona one is Mark. Mark has two children, aged eight years. They are in primary school, and they come home
with a letter from their school, with an invitation letter to join a study where scientists from the behavioral science faculty will join the kids on the playground
and take observations. So they will take notes, paper and pen. They will also send around questionnaires and surveys that the children can fill in and that also the parents will have to fill in. And the researchers say that they are going to outfit the children with a variable
that they will wear for one week during the day. So when the children come into school, they will get a variable, and when they leave school, they will leave the variable at school. So this is persona one. Persona two is Janine.
Janine is just out of prison. She has a story or history of being in and out of prison, actually, and her coach, so she's working with a coach to sort of get out of the spiral of criminal behavior.
With her coach, she found out that there's a few triggers that might lead to her behaving in a negative or unlawful way. And those triggers are that she does not get out of bed for several days in a row,
and that she is in touch with a set of people that actually are sort of not good for her, that will lead her back into criminal behavior. And so the coach tells her, maybe you should join a study where we are going to actually look
and use a wearable watch and a smartwatch. And whenever these triggers, so this not getting out of bed and contacting those set of people happens, someone that you trust or the coach will be notified.
So it's a system where she wears the watch and a coach gets notified so that she does not spiral back into criminal behavior. That's persona two. Persona three is Carla. Carla's mom is actually in a nursing home
because she has dementia and cannot live at home anymore. And also here, it's the case of an invitation letter by a university saying some researchers are interested in how or how much activity
patients with dementia show, if it's more sedentary behavior or very active behavior. And they want to use, as well as with the children, they want to use a wearable device that Carla's mom will wear during the day to track not only the activity or activity levels of hers,
but also where she is during the day and if she is using the park that surrounds the nursing home. So I hope the three studies are more or less clear. So we have the left one here, Mark is the dad with two kids in primary school.
There, we want to use the variables to check how children are playing on the playground, what they're doing the entire day. We have Janine who might wanna use the watch to check if some criminal behavior might reoccur or not.
And we have Carla, where it's about the dementia care patients and how much activity they displayed during the day. So my question, and I'm gonna leave this here in the chat as well, is what do you think are the biggest alarm bells?
Like if you were Carla and were asked to outfit your mom with a watch, what would be the big alarm bells that would start ringing in your head? And would you give people,
would you give the researchers, would you allow them to use a watch on your mom or on your child? Would you yourself use a watch if you were in such a coaching process? And what do you wanna know?
What more information do you need to decide? So if you, I see that one person definitely has found a way to the HackMD. You can just click on the link and edit the MD.
And I can see that for the Children's Behaviors Project, there's already people editing.
The behavioral impact on the child, yeah. Children getting used to tracking. So that's an interesting one, actually. I haven't thought about that one. The children could just get used to wearing such a watch and being like followed, so to speak, by someone else.
Yeah, privacy data sharing is something that is definitely interesting. And regarding the Dementia Care Project,
someone is writing about the consent. Will she understand or will the mom then understand the consequences of the research? Yes, interesting. Okay, I'll leave the HackMD.
We can leave it open and see who else or if people want to contribute later. Oh, there's one more. Yeah, so I guess someone wrote and split between groups of parents. So probably that one parent group says, yes, let's do the thing with the watches
and another group says, no, let's not do that. You might be wondering if people actually come up with such studies and really want to conduct them. Well, I'm one of those.
So I'm working on the Dementia Care Project where we're interested to find out how much dementia care patients actually move during the day, so how much activity do they display. And we also want to know if and how they're using a park
that surrounds the nursing home. And the clue or the interesting point here is that this nursing home is actually going to open the doors of the dementia care unit so that patients that are usually locked
into their care unit and cannot leave that unit without someone accompanying them are now free to leave that unit or that housing residential complex and can use the park. And so this is,
and no one knows really what's going to happen. So it's interesting to look at that. And we are looking into using variables for that research project. The children is also a project here in Leiden where up until now they're using mostly proximity sensors or only proximity sensors to find out
which children are playing with each other on the playground. And they also want to incorporate more information about how children behave on the playground and also track where they are on the playground because it's interesting for them to find out more
about unstructured playtime for young children and what that does to the children. And the third one, oh yeah, the third one was the ex-detainee. We have here an organization called Exodus
and they worked actually with the Hucho School in Amsterdam on a project where they wanted to find out or design a variable device that would give ex-detainees sort of a feedback
on how they're doing with their own goals and the goal being mostly staying out of prison. So these projects actually exist. This is what I want to do. This is what I want to get across. The baseline is those projects assist.
