G Suite Plus GCP: Building Serverless Applications with All of Google Cloud (Cloud Next '18)
My. Name is Leslie Cheung and this is not my coworker, ace. Hello. And wait. A minute what are you all doing here what. Are you doing here. I'm. Just gonna sorry, I'm gonna be a little upfront with you but we're not here to launch a product okay. We're not gonna tell you about some new technology, all. Right in fact we're gonna show you things that you can already do today, but, maybe you weren't aware of it or you, don't know how so. I'm excited that many of you here because I suspect you're. Curious about how to make G sweet and juicy please play well together all, right and that's one of the goals here in addition to talking about server lists so, your that are you totally obsessed with server list and you're going to every single server let's talk all right, there are people like that out there, okay. Cool, awesome. So a little background about me my job is to make you all successful. Using. G suite GCP and, all Google API s I am, the host of the G suite dev show on YouTube so if you do a lot of G suite developments, each of these videos is like five or ten minutes really quick to show you about one actionable. Learning thing and then you can take that with you all right I've, written some Python books and I'm also one of the original engineers, that built Yahoo Mail 20 years ago it's this, October will be the 20 year anniversary of, Yahoo Mail and. What, about yourself so. A little bit about me I'm a developer, programs engineer on the Google cloud team so, my mission here is to enable developers, to maximize, their productivity using server lists so, I'm. A full set of who kind of wandered into cloud I've had other lives doing dotnet, desktop, dev mobile, dev so, on and so forth and you can ask him about that afterwards, all. Right so, this. Session is basically what, I just told you guys earlier so, GC P&G sweet and serverless we're, gonna show you how to build solutions using all of Google Cloud and so, we're gonna give two demonstrations, one showing you how to go one way and the other showing you the vice-versa all. Right so how are we gonna do this so. This, talk is broken up into six major parts we're gonna go through some introduction, talk about what serverless is in general in case you're new to it then. We're going to show you. Describe. Each of the different products on the GCP side for service as well as on the GC side and then, we're going to go into the two demo apps where one we're going to analyze G suite data using G CP and then, the other app is how to access g, CP technologies, from the g suite side and, then, we'll give you some inspirational, ideas as to where you can take this you.
Know What's your next step and then we'll wrap it up after that, okay. So why, you, are here, so, overall. Just to kind of like set the stage Google Cloud is a combination. Of both g CP as well as g suite so that's pretty important, to know and even. Though they are their own products, we want to show that they, can work well together. Now. Who, are you and what is server list those are those are the big questions there's, lots, of different use out there so maybe, you're an engineer you work at a large company or maybe a smaller ones medium sized business a start-up or perhaps, a hobbyist or you work at some independent, software house or. You're in DevOps you're some kind of admin a DevOps person, or you're, a consultant analysts, or some kind of integrator, or any, other type of technical professional, or decision maker on, the. GCP site you could be a full-stack developer fully. Living in the cloud but maybe you have some apps or data that's in G suite and you want to be able to take advantage of that on. The other side if you're, an enterprise developer, and you're fully immersed, in the G suite world well, maybe you have some big data stored in DCP or you have big data period and you don't know how to take advantage of GCP so. We want to show you how to do that as well -. All. Right, so what, is serverless so. Wikipedia. Defines service. As this cloud computing, execution, mountain which a little bit okay, there's a lot of talk, right. So, I'm gonna try and take a chance at simplifying this or both of us are we have our own takes on what server this really is my, take is very simple, I'm. A Python, developer you probably already know that so I take a lot of code and I reduce, it down to very few lines of code so I have shrunk this down to serverless. Computing, is really just logic, hosting in the cloud because. What is logic, it is an app or it's, a function, and if. You're hosting a function, or an app in the cloud, that's, serverless, because, you're not spinning up resources and things like that you're not managing machines or VMs or anything like that it's just hosting, apps or hosting, functions and ace, but, what's, this with a molecule. So the core idea here is that service, isn't all or nothing so you can think of your app as a molecule that consists, of one or more ad atoms. In the same way rather that your app consists, of one or more services, so, like a like, a molecule, you can put some of your atoms so to speak on service. And some of your atoms on vm's or whatever hardware, you have in mind depending on your unique situation ah, I. Think I needed to take more science classes in college, thanks. Haythe all, right, so let's take a step back with a true story at the end of the last millennium vm's. Became, a theme they, exploded, the server market there was really no need for bare, metal anymore. All right but now what server Lazar we're, saying you don't need vm's anymore, what gives oh well. So what really happened, is that vm's were the revolution. And server this is just the evolution from, that point so, when you're at a higher level you can solve more challenging problems, all, right well that's enough history let's take a look at some server salute serverless solutions, starting, with GCP and i'll hand it off to ace, to tell you a little bit more about that, so.
