Brandon Cannaday: Hello, everybody, and welcome to Losant's deeper dive webinar series. The deeper dive series is designed to provide an in-depth overview of some of Losant's functionality in order to give you a lot of information when it comes to building your own IoT applications on top of our platform. If you are new to Losant and looking for an overview or maybe a demo, feel free to reach out. We'd be more than happy to set that up for you. But I would recommend you stay on, because you're still going to learn a lot with this webinar. Speaking today is me. I'm Brandon Cannaday. I am the Chief Product Officer here at Losant. I'm going to be acting as emcee, or Taron Foxworth. He is our Education Lead and will presenting the majority of today's content. One of the major benefits when you join us live is that you do have access to some of Losant's foremost experts, and that includes Dylan Schuster today, who is our Lead UX Engineer. He has joined us to help answer your questions at the end of this webinar. Speaking of questions, I would recommend you use the Q&A feature in the Zoom webinar software. You'll find that button along the bottom of your screen somewhere. Feel free to put questions in there at any time during this presentation. I'll be watching those during Taron's content, and then we'll answer as many as possible at the end. This webinar is recorded. So, if you're unable to join us for the entire time or want to share this content with a colleague, we'll send that link to you after the webinar is over. You can also find this content on YouTube. Before we dig into it, I want to provide a brief overview of what Losant is. We are a cloud-based application enablement platform specifically designed for IoT. We provide that foundation and those building blocks on which our partners and customers build their own IoT applications. There's a lot of functionality that makes up Losant, but we divided it into these five main components, which we feel represent the complete technology stack that is required on the software side when developing an IoT application. That includes edge computing, your devices, and incoming data sources, the data visualization which includes dashboarding and batch analytics, our visual workflow engine, and finally, the end-user experiences. That's what's used when building and publishing your own completely custom and branded applications. Losant's a leader in the industrial and smart environment spaces, and we've offered this foundation for a number of customers ranging from start-ups all the way to Fortune 100s. So, if what you see today seems like it might work for your use case, please reach out, and we'd be happy to set up a much more in-depth conversation. While Losant provides the software foundation for IoT, there are many other components that have to come together. So, we surrounded ourselves with a wonderful ecosystem of partners. Includes technology partners that provide sensors, gateways, connectivity; better solutions partners that actually work directly with you to develop that application on top of our platform; and our strategic partners, where we have a very close relationship, and we share go-to-market and sales strategies. I would encourage everyone to explore our partner page and see if these partners do offer any products or services that you're looking for. Today's webinar is really going to cover application templates, and that is a feature that we have recently released that brings together all of the functionality that makes up Losant into reference implementations. We've done a lot of work to make sure we provide good documentation and educational material around each specific feature, but templates really provide you a good source of information on how to combine all of those features into something that's more real world. Templates provide that architectural guidance, implementation best practices. There are a lot of ways to develop similar implementation. So, our templates do guide you into some of those best practices. If you are building an application that looks a lot like one of our templates, feel free to use these as a starting point and extend it into the application with your own functionality. These are demo ready. So, many of these templates do come with built-in simulators. So, if you want to demo Losant in your organization or to your customers if you're a partner, you can start with these and show Losant's capability using one of these templates. Today's webinar is really going to talk about the huddle room application template. We developed this template because a lot of our customers have adopted the huddle room strategy. Huddle rooms are essentially non-booked conference rooms. Allows your employees to go find a quick spot to have an ad hoc meeting, but with that strategy came some issues. Our customers found their employees spending an enormous amount of time walking around looking for an available huddle room, and that caused a loss in productivity. So, this template and our customers, what they have started to utilize is occupancy trackers. So, understanding the real-time availability and historical utilization of conference rooms so that employees can immediately look at an employee app or maybe a kiosk within their space and see which rooms are available instead of wasting that time walking around. This template really does show a broader smart environment use case around utilization understanding. So, as Taron is walking through this, keep your imagination open that although this specifically applies to conference room, what else in your space could be tracked? Where is that benefit for understanding utilization within your environment? With that, I'm going to hand it over to Taron Foxworth who'll be walking you through the huddle room monitor template. Taron. Well, no problem. It looks like Taron is having a little bit of technical difficulty, but while he's getting back in, I'm going to go through some other use cases that we've seen in the smart environment. So, one is certainly the smart conference room and smart office space, and that's where we really talked about when it comes to the huddle room monitoring. So, adding sensors or retrofitting inexpensive sensors into existing spaces really gives you an in-depth understanding of what's going on within your own environment. The other one is entire buildings. So, you have conference rooms and floors. You can extend that to an entire building and see where assets and people might be moving