People want to use activity trackers for various reasons. Some of these reasons are on this slide. A, the tracking of activity of heart rate, location, interactions, but also something that is called momentary emotional assessment,
which is short questions that are being asked at various times per day or various times per week. And people are asked how they're doing basically. Then the trackers are attractive because they are passive and almost non-intrusive. Wearing a watch is something that you can ask people to do.
You can use them for longitudinal studies or for several weeks and they will gather real life data and real life data is data that is collected outside of the lab and that is very rich. It's also not the most beautiful data to work with
because it comes with all kinds of noise, but it's very, very rich. So yes, people want this data and people want to work with the variables for these reasons. Let's look at the three projects that I just mentioned
as to what kind of data we want to collect and what the participants are. And the data that we want to collect is some sort of activity data, which we can infer from accelerometry. EMA, so the emotional momentary assessment
would be something that might be implemented in the children's project. Location information would be something that we're looking at in all three projects. Then for the ex-detainees, something like call logs or messaging logs would be interesting.
When you look at the participants age, we can say that the children are generally younger than 13. They're all primary school kids. For the ex-detainees, it's anything, the range is pretty big. And then we have our geriatric patients
that are well over 60 and 65, most of them anyways. And then I have two more points of information that are coming in at a later stage, which is somatic health and mental health. For the children, we can assume
that they're mostly healthy. There are hard of hearing children in some of the classrooms. And then for the ex-detainees, we would say they're probably generally healthy. In the nursing home, we have old people. So they are geriatric patients.
They will have heart conditions, for example. They will have mobility that is somehow limited. Then from the mental health aspect, again, the children, we have some special education schools in the data collection process there.
So that can influence how people are moving, what kind of movement patterns they show. For the ex-detainees, the chances are that you get a psychiatric patients in there. Not all of them will be, but some might.
And in a nursing home, it's a dementia care project. So there are people with dementia in that group. Now, before I present what kind of solutions there are,
or what kind of say big tech solutions are often used, I want to say that not, I want to say that there are good reasons for using big tech and some, and we need to know the reasons why people are using those big tech solutions
to do better with an open source solution. So I'm not here to bash Apple. That's basically what I want to say. First up are our medical research devices. Medical research devices give us high quality of the data.
They give us access to the raw data. They are tried and tested. The biggest problem here is that they are really made for lab experiments of controlled environments.
Plus the medical research devices usually only track one, like they have one sensor or two sensors max. They are not smartwatches. They are usually very bulky. Yeah, bulky boxes basically that you wear on your wrist.
On the other side, we have consumer grade devices and I sort of summarize them, per platform that they're using. I hope that makes more or less sense. So we have Apple watch, Fitbit and Garmin, the big three.
And then we have Android watches. We have Tizen, so that's Samsung watches or the Tizen platform used on Samsung watches. And we have asteroid OS. First up are the big three, because most research that's being done
is done with those big three. And I'll be very brief here because also time. So Apple watch, you can use an Apple research app, which some universities actually using
or collaborating with Apple. It's probably pretty difficult to get into such a program. Then we can use the health kit and care kit frameworks that they are offering. Here you have access to all kinds of health related data from the users. It is limited to a specific set of tasks
that users can do. And yeah, limited to the health data that's being shared throughout the platform. You could also build your own application for Apple watch. Here again, you're limited by the certification process
that you have to go through to actually launch the app. So for example, background, logging in the background at high frequency could drain the battery so much that they would not allow you to certify the app that you're developing.
Fitbit works together with Fitterbase. And Fitterbase is a company providing help to researchers to use Fitbit as a research device. You can use their web API and you can bulk download information per user.
So you could basically open a user account per participant and then ask the participant to send you the data or download it for them. That would be an option. Both the web API and the Fitterbase solution have the problem or the limitation
that you cannot get data at a very high granular level. It's actually quite coarse. Then Garmin, also Garmin works with Fitterbase. Also, they have a web API, it's called Health API.