The Core idea here is that service, compute on GCP can be thought of as logic. Hosting, so, basically. X hosting, in the cloud for whatever your X is now. If your X is a function you're going to use, something like Google cloud functions, if your, X is an entire app you're, gonna want to use App Engine and if, you're trying to host a container or cluster, of containers in the cloud you're going to want to use either gke, service, or Kay native. Oops. Sorry. Now. The thing to bear in mind here is that service is more than just compute, there's a widespread of, examples. And service tools that we have on GCP that, are not exclusive. To the compute, space. So. Take, cloud datastore for example you probably know, what this paradigm, of a no sequel database looks like maybe you've worked with something like MongoDB but. Let's take the MongoDB example. When you're setting up a MongoDB instance, you have to worry about say the underlying, security, of that instance, you have to worry about like. OS scaling, load balancing, things like that with that particular, MongoDB, instance with. Cloud, datastore you get a no sequel database. Api ., you don't have to worry about the underlying VMS. Or the underlying OS or anything like that it is completely, abstracted, away from you next. Slide, so. Without. Further ado I will hand it back to Wes who will talk about the, G suite aspect, of service great, so, so, for. The G suite side we have two major ones because. GCP. Lives at the platform, in the infrastructure, level J, suite lives at the SAS level so of course we're only dealing with applications, and applications. Hosted in the cloud so, we have two platforms for that the first one is Google Apps Script and then, the next one is Google app maker so you can check out the links for the products but let's dive into each a little bit more deeply so what is Apps Script it is a server-side. Suicide. A runtime, just like node however. It's customized, for G suite automation. Extension. And integration. So whether you're trying to add custom, features you know for example one of our samples is G, Google, slides doesn't, have progress bars when it'd be cool to find out how far along in a presentation, I am so. With one of our samples let's you actually build progress, bars so in that sense you're extending, one of our applications maybe. Your, company has a well known product but you want to give them access to that, product from within sheets or Docs or slides or Gmail or anything like that so, add-ons allows you to do that okay, so it's native. Up service, object access versus, using API, so what that means is that instead of you know having to worry about OAuth, thinking.