around a space in a much larger area. Then finally, campuses. Smart environment is made up of more than just conference rooms and spaces. There are assets outside the environment, as well, including parking and sometimes even bicycles. Think about campuses that offer employee bicycles. Do you get the right bicycle? Do you have the right sizes? Where are they? Can employees access those assets within the broader campus itself? So, while Taron is still joining—it looks like he's having some technical challenges—I'm going to go through the more architecture of a smart environment. So, I mentioned a little bit about retrofits into existing spaces. This is a really common implementation approach that our customers take. It can be very costly to build out an entire new space. So, a lot of people just want to retrofit the space that they have. That's that top area, the top section on the left that you see. Something like Aeotec multisensor is a very expensive battery power option, but there are a lot of Z-Wave or Zigbee powered sensors that you can find off the shelf and install in your office space. Each one of those do talk to a local gateway. Z-Wave, Zigbee, Bluetooth, they're unable to talk directly to the Internet. So, they're going to talk to some kind of local gateway. Again, these are all available off the shelf, and you can find these and install these into your own space fairly easily. Those gateways then talk to the Losant platform. So, those gateways provide that bridge between the local environment and the cloud infrastructure. Usually, we see that over MQTT or HTTP. Alternatively, there are a number of partners in our ecosystem and vendors of existing smart building technology that have the sensors and motion sensing built directly in. One example is Cree Lighting with their SmartCast technology. Cree Lighting does have motion sensors built almost on every light fixture through the SmartCast gateway. The Losant edge compute agent is able to access that information and bring that data into the Losant cloud. The Cree option is going to be a little bit more popular if you are in the middle of rebuilding an entire floor or space, maybe a renovation. So, think about existing vendors that might have technology available for you if you are doing a renovation. Looks like Taron might have been able to join us. So, Taron, are you there now?
Taron Foxworth: Yep. Are you able to hear me?
Brandon: I can. I will pass it off to you.
Brandon: All right. Thank you very much, Taron. So, we've got a number of questions already coming in. As we do a little bit of wrap-up we're going to use the rest of the hour for those questions. So, if anything pops into your head, please make sure to put those in the Q&A section within the webinar software so we can get to those in just a moment. If you do start an application from a template check out those read-me's. They provide a lot of useful information about how that specific template works. We do have two other webinars already recorded. Replays are available for the industrial equipment monitor and the asset tracker templates. They cover the in-depth details just like this one did. You can find those replays on the deeper dive page on Losant or on YouTube. So, from here the next step, if you want to explore this template and maybe use it as part of your own smart environment application, the first thing you're going to have to do is start defining the model that represents your device and find some hardware. So, these templates use simulated data, but in the real world they're going to be powered by some kind of hardware or sensors, and really, that's up to you and your specific use case. If you do need some help or guidance on hardware that seems to work, we've got a ton of experience with hardware that both works really well and hardware that has not proved to work very well. So, feel free to reach out. We can get hooked up to one of our solution engineers or solutions architects, and we can walk through some hardware choices that might fit your use case. Then once you've got some real data streaming in that matches the model of that device's attributes, you can start applying that to your version of the application. So, you can start from this template or use this template as inspiration or guidance for your own application. We've got a couple other webinars coming up. So, make sure to save the date for those. Taron did a brief dive into experiences in this webinar, but Dylan is going to lead a much more thorough overview of Losant experiences. Experiences are one of the more challenging parts of Losant for people that might be new to the platform. So, if you are using that functionality, this is a webinar I would really recommend. Taron is coming back in June to do a webinar to cover a lot of the learning and educational resources that Losant offers. There's a lot of ways within our ecosystem to learn how to use Losant. So, Taron is going to take a deep dive into many of those and help guide you when you're exploring building your own application. You can register for all of these at losant.com/deeperdive. If you are ready to get your hands dirty, we do offer a ton of learning material. So, our documentation is very thorough. We make sure no feature is released without being documented. We do have our forums. So, if you want to ask general how-to questions on building something in Losant, we would really recommend you go to the forums for that. We do have Losant University. So, University is an in-depth video-based training tool that guides you through almost all of the major components of Losant. It's a great way to get up to speed quickly, and we do also offer a certificate of completion. So, if you do make it all the way through Losant University and do pass the exam at the end, you can print yourself a certificate and, hopefully, hang it on your wall in your office. Then our blog has a large number of tutorials. So, if you are trying to solve a specific problem using some specific technology, go check out the blog. There might be a tutorial on there that is pretty close to what you're looking for. So, with that, we're going to open it up to some questions. So, first off, let me bring open some questions in chat and the Q&A. Let's start with some simple ones. So, Dylan, if you're still on, can you do a couple of questions about limits that came in? Is there a limit to the number of experience users? Then follow that up with: is there a limit to the number of levels that a system can have?
