Write Looker-on Currency with airSlate SignNow

Save initial looker-on

[Music] good afternoon everybody wherever you are in the world and whenever you're watching this my name is Imran Hussain and I'm very pleased to welcome all of you to the first online session of the Google Cloud data user group a meet-up that's all about sharing learning and being inspired by uses of a data on Google cloud so firstly I hope everyone is okay and well safe and well in this covered like nineteen lockdown I understand I realized that all of us are in different stages of this and it's been a very challenging time for everybody but I hope that it's not an opportunity for all of us in our own little way and her own capacity to try and change things for the better and I thought as a community that has a very specific professional interest starting up a user group like this would be a way of reaching out and coming together and going forward on some some common vision and values so firstly housekeeping this I'll be wrapping this webinar up within the hour I want to make this session as interactive and as engaging as possible really get that buzz of a physical meetup going so you can ask questions at any time using a online software could slide oh excuse me slider comm so it provides a way of asking questions interactively and will also be running some fun interactive polls GCB data is the event code I suggest everyone gets in there locks in and has a window rocked up ready to go this session has also been recorded and it's gonna be posted on this channel and on that note I'd really recommend and in fact ask that everyone subscribes so you can follow the videos that we posted here we've got some amazing tools car happening over the next weeks and months these include the use of data vault on bigquery getting savvy with CI CD on on a composer data pipeline using PI test and running a big table in production some of the issues that come around arise from that so a real mix and the others coming as well don't miss any of them subscribe I hope I'm putting the right direction but anyway subscribe on that button thank you we have a great agenda today I'll be starting with a brief intro to the user group and how you can get involved well then kick off the main topic I'm looking at look at no pun from a bi perspective specifically three ways in which I think look at it really stands out and taken together makes it a really unique and game-changing bi platform we're very fortunate to be shared to be joined by Sebastian Fabri from Google a looker expert who will take us through looker from an architectural perspective and also show a really interesting retail demo and we'll wrap up with the Q&A and to reiterate the point slide Oh enter any questions as they come and we'll try we'll address them during the session we'll certainly try and get through them at the end and finally we'd be closing out with details of a great master class of looker GCP so a packed session to get through we'll start with a poll to really get a feel of what interests you and that's going to be very useful for us as a community there's a what are the subjects across the data disciplines that are most interesting to you and business intelligence and data in generic pop out as the pretty much in Nick and Nick leet trading off for going further dough and machine learning data architecture and see ICD and so forth so really interesting mix I suspect B is probably splitting the warehouse vote data warehousing votes will go to the next question so thank you for those who enter George you you're entered entered into that which is what are the top three GCP services that most interest you there's quite a few but I suspect that most of us will have an intuitive draw to the services that we use the most so I'll give everyone a few moments to to put in their items and interesting so there's a it's ml's in the lead MLA our platform bigquery following closely composers taking the lead okay some data engineers in the house neck-and-neck bigquery composed the ML engine followed by data flow and then datastore cubed flow and some of the other services but bigquery takes the lead no surprise there and I leave the poll open just afraid one today finish off so great thank you for that again it really gives a good idea of the distribution of people's interests and usages on them on GCP so moving onwards a quick intro I've been working with data for twenty years I studied on s ap spree much where I cut my teeth the data warehousing BI and since then I've got on to use and build solutions on most of the or all of the major eco state ecosystems in some way and one of the things that's drawn me and I suspect many of us to the Google cloud is the way it places data center stage the services are constantly forward-looking and they make sense that said they are still a lot of services out there and they're growing and evolving all the time and that can be confusing because there's so many ways of using the platform it can be a particularly challenge it can also be a particular challenge given the pace of Google's innovation so I would say the first objective of this user group is to build a practical understanding out of how to make the best use of data services on Google cloud and in that sense I think we all have interesting perspectives and knowledge and experiences that we can share and we get some wisdom of the crowd going to see more of the connections and possibilities of using this tremendous platform to move forward the second dimension is of the user group is us as state of practitioners were often working in silos and one of the peculiarities of data is that is a deep into dependency that exists between our different respective disciplines so the second objective of these group is to try and build up a unified view of how these disciplines work around Google cloud services so the more we understand of each other's context the better we become in our jobs how will the group work the primary focus is data so that's going to be the lens through which we explore and present topics and usages on this meetup group around the Google cloud we'll be doing that through the usual community activities sharing knowledge