And they also offer a health SDK. The SDK allows you to write native apps for Garmin watches. I did not look into the SDK. So that's that.
The big downside of using these big tech approaches is the granularity. On a temporal level, so for example, you would only get a heart rate measurement or estimate per minute instead of per 10 seconds.
You would get a maximum of accelerated data or no access to raw accelerated data and only to the compounds. And with the compounds, I mean the activity classifications that are usually done as in the activities such as swimming, walking,
sitting still. And so those are the compounds that I mean. And so the granularity is very limited in these approaches. You have web APIs that you can use, but then you depend on the tech companies actually allowing you to use the web APIs.
And if they are shutting down those APIs, then you don't get your information anymore. And possibly certification issues. For example, if you're draining the battery too much, though you might not care draining the battery because you only need information
for one and a half hours. But yeah, so this is just certification issues that can come along. What is it that draws people still to using Fitbit or to using Garmin? One is the availability and scalability of research.
So for example, in Germany, you have Corona DARTENSPEN with the A, which allows people to share their health related data with a national institute. And they want to estimate how Corona is spreading
throughout Germany. On a temporal and spatial component. They can use this. So the idea here is that you have, if you have enough people joining this Corona DARTENSPEN with the A data collection,
and if there is a signal in all of these compounds, then you will find it. That's basically the idea. And they can use this because there's just so many people using a Fitbit or a Garmin.
Another reason for using a consumer variable is a design, is the design choice. And here I wanna quote actually from a paper
written two years back where they worked with psychiatric patients. And so they write to enhance acceptability and minimize user burden and stigma and widely available consumer oriented technologies were therefore considered. So they talked to their participants,
suggested a few variables amongst which also the medical research devices mentioned earlier. And then the user groups favored the wrist-worn Fitbit charge due to its appearance as a lifestyle device, as opposed to a medical device that is acceptable
to both younger and older users and the ability to view metrics related to sleep activity via the Fitbit app. So here they work together with the participants and decided to use something that gives the participants
more than just being a tracker. On a similar note, I wanna mention again, the compounds. So what I mentioned before that you don't get raw data, but you get like activity, sleeping, activity, walking
or running or swimming. So is it that we can use those compounds to do research? And there the question is, what is your intent? So in the same paper, the authors write,
we suggest that those variables, those devices actually work with clinical prediction, depending on the questions that you're asking. Our goal is not to draw conclusions about, so in this case, it was only about sleep parameters,
such as total sleep time or sleep efficiency per se, because if they were interested in sleep time or sleep efficiency, they might use something else, something that is validated, but rather our objective is to ask whether changes in longitudinal rest activity patterns
at a within person level, captured using a verbal device, predicts deterioration in clinical status. So they are wondering how changes over a long time period predict deterioration in the patients.
They were not interested in how someone is sleeping, but how sleep patterns as approximated by the Fitbit, actually help them predict behavior or an episode of psychosis.
So it's about the intent basically. Same goes for the Corona Datenspände. So here you can see that they're collecting all kinds of information here, how many steps you've taken, how many calories burned, how many flight of stairs you've taken.
All of those are a compound information of something that you get out of the watch that you cannot really validate or you don't know if it's validated. It's an approximation of activity in general.
And here at the Corona Datenspände, as I mentioned before, if there is a signal, you will find it. That's the hope anyways. So in summary, we have solution for lab studies. So that's the medical devices.
They are bulky, but precision technology. We have solutions for big data studies, widespread consumer grade devices. You'll have access to summary statistics, perfectly fine. For real life data collection,
yes, if the intent is in agreement with what you're getting from the watch, then you can use the technology that is available. Now, when we look at our case studies,
the biggest problem I see is that we have age groups here that are below 13 and above 65. I don't know how the compounds or these activity classes are generated. I don't know if they are trained on kids' data or on elderly data.
For the elderly, we have geriatric patients. So if we're looking at heart rate estimations, for example, might be difficult if you have a heart condition or if many people in your study population have a heart condition. We have dementia care patients, so their movement patterns might be completely different
from a healthy population. So when we're looking at the compounds, not much we can do in these case studies. And then there was this privacy issue, right?