Dylan Schuster: Sure. Just start with the experience users. Yes, there is a limit. That limit is pretty low when it comes to a sandbox account. I think it's only two or three users, but that, of course, goes up when it comes to an organization. If you do find yourself building an experience or having an application experience out in the wild that is bumping up against that limit, you can, of course, reach out to your contact here at Losant, and we can talk about increasing that number to whatever it is that you need. Second question: is there a limit to the number of levels a system can have? There is a sanity limit, I will say. I don't know what it is off the top of my head. I want to say it's 100 levels deep. For all practical purposes, there's no real limit; it's just a matter of keeping the API from getting itself into an infinite loop with that number, that limit, on how deep the system can go.
Brandon: Awesome. Thank you. Taron, here's a perfect one for you. The template tracks essentially utilization, or maybe motion detects a presence of a person. Can you step through what it would take if you wanted to add temperature and humidity to those devices, which seem to be pretty common in a lot of the sensors that we find?
Taron: One hundred percent. You mind if I share my screen? So, as Brandon mentioned, and as we talked about in this diagram here, there are a ton of ways to get this data to Losant. So, the first thing that you want to do is figure a way to get the data. So, in this case, the Aeotec multisensor, that sensor actually tracks motion, humidity, and temperature. Has a bunch of data with it that you could actually use and leverage immediately, but the Cree Light is just a light and a motion sensor. So, step number one is going back to what Brandon mentioned, find hardware for a use case. So, you got to have the ability to track temperature. But once you do, and you're able to send that to Losant over MQTT or HTTP, integrating that to the template is pretty easy. We can create new devices to [Inaudible 00:53:06] temperature or we can extend a hardware device like the one you see here and add a new attribute. In this case, it would be temp or tempC to represent Fahrenheit. We can track those directly on our devices, as you see here. That would be how I would integrate temperature into the application.
Brandon: Awesome. So, here, pass that screen back to me, please.
Taron: Of course.
Brandon: All right. Then there is a great question on some deeper analytics. So, I'm going to answer this briefly, and then I'm going to ask Taron, can you bring up our Notebook example for the smart environment, and then I'll toss it to you to step through that. So, while you're bringing that up, I'll read the question and do a high-level answer. So, the question is: how can we pull the time series data into an ML training application for occupancy forecast and prediction? The use case is specifically for commercial buildings such as shopping malls. That is a great question. So, this really piggybacks on our batch analytics capability. So, I'm going to talk about that briefly while Taron is bringing up an example. So, within Losant there is a feature called Notebooks. Notebooks is 100% designed for that large scale batch processing of historical time series data. So, what you can do in there is if you've got all this occupancy data, we'll take care of pulling all that information out, placing it close to a Jupyter Notebook. We piggyback on Jupyter, a very common data science tool. Our Jupyter execution environment does have Google TensorFlow. So, you can use that to actually train an ML model. Combined with our workflow you can make that scheduled. So, maybe you want to retrain it weekly, or monthly, or what duration it would take an interval that makes sense based on a meaningful amount of new information. Then that model can be sent up to Google's cloud execution side, the ML engine, which our workflow does directly integrate with. So, we can do the real-time training and the real-time execution of that model entirely within Losant. So, Taron, do you have that example up? Maybe show a little bit of the Jupyter Notebooks?