and examples through online presentations facilitating micro and macro and master classes and micro classes and doing hackathons on big query turns flow I know other user groups doing that as well but if there's anything that specific particularly data specific we can we can rock something up and finally we try and stay practical often theory and reality can be two different things and by focusing practical usages we can avoid gotchas and be inspired by extending our designs in several and in different ways I suspect this so this meetup group is going to be online I'm going to remain online for the foreseeable future given the current situation and I'll get an again I'll give details at the end of how you can can take part and present present on present on the group excuse me moving on then to the main piece of the evening which is Luca so Luke has a pre new BI platform it was formed in 2012 which is like yesterday in enterprise software terms and it's gone from its inception through to being part of the world's largest data platform in almost Nix in no time I thought it's a great it was great but I thought it's a really interesting acquisition by Google because enterprise bi in my opinion anyway was always the missing link in Google's offering so it's it's a really exciting and a really exciting capability that's now common board so before diving into the look of subject I like to ask one more question on the poll and that's really good understanding of where all of us are in our unarmed lucre journeys so going back to slide oh I'll open the next poll which is where are you on your Luca journey pretty straightforward options I hope we'll be able to answer pretty very clearly as to which which cat will you fit into and again I'll give a few moments for the submissions to come in and we have a interesting to play there so the buyer the buyer seems to be towards interest being interested and want to learn for personal development and I think that makes sense I suspect many of us are coming at this from a Google cloud perspective and sending practitioners of Google cloud unless you've been using looking for a marketing perspective it will be a relatively new tool so great that's that's interesting to know I'm so interested personal development already using it and what to learn for my current work so to begin then I'm going to show a very brief snippet that I've put together from last year's Google cloud conference oh excuse me look at join conference in cloud next this this year and here Lloyd taught the founder and Frank being outlined their reasons for joining joining Google [Music] you know this particular relationship has been going on for four or five years I always say like for these to be successful you have to date before you get married and and you know there was a lot of you know just real value for customers created in the field between the two companies so that really said okay this this is something that could work and then I kind of went through some of the other synergies but I think those are very real you know it's it's it's interesting products that we can build together you know it's this culture that's shared in a lot of ways I've always been impressed with Google because um they've build great software that serves that serves users really well their their focus is organize the world's data and make it useful right and and ours is empower people through the smarter use of data all right so the the missions that are really aligned you know they're an intellectual culture they're a nice culture they're they they've I'm super excited about about actually working with them and then they together I think we're abeam it'll be amazing I hope you found that interesting maximize slide beyond having the reasons from the horse's mouth as it were I picked that up because I thought the philosophy of the tool which is really people and an organizational culture centric comes through in their personalities and Lloyd made a really interesting point on the smart use of data and that I think sums up the way looker approaches data which is as a data platform which is being smart I'm going to I'm now going to talk through three aspects of looker I'm going to steal these into three aspects which I think make look as stand out as a data platform as I said earlier taken together they do something which I think is really really hard to do and that's instill a data-driven culture and vastly accelerate the value which with which organizations can get value from data and that's why it's such a revolutionary platform unconscious that sounds salesy so let's make it concrete so setting a stage for bi today you can pretty much generate generalize all bi applications look included into some mix of a server cloud side component some form of developer tooling and the front end usually web-based to present the outputs let's now shine a light on what what look at what what makes look at different and what it brings to the table so the first point is that it introduces a cloud native semantic layer so and excuse me for those who are familiar with this already this is essentially an abstraction of data as expressed in a in lookers case a fairly high-level easy to learn decorative language called look ml that language is used to model the data and it's these models that act as data sources by front end outputs such as dashboards by introducing a semantic layer to profound things have happened one is the form and storage of data as me decoupled from its usage and secondly the focus has shifted from the physical structure of the data to its meaning in the business so on the consumption side the bi team is free to model the data directly in Luke ml layer without necessarily having the final dish definitions materialized in the data warehouse so that allows modders to be prototyped and release production quickly without necessarily the data engineers getting involved and then at the right time that gap that can be pushed down into the data warehouse there on the data engineering side the - engineers have the flexibility to manage change - their pipeline and in fact they can even swap out the entire database that I entire data warehouse with minimal