About How to use HTTP, to talk to an API they're, just objects, that are already there available for you in Apps Script so if you want to talk to a Google sheet you would do spreadsheet, app dot and then, the name of the method that you want to call okay so it's it's basically like that and of. Course it covers more than just G suite products you can use other some PCP products, so we're gonna have that in one of the demos and you can use other Google services, as well too so in Apps Script you can also reach Google Analytics, Google, Maps YouTube. And many. Others of course, we don't want you trapped, inside the Google world we want you to be able to go outside and talk to other services, that are online or on Prem so, we have the URL fetch, service you, can see here as, well as JDBC, if you have a database, that you want to reach as well so, all of your applications. Hosted and executed, in the cloud there for service and if. You've ever heard of add-ons for Docs sheets slides forms, in Gmail app script, powers add-ons, if. You want to find out more about Apps Script there are four talks, that you can take a look at just. Go back you know go back home and watch the videos I think most of them have have have already happened so that's why the. Next product is app maker this, one is for an. Audience, who may not code all the time so it's a low code assistive, development. Environment, it has built-in templates, so if you're ready you know you want to write a web app for you know purchase orders or something like that there's a whole set of templates that are already made for you you, basically build an app by dragging and dropping from a pallet onto a canvas and then, you kind of connect, to the data source and choose, the columns that matter to you and then that, allows, you to go from an app an, idea to an app in just minutes instead, of requiring, engineering, resources or taking more time under, the covers it generates, Apps Script code okay, okay and since Apps Script is server list then you automatically, get that the. Default storage mechanism, for AppMakr is Google Cloud sequel, so you are kind of straddling, both worlds between G suite and GC P and we, just launched it publicly, last month so it's very very new and there, are a couple talks that, are related to at. Maker that you see listed there you feel free to watch those videos as well now. If for your work you need to actually have hands on or for your engineers, that need to have hands on doing this stuff we do have one more bootcamp left on Friday afternoon, so just end. Of the day from 1:30. To about 6 o'clock where you get hands-on with all three you will learn how to use G suite REST API s you will learn how to use Apps Script and you will learn app maker and you will have working code before your before your done so, it's four and a half hours and we teach you all three and you get hands-on and doing multiple code labs so, that's, so.
That's The training session we still have a few seats available so, if you have time and, an. Extra registration, fee for the bootcamp then feel free to join I'll be leading that session along with my colleague will be doing the app maker session. Okay. Now. Is, it. Hard to get started playing. With server lists on GC P and G suite and making them play well together are there not enough documentation. Or code, samples for you to look at or do, you know G CP but not G sweet or vice versa and you want to see how it's done well, we have lots of links to share and then we're gonna have two simple apps to demo so you, won't look like like. You're confused, alright, so, first step is called custom, intelligence, in Gmail and it, is going, to allow you to analyze G, sweet data using. GCP so. The. Main premise behind this app is we. Get email we, get a lots of email and. Services. Like Gmail and others do its best to get rid of all this spam however. Even, with, the spam gone, there's, still a lot of email messages that come into your inbox and, not all of them are totally important, so it would be great if we could sort of like single out maybe the dark messages, here that we care about and we want those specifically. Annotated. So we know those are specifically, important, messages to us right, so, let's build something like that so the, way it's gonna work is on the GCP, side you. May have noticed that there's a lot of data, that you have in G Suites so we want to use GCP you know whether it's analyzing email messages, or other data you have in G suite. We, have UCP, has a lot of ml, and big data analytics that you can take advantage of so it's important to do that so. One of the tools that we can use is a Gmail API you. Can get an alert for every new message and the, way you do that is you can tell the API to set up a push notification to, cloud pub/sub and then. Cloud pub/sub can trigger an appropriate, cloud function, cloud. Function can take the, email message you, know scan, the images, that are in attached. To your email message and then send that to cloud vision and. Cloud vision can identify the images, that you care about and then, label them and such. That if a matching, label comes in you can talk to the D mail API again and star those messages which match that label okay, so that's how it's, going to work so. I'm gonna hand it off to ace and he's, gonna describe to you a little bit of about the code and how it works and then he'll do a demo for you so it's, the middle button so. Diving. Straight into this code just as a little bit of background the, way this code works is this is the function, that is triggered by the pop subtopic, whenever, there, recipient, receives a new email so. Starting, off basically. The first two lines of this code are all about extracting, the recipients, email address, from that from, that pub/sub notification, then. We fetch their oath taken, which is stored in datastore and once, we fetch that token we can carry out operations, on the Gmail API not, necessarily, as them, but on their behalf so then. We have to list the message I the most recent message IDs which we do with this less message, IDs helper, function then. We take that we get the most recent message ID and we. Get that message as sort of object representation. Now, when we get that object representation. That includes, the actual contents, of the image so, we can go through those contents, and extract any mentioned, images from it, finally. We take all those images and we say we're, going to pass this off to cloud vision using the get labels for images function and then, that passes. It off to cloud vision cloud vision, runs. It's machine learning models and categorizes. Those images and. Sends. Those categories, back to cloud functions, now, the.