So when we looked at the alarm bells, where is the data going? Where is the data going? With whom is it shared? Do I have access to the data? Who else has access to the data? And so I think for the privacy part,
there's two questions that we should ask. Is which path does the data take from the variable to the computer of the researcher? And then another question that we have to ask is the privacy statement between the participant
and the producer or the company, or is that a privacy statement between the researcher and the producer? And this is something that any researcher has to take up with their privacy officer. And not all privacy officers know a lot about clouds,
for example, about data being sent around all kinds of service to then end up on the researcher's laptop. So from own experience, I would say, take time and plan ahead to talk this through
with your privacy officer. So at last, Open Science. I promised it at the beginning. And now I have another 10 minutes or so to actually dive into the Open Science part of my talk.
So the definition by Foster is, Open Science is the practice of science in such a way that others can collaborate and contribute, where research data, lab notes, and other research processes are freely available
under terms that enable reuse, redistribution, and reproduction of the research and its underlying data and methods. Let's look at this definition or take it apart. We want to collaborate and contribute. Big question is, can you collaborate,
can you contribute if a research is really locked into a closed project, basically, where you don't have access, such as an Apple Watch, an Apple Research project? I would say it's very difficult to collaborate
and contribute on such a project. Next question is, or the next point here is, under terms that enable reuse, given that we're talking here about proprietary data and models, proprietary, yeah, access to the data.
So that, I mean, that's the definition of proprietary. The terms are not open. That's basically it. And then the question is redistribution, reproduction of the research and its underlying data and methods.
Next question is, trying to reproduce something that is built on a black box algorithm on a model or on a compound, whatever you want to call it, will be very difficult, because you don't know if, for example,
this black box algorithm changes with updates of the watch. So say you have a longitudinal study, you have users use a Fitbit for like two years and they update their watch. It could be that the model that predicts
how well you sleep changes over those two years. So you don't have a consistency there. That's not what you want if you want to estimate any effects there in the data. So it's a no-no actually
for when you look at it from a robustness point of view. That being said, not every research project that, so research projects do not need to be open science to be good research projects, right? So there are reasons or situations
where you cannot adhere to the standards that I just mentioned, and they are still good and valuable research projects. But this is under the assumption that you want to do open science or open science as much as possible. I'm not the first one
or the only one thinking about these things. There's, for example, a preprint by Nelson et al. from this year on how to use heart rate in bio-behavioral research. And they have also a few good, a list of considerations that is very good.
So if you remembered the very one of the very first slides, there are three more hows that I did not mention yet.
And now we're getting into the discussion with you as listeners, as my audience, how can we open up those trackers and how can we open them up in a way that they allow for open science methodology
or open science approaches? For this I'm trying to give the presentation in moderation so the microphones can talk if you like. Yeah, give me another two minutes and I'll just use the three things
and then we can open up the discussion, I think. Okay. Yes, cool, thanks. So the first one is Wear OS or Android for wearables. It's kind of, I'm saying it's a cheapest option because there are just many wearables out there that run Android.
I found this app here, Vada, which can track accelerated gyroscope and light sensor. It's just a package and can be installed offline with no accompanying phone. So I found that very attractive.
I didn't try it yet. So maybe there's people in the group that have tried this or another app on their Android watches. And I'd be interested to know if this really works without having to connect or to send information
via an accompanying phone app. Yeah, well, you will be locked into this, the version of your, into one version of the operating system that is on your hardware.
But that is probably something that we have to get used to. The hardware and firmware is a bundle. It's a package that like sort of, yeah, you don't really get it out of each other apart, teasing apart, it's difficult. Again, Android watches,
you could probably also bulk download health data possible. Question again is, do you want to force your participants to have a Google account just to get their health data and share it with you? My guess is no. That would be my first reaction to it anyways.
Then I want to present our little homebrew. We used a Samsung watch and had, oh, there is text missing here. That's not nice. So to the left, you see a watch.
This is supposed to be a watch. To the right, this is supposed to be a laptop. And we have watched a device app and we have a command line interface for the laptop. And this is how the two communicate with each other and they communicate via a local network.