Taron: Yep, sure can.
Brandon: Let me pass that back to you.
Taron: It's like hot potato. I love it. In this case, what Brandon was talking about here is the Notebook. This is the Losant Smart Office application. The application looks very similar to the one we just demoed in the template, but this one has a Notebook. As Brandon mentioned, Notebooks are for batch processing. I like to start my Losant learning with questions. So, the question to that Notebook answer is: if we want to look at data over a large amount of time and derive some sort of insight—so, the batch processing—the workflow engine is stream processing. So, as they're coming in, let's do something with it. Notebooks, on the other hand, are let's look back at data. That's what Brandon was talking about. So, when you're configuring Notebooks, you can configure your input. So, in this case, my input is simply my devices data. Brandon mentioned Notebooks are all backed by Jupyter. So, the entity that's performing the batch processing is a Jupyter Notebook. Every time a Notebook runs, we get this super clean execution log that tells us when and what happened and all of the outputs of the Notebook. So, if Notebooks are for batch processing, that batch processing has some sort of result. We can get those results back, integrate those into other dashboards, trigger other things. The nice thing is that we also see the Jupyter Notebook and what it looks like after it's done running. So, in this case, here's a Jupyter Notebook for this application. It's pretty simple. All it's doing is calculating how much motion is on each day. I think this one is answering the question what is the most popular day in the office, which is a really tough question to answer in a workflow engine but very easy once you have the tool that you're familiar within the data science world. Then Brandon mentioned all this is directly integrated into our workflow engine. So, in this case, we can trigger workflows when our Notebook is complete and use that result somewhere else, or we can tell from a workflow trigger a Notebook itself. You wouldn't actually connect these in practice but... That is how Notebooks work in Losant.
Brandon: All right, awesome. So, just a few seconds left. I think we've got time for one more question. It's poignant with the atmosphere we're in today. So, Dylan, maybe you can talk a little bit about how you might use an application like this to track some social distancing. So, can something like this be used to make sure no more than a certain number of people enter maybe a conference room, or could be a commercial space, or something like that. So, can you talk to some of the technology you might recommend when making sure no more than a certain number of people enter a space?
Dylan: It's a good question, a good topical question. It's going to depend on your hardware, of course. The occupancy sensors range dramatically as far as how granular data can get. So, at the bottom of the spectrum, it's just going to be a motion detector. If there's motion in the room, we can assume that at least one person is in there. We can't really tell much more than that as far as how many people would be in there. Of course, we also don't know the dimensions of the room. These sensors go all the way up to facial recognition and everywhere in between. Not so much this is Brandon, and Fox, and Dylan in the room, but there are three people in a room. Something like that we certainly could track. So, in the case of how you would set that up in Losant, assuming your hardware can at least tell us how many persons it is picking up, I imagine... Well, there's a few ways that you could do this. The device that is reporting this information would count how many people are in the room. You probably have established some sort of threshold given the size of the room. You could do anything from turn on a light on the wall if that's connected to the platform via a different device, a little sign that says, "Hey, too many people in here." You could send a text message to the building operator. You could log an event, which is a little less intrusive, and just come back and use it as a training mechanism later that we had too many people in the room from this time to that time. A whole lot you could do there.
Dylan: Oh, go ahead.
Brandon: Oh, if you had more, please continue.
Dylan: I was just thinking out loud. There's also possibilities where... This came up in a conversation I had with somebody, actually, earlier today, where they could use camera sensors to determine whether or not somebody is wearing a mask. That, of course, comes back to how high quality the hardware is. But if you have an image on hand, you could also use the workflow engine to send up to a third-party service that does image recognition such as Google's Vision API, and perhaps that can deduce whether or not people are wearing masks. If it can detect a face without a mask, you could create the same alert that I mentioned as far as number of people within a room.
Brandon: Got you. No, some good insights. Thank you. That reaches us to the end of the hour, a little bit over. Once again, I want to thank everyone for taking time out to join us today. As one more reminder, there are two other webinars coming up in May and June. So, add those to your calendar. If we weren't able to get to your question, we do try to answer those offline. So, look for those in the forums if we're able to, and then look for a link to this replay that you can share with any colleague that wasn't able to join live. With that, I appreciate again you joining us. Thank you.