Absolute, Final, step here is, if. Cloud. Functions, gets, this set of labels and that, set of labels contains the word bird then. We mark, the messages start if that, set of labels does not contain the word bird we simply exit the function and don't, star the message. So. This, is the function that talks, to cloud vision to get those image labels at a high level basically, what it's doing is it's. Sending a, request. To cloud vision using the cloud vision client, library and in, nodejs, and then. Once, that process is done it takes that response, kind, of massages. It into, an easy-to-read format and particular just a list of categories for all the messages and or all the images in the message and then. Returns, that back to, its original calling, function. And. Then. We can star interesting, messages, for some definition of interesting, in our case containing, an image. Of a bird sorry. Using. This label message helper function which simply wraps the. Users. Messages. That modify, G, API function, and we. Can simply pass a, label. A list of labels that we want to apply to the message which conveniently, can include the reserved. Label starred marking, the images or the message sorry s starred. So. Walking. Through this this process in terms of what happens, when a message. Actually arrives, in the recipients, inbox. So. Once. That message arrives in the recipients, inbox a message. Notification, is kicked off to a pre-configured, plot cloud pub/sub topic, now. That, cloud pub/sub topic, is then used to trigger a google cloud function, and that. Cloud function, can then retrieve the comments of the email from, g suite from gmail, so. If there, are any images, in that email they are extracted, and sent off the cloud vision which, then, categorizes. Those images and returns the results back, to that invoking, cloud function. Now. If the, proper category is, present, in that. List, of image, or categories. Of images then. The image is marked as starred using, the Gmail API and. To. Demo this. What. We'll do is walk. Through it here. Now. You want me to drive it. Sorry. About that that's right. So. Here we have some sample messages. And if we click through them we'll notice that the ones with photos of birds, are starred. So. That's a bird it starred. That. Is my cat Ethel he's, not a bird she's. Not starred she's, a very lovely cat but she's a cat not, a bird, this. Is a bird and it's been started by a claude function, and. This. Is the fish, it's. Not a bird so, it hasn't been started i. Just. Like that demo. We. An in Chester's email inbox by hand I just made it can you tell there's a welcome to Gmail I haven't. Actually used this as a user so and. Of. Course you know you you can use birds you can use whatever, sort of filtering. Mechanism, you want you might say okay maybe, if this contains a mention of I don't, know someone in my family I want to automatically, start or, something like that. I. Know how it works so. To, summarize, this app at a high level, basically. The first step is to use the Gmail API to set up that notification forwarding. To a cloud pub/sub topic, the. Second step is to have pub/sub when a message, comes in trigger, that logic that is hosted, by cloud functions, third. We use, cloud functions, to provide us with server lasat scalable access to the various GCP sort of machine learning api's, and whatnot and we. Can combine these three things to provide us effectively, custom, intelligence, in Gmail and when I say custom I mean it's code you can write whatever logic, you want now. The applicant, the source, code for this app is hosted, on github at github, comm, slash, Google cloud platform slash. Cloud, -, functions, - Gmail - no js'. And. Without. Further ado I'm gonna hand it off to West who will talk about the other direction so going, going. To GCP tools from G sweet cool, thanks, ace that was an awesome app. So. Next, one is, called. Big data analysis, - slide presentation. So. What does that mean so, what that means is that big day as a thing now data. Scientists, is a is a title, now they, offer data science degrees in college now so it's a thing there's, a lot of tools out there for data scientists, to help analyze datasets, machine, learning to give you insights, and analysis, and so on and GCP, is always releasing, awesome new features. So. For example suppose you are doing some deep analysis, on some big giant data corpus and you, choose to use a GC, piece bigquery, it's a great tool to do the job now.