disruption to the semantic layer or the dashboards so the to have huge freedom to operate around the respective concerns the second element second profound change introduced by the semantic layer is that as I mentioned on the data consumption side the conversation has shifted from building reports and dashboards to the meaning of data because look ml is so easy to use the business is getting more engaged with shaping the data and their ability to experiment with it so there's a far greater engagement and fusion of roles between bi and business so in summary excuse me in summary the semantic layer gives freedom for the business to get insights quickly and do that through rapid iteration and that the point of rapid iteration equally applies to on the data engineering side as well [Music] the second element in which I think you'll agree is unique is its open architecture so Luke has an architecture which allows embedded dashboards to be embedded in web applications that's pretty much standard stuff in terms of how Luca goes beyond that unless look at the final mile in terms of data refinement and how other other have the external API is can tap into that so data scientists and statisticians can now use Python on our through native API to tap directly into the BI layer within Luca so this pattern has huge implications it means we're only we're only having to wrangle the data once and it's ready to be used everywhere data scientists typically spend a lot of time and energy getting hold of it and getting hold of and preparing data usually the navigating around the data warehouse with either too much or too little access and they now have easy yet step complete still completely governed access to clean data at any level of aggregation with any degree of richness supported by the underlying data models and for us data scientists that's nirvana and because the data is so clean and consistent coming from the BI layer with basic with business processes any models cat with with the underlying physical physical link can be plumbed into operationally operationalize usage really quickly developers can use a native Java and types create API to touch dial key the game directly tap the the bi layer so you can now build bi native application bilk sorry cloud native applications that directly interact with customers external suppliers and other partners using your your primary reporting a set of percent of data from your data warehouse so you can really go to town in terms of building data different products and that's a huge enabler for innovation and finally there's also native API for swift and cotton allowing an organization to tap the bi layer directly for native mobile experiences it's again that's massively powerful for companies undergoing digital transformation or trying to provide seamless customer experiences so you can see how that single looker male model is the gift that keeps on giving and these are patterns that no other bi tool can deliver with anywhere near the same level of ease and efficiency and going back to the cultural point you now have the foundation of a community which is a common language you've got stakeholders right across the organization interacting with the data using the same terms the same expressions and you can see how a data-driven and a collaborative culture is being formed by the ability to use data in that very Universal and democratized sense and you have conversations happening that haven't happened in the past so insight you summarize the open architecture brings massive value from data and allows organizations to really collaborate around data in a way that simply hasn't been possible using some of the more traditional patterns and finally reasonable blocks so this is something which is unique to look to look at in terms of the granularity and the modularity of which you can take off the shelf blocks of logic this can be business logic data sources public data sets transformation patterns custom visualizations embeddings and plug them straight into your bi solution so that means a team can get up and running in next to no time and not only only are you now creating a way of Russia hring content within your private organization community but there's a whole look of community of content that you can tap into and it's a pretty good chance of whatever you're trying to do some has already done it and shared it so you can just take that off the shelf customize it tweak it as needed and you've got an you're up and running with a ready to go solution and that's the revalue of creating something once using it anywhere and having fewer overall assets to manage and the ability to show them broadly and easily in summary massive value from data so without further ado I'm now going to hand over to Sebastian Fabri I'll bring him into the studio he's already here I Sebastian good afternoon I'll do a quick intro so fast is Sebastian Sebastian Fabri is a sales engineer for Luca and works with strategic partners so we actually worked together in my day job he's worked directly with enterprise adopters to help them change their approach to BI advising her date analytics and data strategy Sebastian welcome all right thank you for having me a pleasure to have you yeah I've been I've been about five years with looker now so hopefully I can bring a bit of context and also kind of show you guys how everything works both from the pure looker point of view both so how they can work with with with DCP now for me there's really two goals in this demo and when we cover that in a second so first of all kind of the quick summary of of what looker is right and so since the acquisition of liquor by by Google which was finalized a couple months ago now really where Google Cloud is positioning looker as the cloud native enterprises a platform and the goal there is to obviously secure access but we wanted to do that for near real-time data wearing where it's needed so that really just goes alongside what Imran was explaining already earlier on the code so the kind of short and sweet statement one slide and done now if you look at a slightly bigger picture of what GCP is offering it's a much broader list of solutions and looker really fits sort