And the nice thing is that this gets us raw data. So we can actually do data science, which is great because that's what I'm supposed to do for my reset. And the cool thing is that it's possible to extend
to Bluetooth or proximity tracking, something like a heart rate and emotional momentary assessment because it's basically a smart watch and people can also interact with the watch. And the negative is that it's still under construction and it has taken us a long, long time.
You need good programmers to come up with a good solution here. It's all programmed in C. And so the first bulk of the work was done by third year bachelor students at our institute as part of their software engineering course.
And then we have secured two of the students to work for us as student assistants and continue work on the command line interface mainly. But yeah, it's for behavioral researchers,
behavioral scientists, it's not easy to find programmers who can do that work for them. And also the two students that we included in working for us, they are doing that not for the money because we can not afford too much.
Like we have a standard pay rate for student assistance and that is much lower than what they would get outside of the university, but they are doing it because they liked the open source idea behind it. So getting people to support you here is very difficult,
certainly for field researchers. Well, and with our solution, we are locked into Samsung devices. That's, yeah, that's a problem or maybe not, but that's how it is. Question, asteroid OS is also an option.
From what I understand, there's only limited lifestyle apps there. What other disadvantages are there regarding asteroid OS? This is something that I really wanna know from the community or from people here that are listening.
We are, there is, in Leiden or in the Netherlands, we're kind of working towards the idea of having such an open platform. And I already see that someone has shared, did you hear the rain in the background?
I have lots of rain and it's really loud. I hope, Ormole, do you hear me well still? Yes, I'm hearing you well and I don't hear the rain. Okay, good. Then I just continue working, it's really hard rain here. So just very, very quickly,
the mission of our little group that we came up with is to be or create an independent community of researchers and other stakeholders evoking a cultural shift towards more sustainable research. And this community works towards a common toolkit, whatever and however that will look like,
which is transparent, flexible to use and open for improvements and change. Sustainable research is better privacy, adaptable design, affordable and transparent. Transparency allows everyone in the community to take their own decisions and draw their own conclusions about the product. So have this control about their own,
yeah, the control is basically important. And this is a mission that we worked on together with a few, a bunch of people from the variables and practice community in the Netherlands.
So now I talked a lot for about 45 minutes or something like that. I hope my story made some sort of sense. And I want to ask you another poll, which is a truly open activity tracker an option?
Like, is that something that we should look into or should we just forget the idea? And let's make it a yes, no. And you can fill in open as you want to fill it in.
Yay, people are for it. So you know that I gonna ask how to do that in a while, right? Yeah, so we have like eight people saying, okay, this is absolutely an option. So now the question is,
is it worth exploring the asteroid OS plus custom small watch? Do you think that this is something and now how do I close the poll and open a poll again?
Or more help? How do you do it? By the last time. Yeah, but then now I getting the published polling results maybe once I publish them, maybe I can open a new one.
Yeah. Yeah. Okay. So yeah, so eight people said, yes, it's worth exploring the asteroid OS an open approach. So now I wanna know if it's worth to explore the Australia OS plus custom smartwatch.
And it's again, a yes and no. And please also use the public chat if you think that maybe another combination is useful.
Absolutely, asking the echo chamber here, everyone is like open, open, but yeah. I like echo chambers, then you can get, you hear what you wanna hear, right?
Anyways, let's go back to the poll. No, there is someone saying that the asteroid OS plus custom smartwatch is not a good idea. Why is that? If you want to share that,
or maybe you don't wanna share it, but I'd be very interested to hear who said no here.
So the person who said no did not, doesn't want to answer. And then the third question has already been answered in the public chat, I think.
So I see here open hack as an idea and GNU Health, which I don't know, but maybe we can open the floor to questions
and everyone, because we're only 10 people here, so. Yes, can we do? I can enable the participation. Yes. No, you're unable, you can talk if you like and or you write something in the chat if you like.
Yeah, or raise your hand if that is an option in here. Yes, I think it is. So what I would be interested is Cy Revolt, if he, she is still here, they are still here. The GNU Health, is that a project in Europe
or in like, how do you know about it? Okay, so you actually know a person who works on it. Okay, perfect. I'll definitely dive into that.
Is that in Germany, is that a German project? Okay, super.