It, Has a UI so you can craft what you know use whatever data set you want craft the the. Query that you want but. You, know you're happy with your query but now you want to put it in code so what does that look like so. This is an Apps Script app alright, so this is a G sweet serverless app that does that so, in order to talk to bigquery, you do need your cloud project ID that you get from the API developer console just. Enter it up there but once you have that then here's an example query we're just going to look at you, know sort of like the the most common words in all of Shakespeare small, data set not too too large but, gives. Results pretty quickly so that's the query that you would send it's very sequel like and what. You do is you the, the request to bigquery then. You get its job ID then you wait for the job to finish and then once it's done you get the results and then you have some rows. So. Yeah. We got rows now. What, like, what do you do with this data sitting, in a variable I haven't put it anywhere and if, I turn off my computer goes, away then I have to run it again right so you could I'd put it somewhere so. What's. A good place for data. So. Anyway the data and, the analysis, is only part of the story so my. Contention. Is that this is only part of the job it's like these stairs well. It it, works you can use them but it's really better, if the job was complete, right and what, I mean by complete, is at the end of the day with, all the hard work that you've done with your big data analysis, and stuff don't you have to justify. Your, efforts by showing your results to your managers, right. I mean you got to get funded right and if you're in if, you're at a university this, is even more true because need to get really get funding right so, why, don't we kind of like go the last mile don't, quit just, because you've got results and you can stand up take, it one step further and, you. Know make, sure that, you can take, advantage of gee suite and most. Importantly, its api's so that you can visualize the, results better and make your results more presentable, to management, okay. So. If. You've got data results, how are you supposed to give it to management ok there's a little button there that says download as CSV do, you really want to do that do managers know how to read CSV no what do they know how to read Jason. Maybe not, really maybe. Save it as a bigquery, table no not really but, guess what managers, know how to do spreadsheets, okay. A spreadsheet. Is just a UI on top of a database isn't it right. Okay, okay we can have a debate about this later okay. But, yay now we actually have found a home for our data to go to that our managers can understand, so how do we do that so, if we save the rows now. Let's go make a spreadsheet so, the name of our spreadsheets, gonna be your query name okay. And, then I'm gonna create a spreadsheet with that name, get the current active spreadsheet that I have and. Then. I'm going to convert, all the names of the header to. Uppercase just so I could see it better compared to the rest of the data and then I'm gonna add that data to the rows and. Then what I'm gonna do is I'm gonna grab all the data from the rows and then I'm going to stick it into the spreadsheet so we're gonna skip the header row so a starting row on row two because, the headers are on wrote one and then starting column, one, okay. Because, we want to go from bigquery which. Shows you the word in the count. To. A. Spreadsheet. Has the word in the count so that's why I capitalized, it so you could see it and distinguish, all those words that are different from Shakespeare's, actual words and then how many times they show up okay. So that's that part so. Yay we have this that's. Great now what while. Having this data in a spreadsheet is more, visual, than having. The data sitting, in some database somewhere we, can do better, okay. Like, a chart it'd, be cool if we could take that data and build, a chart in the spreadsheet that's even more visual. For your managers, okay, and because remember they're not technical. Okay. Alright. So how do we do that so, here's how you do it so. We have a create column chart function. Which will get, the. Sheet I'm going to start. Looking. At the data I want to skip the header row because that has just column, names it doesn't have any data so I want to start with the data so, start with cell, a2 going, up to cell b11 so, if you go back mentally, you'll see that and. Then we're going to start putting the chart at row row, five column, five, and. So all we do here is we just request a new chart it get create we want to columnar, chart we. The range is gonna be start cell from a2 down, to be 11 and then I'm gonna set the position of the chart build a chart insert the chart and then I'm done, there. We are yay. That's, the next step, but.