of like at the end of this offering you're able to get all this data from wherever you want wherever calculates into the DCP environment you able to enrich it you able to understand so treat it and use AI et cetera all the way from the sourcing of this information to the delivery into bigquery or other data stores and even once it's there you can keep this journey lucre is going to be there to really allow you to go further in terms of where do you want to make this data available do so in a way that's very governed so that everyone speaks the same data language and also do that in a way that's a lot more easy for non-technical folks for business user to understand actually be able to leverage all the world that you are doing right somebody needs to be using all the world that you're putting in and so you can really simplify that so now we're saying this two things I want to achieve here those two things are a give you an understanding an example of how does looker actually fit within all the DCP services and two is how does looker actually work so we will inevitably go through go through some dashboards that will give you a flavor of that a flavor of the integrations with DCP will also win definitely spend some time in the modding layer in the leukemia where you can actually see the code that's enabling this that's active that's acted on that's used to actually leverage your services to actually access the data and then deliver it to you in whichever form and format it needs to be now this slide is kind of one of the one we use and it's very close to what I was just presenting right you have lots of data sources that customers are going to be analyzing let's move that let's centralize that into something like bigquery right so the lots of you are very interested in bigquery that's great like it the scalability the power the speed and the number of features of bigquery makes it a really great fit there for centralizing that information looker will be able to leverage all the features of the query and we'll be able to do that simply through sequel right so it's what it's when it's meant for so look up connect to be query as well as a whole host of other databases we are part of this multi cloud approach the GCP as you're connecting to bigquery sending those queries the modeling layer allows you to describe a what is in bigquery be any transformation any business transformation you want to apply to it you know what is revenue what is active users how do we can calculate your retention rate all these things and then it's a data platform because it allows you to use that logic wherever you want it to be yes there is a bi dashboard aspect of it but also there is the the self-service that allows people to actually access data directly there is the publication to external websites there is the ability to send information to much learning algorithms and then leverage those back inside of lucre getting you're getting looker to alert you when something happens so that you don't need to worry about this look I will let you know when this has happened and you can come in and look at the data or act on it being able to do all of that programmatically right leverage api's or Flickr so let the look API but also leverage the API of other applications from looker and do all of that from from the platform that you choose in in the in the platform a cheat sheet so it's kind of Richard I wanted to sort of as a last slide gives some ideas some example of how we can leverage some of the other features of GCP this is far from an exhaustive list but this is gonna speak actually to I'm gonna show somebody's in in the live demo right - kind of a big query straight forward the data is in the query we can access it well the feeds are nested whether you crush on your table whether partitioning them no issue there we have that very neat way of writing the sequel is going to be able to leverage that you can combine sources combined eight command different data says that GCP makes available to you for free using App Engine to create more complex applications either that look provide data to for processing or actually as the dekat you want to provide and show to your customers or maybe a page is actually enriching the capabilities of liquor and I'll show you one of those but you may be quite much learning you can use that directly from liquor right thanks to the sequel interface tensorflow look okay and sending utterances flow and you can run your run your machine learning model or maybe actually you can also call tensile models directly from the query ml and then make it available in looker Cal function is really where we can dream up right right your database create slide deck figure more alerts have special like sending SMS or changing the the operation of the machine or of certain keywords away for whoever it is and then you kind of jump a little bit of GCP maps right for example again what we'll be using there is there is the Maps API prove your route see the street view at a distance is excetera even though some of these can also be done directly or it can be queried yeah this is we're lucky were lucky enough to be uniquely positioned to be able to leverage all these things and enrich the information and rich experience that you're delivering to your to your user so simply I'm gonna try to show you some of that now over the next 20 minutes or so so here's an instance of looker alright I'm connected to my browser it's obviously cloud-based and essentially everyone's going to log in here everyone will be able to go to the same URL and depending on who you are you'll see different reports you'll have access to different data you'll have access to different features and areas on the tool as the first step I wanted to highlight those integration that's actually probably more easily done as someone that is a business user and then I will link that when we jump to the model who just really came out I will link those feature back and show you how easily it was to implement now obviously I'm gonna show you the looker part of this there are go desk there's a the flipside of the coin which is the GCP side of it where you've actually where have you created the appengine