Yeah, so thanks a lot for sharing that. Also, OpenHack is being shared here. So yeah, I don't really have a slide anymore, I'm done. I didn't even attach a slide on how you can contact me
or if, how to contact me is not included, but you can send me an email and my name is on the first slide and I think there's not so many Daniella Garvains. Certainly not in Leiden at university. Ormo, do you, I don't know how usually people
are doing this with questions. Yeah, normally it is so that the people can write something or they can ask something if you like. So they can unmute their microphone and ask the question or write it down on the chat, please.
Yeah, so I just left my email address in the chat. Perfect, thank you. For that, I hope my story made a little bit of sense. I honestly had problems in how to structure it in a way that like behavioral scientists, developers
and people doing data science kind of understand the entire jazz because I find it a quite multifaceted problem. But yeah, that was the challenge today, I guess. So if nobody have a question, maybe I have one.
Which thread are you using personally? No, so I don't use any device, any trackers because I don't know. You're not trusting it. I might trust them, but I don't wear watches. So I never wear one, that's the main thing.
But would you recommend one of the trackers? For private use, I never looked into them. Okay, there's a question that you answered.
I'm wondering whether we have more people in the group who have used or programmed variables for research. Yes. Is there a research software engineers maybe in here in the group listening in?
I'm sorry. Not too much yet. Software engineer, yeah. Yeah, I find it very difficult to find sort of,
because app development seems to be really a subgroup of developers and finding them to also communicate with researchers, I find that very tricky. So yes, yeah.
Okay, so I have to read out the questions from the chat. What I can say is that the closed state of the hardware doesn't help very much. We would need more lower level access. And then yeah, the firmware project
and the firmware or open firmware. Definitely. I think, no, I might have skipped that. Which is, so whenever you're deciding for a platform, you're deciding basically for a bundle that combines the hardware, the firmware,
the what else is there, the operating system, the software stack that is on the operating system. So you sort of decide for such a sandwich or some, yeah, an entire bundle that you can very hard take apart or tease apart.
And I think that is the main problem. Or this is the main problem that you're not use, like you're not deciding for, I want sensor A and operating system B, but you're really deciding for bundle.
Great. Okay, so yeah, thanks for your feedback. Oh, how much? I don't know what SPS is.
So someone is asking how much SPS do you need for analyte? Oh, samples per second. Okay, the data scientist in me says as much as possible. So go as high as possible.
So for the accelerometer data and the gyroscope data, we're going at 50 Hertz, I think. So 50 pieces of information per second. For the GPS location information,
that depends on how quick the people are walking and yeah, how quick they are. For the children, we're going probably at a higher sampling rate than for the older people for our residents at the nursing home. Which by the way,
doesn't mean that all nursing home residents are very slow. Some of them are very quick. So yeah. Okay, then I have one more comment here. Agree, close statement, close to say problem.
We need more funding for people power to do the open thing in this echo chamber. Are there any ideas for funders? And yeah, so if, do you know of any possibilities
anyone knows about funding possibilities for open hardware projects? Or yeah, probably open hardware for science projects. I know that there's Mozilla's open hardware program,
I think it's called. Any other ideas for funding?
M&T reform. And just one question again for Ormo, the public chat is not recorded anywhere, right? So I would have to just- That's right, correct. You have to write your notice. Oh, you can copy paste it if you like. Okay. I think you have the rights for this.
Okay, perfect, thanks. So, but you have to do it before, I have to close the session so that you have it in mind. But yeah. Yeah, okay. Oh yeah, there's someone taking notes already. Perfect. I think my time is up. So I really don't wanna keep people here. And I think there's also people already dropping out
for longer than needed. Yeah, thanks a ton for the feedback and or the information shared in the chat. That was very useful, definitely for us. Yeah, you can send me, drop me an email. You can find me on Twitter if you're interested
in our research there. And then you pretty quickly find all the other people that are in that say working group on open hardware. Open trackers in the Netherlands, that is. And yeah, yeah, thanks for FrostCon to invite me or to allow me to give a presentation Saturday also.
Yeah, thank you very, very much for your lovely talk, Daniela. And normally it's given applause, I'm sorry. But maybe the person that have to like to make your microphone on and can clap, but nobody have to, so yeah.
Okay, thank you. Thanks, bye. Bye.