Now. What we can do even better than, this yes I know you're really, really, hand-holding. Your managers, here but, we can do better than giving them a spreadsheet okay. We. Want to make it presentable so that's where Google slides comes in so, we want to build a slide deck so we have a title in a subtitle, okay so let's show, how to do that. So. Here is create slide presentation, same query name as before so this is all part of the same app by the way and. Then I'm, going to get the page elements, which are the title and subtitle then, I'm gonna set the title and then I set a subtitle, because, subtitle. Is here the query name was a variable because we also had to create the spreadsheet so that's why that was a variable otherwise I would post a string here okay, so that's how you create a slide deck now we need to ingest the data from, the spreadsheet so. That it is in a table on a slide, so. How do we do that, alright, so this time we do want the headers because we want to bring everything over from the spreadsheet so we start at cell a1, and go down to B 11 again and so, the first step is to insert a blank slide then. I get the range from start cell the end cell which is that entire range and then I'm gonna insert the table and then I have a double for loop to populate. Every single cell of that table so I have to say. I want to get the cell I want to get the text box that's in the cell that I want to set this text with the string value okay. So that's how you put the data into, the table now. Let's, link to the chart in sheets. Only. Two lines, because. The object already exists in the spreadsheet all you need is a chart ID so, when you created the chart in using. Google sheets you, save the return value which is the chart, ID and so all you have to do is do a pen another blank slide and then insert the chart into it and then it's done okay. And so we're, able to go from this. To. That so. You have a slide deck that has three. Slides, that. We just went through okay. Now that you can go and present to your managers right away right. You ran the big data Corre all the way to having a slide deck and you didn't lift a finger in between okay. So this is what I mean by going the final mile yes, you got your results but now let's put it in a place where we can actually show our managers. Okay. So how did this app work so this is an Apps Script app so, what happens is you, can start your app script back from a desktop or from mobile you. Send the request to bigquery to opportu execute, that query on your behalf it. Returns the big data analytics you, store that into a spreadsheet then. You build a chart then. You call Google slides and. Basically. Build a slideshow that has all this data in it and then, now you have something that you can present to managers so. That's the flow of that app now. Let's see it run, there. Maybe a little bit more action to this one but. Maybe. Not. Okay. So here's the script, and. It's great big query, presentation. So I'm gonna do the most common, words in all of Shakespeare which works again and then. If I do this of course I don't have a Google Drive folder I should have created a drive folder I should, have the OAuth permissions so. In this code when you look at it there is no Oh auth code so that's again one of the benefits of using Apps Script is that although well, a woth code is written for you but, the user still, gets the prompt to authorize, this script regardless, which, is kind of cool, so. I'm gonna review the permissions, I have not run this demo yet from. This account so let's let's. Do a live demo in front of everyone okay, since, I didn't publish this app it's not verified, so the question is do I trust, myself. Depends. What time of the day I'm coding. So. I trust myself and, it, says I want a view and manager data in bigquery. View and manager spreadsheets and drive this is looking promising by the way and then, view and manage your Google slides presentation, so I'm gonna allow this code to run, cross. Fingers it's, running running running. Project. Miss this user does not have access all, right I was, taking a risk all. Right, yes. Not yet not yet not yet not yet. I'm. Just gonna go to drive oh. I. Hope, you shared it with my self. Shared. With me please. Be in here, of. Course I didn't. Okay. My, fault so, anyway it works trust me okay. If. You don't believe him you can go look at the code yeah yeah yeah but. What I was gonna do is I'm gonna show you that deck and now I was going to show you two other queries, that I did that is not the. Most popular words in Shakespeare, so, I have one which I already pre-loaded which is the most popular long works in Shakespeare, okay. So these, are these words. Okay. And here's that and then. The other one I did was to, break away from Shakespeare, what, are the effects of, drinking. Alcohol. With. Newborns, and so this was a corpus, from. The neonate, allottee database from. 1952. To 1965. Roughly, and.
You. Know basically the, the mothers had to say were they drinking alcohol, during the time of their pregnancy, and so, there's a yes. No and, also no answer and you. Can see that there's a little bit of a difference for people who did have alcohol use so. You can run any kind of query that you want on it and. Be query has awesome public, data this is all public data that's available in bigquery already. So. Anyway, so I, don't. I'll, have to check on the project but anyway. Trust, me that it works. Okay. So. The, summary is you, can leverage DCP, from the G suite site for example in, this case we're taking bigquery, and doing big data analysis, on it, but. I'm trying. To encourage you to do is to, go that final mile you did all this hard work but let's just take it one step further and put, it into a spreadsheet not only has data stored but, you can actually use it to generate a chart or something visual, so obviously the I only showed you one type of chart which is the columnar chart but, spreadsheets, can do a lot more different types of charts than that okay, so it's very important, to know that you, can have different types of visualization. Then those of course Google slides that you can use to generate all those slide decks that you can just straight take straight to management and it's all glued together by G suite server lists so. You can build this app on your own using. The code lab that we have so it's available at the shortlink it's, already been live for a while and if, you do come to the code lab and you find the the intro code labs a little too easy you can go ahead and do that one and some, of us will walk you through it as well too and, then of course the the source code is also available on, github as of. Two nights ago yay, so. You can go check out the source code as well and then try it for yourself. Okay. So what, is next so uh so what, we just did was give. You an idea, showing. You two, sample, apps that are fairly different right. But, we're using a combination of four different GCP api's, and four different G Suite api's you, know with it between the two hours put together so, what else can you do so I just want to make you guys think or make you all think at you know at your problems, at work or challenging, issues you have in the office for or for your customers, or even for yourselves what, kind of things that you could do knowing, that you can code both, at the same time so. The, next step for you is using, imagination so we're gonna give you three ideas and by, the way if you actually end up building these apps please share them with me I'm curious. As to what people build with our stuff so, I'd like to see these apps build one. Of them is I like. To ask everybody how many of you out there have a big giant tear of a terabyte, hard. Drive at home where you dump all your pictures. From, your camera's SD. Card and it's, just sitting there right. Yes I know some of your laughing okay, it's true but if you write code you can actually make some magic happen because. You can leverage. G sweet and GCP again to do some work for you so for example the. Google Drive API if, you put all your pictures in Drive the drive API out gives you access to the, image. Media. Metadata, and same, thing goes for videos, but there's a lot more metadata for photos because you have the date and time you have the geolocation you. Can actually use just that data alone to build photo albums, so you never have just pictures, there with no you know no no reason right, so, that's that if for some reason you don't have any image metadata they're just odd folders and, stuff like that you can upload those to Google Cloud storage and then, use the vision API to, give you some information on those or if you have videos you can use the video intelligence, API that can scan, your videos frame by frame to, tell you you know whether you want to know sentiment, whether you want to do a OCR, or whatnot so you can use those, interfaces, and things like that yeah exactly so you can get a lot of information and in fact over all you can not only extract metadata but you can even make videos and photos, searchable, in addition, to creating photo albums okay, so, we really want you to go from this.
To. That you. Can auto generate, collages, photo albums, without. Even touching, the, files the. Photo files or the movie files just by running a little bit of code okay. So that's one example now. Another issue, that we have is can't we be a little bit more efficient with our time you can actually create a tool to make this happen. You. Can skint, you can use the Google Calendar API you, can scan all your meetings all your, appointments, and you can make possible optimizations. Based on using our cloud ml api's, such, as perhaps grouping, meetings to specific, days or searching. For occasions. Where maybe you're not making good use of your time all, right and then, the other thing you can do is you can auto schedule meetings with contacts, and with teams like if you have team meetings every week or customer. Meetings every quarter, or monthly. You know it's up to you to decide but you can use a gmail API for that the calendar API for that you, can find the optimal day in time for all parties, right so all these things are possible. Another. Example is, maybe maybe you have four customers, email you just as you were getting on the plane to come to cloud next and you really want to respond to them but but you don't have time had a chance to reply yet well, guess what you can automate that ace. Already, showed you how to use, pub/sub. To get notifications, of new email messages maybe, you can you know add some code, or add some logic to tell who are your important customers are and if it recognizes that maybe you can auto send a reply going hey, I'm off to a conference, but, we've. I've scheduled, a meeting with us together to chat next week and you, didn't even have to schedule the meeting yourself your code did for you okay. So something like that and, then. The third example is what, do your customers, feel. About your product, all right it's a really important, thing because that actually determines whether you should stay in business or not right okay so you can actually perform sentiment. Analysis, if, you look at all the different feedback, all the different feedback channels, you, can analyze them so you know look at all your support emails look at all the Twitter posts, and direct messages that you get check, out all what, the press and the media are saying about your product or your company and check, blog posts, where you know people are writing about your product you can use the natural language API from, GCP to determine the user sentiment, you can use a club ml engine and tensorflow to build and train your own predictive, models if you know you want more than just language.
Processing, And then, you can use G suite ApS to track all this stuff in a spreadsheet or in a Google Doc and then you can set up appropriate, meetings with Google Calendar again just like in the previous example. Okay. So that's another, example of what I want to see you guys do so, you get the hang of it now so think about some problem that you want to solve and leverage. GCP and G suite to make it happen. All. Right so, to, wrap up. Serverless. Is logic, hosting in the cloud at Google scale it's, mostly my definition, all right not the Wikipedia one, means. That you don't have to provision or manage resources. There's. More time for you to work on that idea you, know that the inspiration. That I gave you to, think about what other challenging, problems that you want to solve for your company, as well as for your customers, and, you. Know using. G suite and g CP by themselves, is great but, I think the combination of, the two is more potent, and I think you have more a more, ability to build the apps that you really want to build by, combining, both of them together I. Again. To kind of like wrap up why am I giving this talk why serverless, ygc, P&G, sweet why all these app ideas, and demos well, ultimately we want you to be successful and if you're a customer, we want you to be a happy customer all, right so thanks, for coming. All. Right before we get off the stage more next steps so what can you do now hands, on code. Labs Apps, Script code lab app maker code lab and there's a whole bunch of gzp code labs so this, is a definitely, a photo opportunity, and. Like. I said we're gonna go, through the, Apps Script code lab as well as the app maker one and one for the, G suite ap is on Friday at the bootcamp so please join us on Friday afternoon. There, are also tutorials. For cloud functions, Apps Script and app maker that you can follow and then. I also, do videos easy you'll probably know and then there's also a lot of DCP, developer, videos as well so, those are some references, for you and. If. You want to look up related sessions that next when you go back home and check out the videos so, besides the bootcamp on Friday there's, also a talk, given by the Apps Script team that, was CP 131, so they announced sort of like a roadmap and new features.
We. Gave a talk on the first day also, about, serverless telling. You how to write a serverless app from top to bottom there. Was another cloud functions overview that's earlier today at 11:00 that you can watch there, will be a talk tomorrow morning about app maker starting, at 9:00 so this is the developer, talk we there was another talk for app maker earlier, in. The conference that was mostly for managers. And decision makers, but this one is more for developers, and we'll show you how you act you can actually build line of business workflows. In, app maker, there's. Also another talk earlier yesterday, on, certain. Integrations. Between bigquery. Specifically, and Google's, Google, Spreadsheets, Google sheets so, you can watch that talk and then there is, a spotlight, that happened yesterday, as well on server list running on Google cloud so you should check that one now as well and please. Give us feedback on on. This session in, the app or on, the website which shows up in the. Slides. That will rotate after, after we get off the stage all, right thank, you all for coming.