where have you created the current functionality credit etc I'm gonna not I'm not gonna show you that that part today but it would be sitting you know in your DCP console alongside everything else as usual so let's in this scenario it starts as a CEO or at least someone fairly important in a in a retail company so we're trying to understand or how the business in general is doing and starting from a very standard you know a dashboard what's interesting here kind of a couple things first of all the way Luca is accessing the data each of these tile there is actually a specific single query sitting behind it which allows me if I modify any of the filters or whenever I actually open this dashboard these queries are being fired back to in this case big query for us to gather the freshest information right so I have to have a preview of the the latest values that were known and now it's going live and updating this with the latest numbers and I can go and try this automatically with with some of these others that was that was that was talking about so I don't need to come every morning morning to look at this dashboard if it's not something that interests me I can actually check and say okay I want to know if I am a certain amount below my sales target then it's supported from each I look at this and so just because we're not quite there so to scroll down and see what's happening in here so we have a series of different stores by doing business San Jose seems to be lagging behind so let's try to I see more detail or maybe I'm kind of busy so I'm just going to send the text or call the store manager and they can dig into that for me all right so here's an example of cooling Google App Engine it's gonna allow me to send a text message to the store owner right and on the other side on the other hand obviously the store manager getting these tags getting this message being able to drill back into looker in this case with a little bit more detail a little context because is so I've got more granular level to the information that's coming from from from store right and I see how things are going there somebody is obviously gonna be very similar to what the boss was looking at but let's shorten to the targets of this personal store and then see how are we doing compared to my peers so you know simply getting data from bigquery being able to calculate it being able to aggregate it in different ways comparing retaining granular access to my central say store but still being able to see the sort of summarized data of the other ones so that means I can see what other year-on-year sales for all the stores without actually having access to a granularity of their sales because I'm just in charge of one story another feature might be to actually combine that with other data sets are freely available in GCP right has the weather impacted myself how does that compare with you know peer stores over here versus my trendsetter it seems to be fairly you know standard so in this case it's not something that that matters but you could we could have maybe looked at you know rain and maybe look at other trends that are being made available through to look you directly through liquor and to be query and then to look around correctly so it's kind of like that bigquery integration working really well during a little bit further where could I save money right well the opportunity with my with my inventory they seem to be fairly small today digitize can updating life so some days I come here and there's 20 options and today there's only three but what's interesting on this slide here is just this forecasting stock all right what we're doing is very simply using bigquery machine learning to forecast the demand of of these items all right and then based on that being able to evaluate watch which one what's the value of missing stuff right which one should i stock more of which one should I stock less of right and it's a way of having much learning actually being very simply leveraged by now a store on a by a business user I don't need to learn they don't need to know exactly the intricacies of this model but being able to have a forecast number here that is agreed upon by the rest of the company and we will circle back to this and have a look in there looking at model how this was actually implemented it's very straight forward going down here same idea more detail for different stores I could go and actually see more information about one of these product if I wanted to - look that based on different segmentation these are based on comparison etc it's pretty straightforward to do I'm not going to walk you through all these examples here selecting back we could be using the cloud function here to look up images of these items for example and adding that directly looker we could be using custom shirts to enrich this information one way or another and then lastly I was mentioning tensorflow so for more complex cases like here we're trying to sort of bucket and understand the behavior of these customers having a tensor flow model that can go and go deeper in that information and still being able to leverage it from from lucre is going to be super valuable so there was a few example here we talked about bigquery obviously talked about cross functions what about App Engine talked about since the flow we talked about so the experience the high level from a business easier right all this kind of work has done been done in the background by technical folks like like yourselves now sort of like you can bridge the gap a user can go and ask me questions they have this ability to go explore the data they have this ability to actually ask those question themselves and leverage the work that you've been doing in the model by getting access to all the field that you've curated all the fields that you've made available to them and they can go and add them to these reports and add them to these calculations it's quite simply if I wanted to compare again let's compare some stores right let's get some store names going let's see total sales over the last 30 days that's already up there as a filter let's see maybe number of customers as a total number store quite simply I can go and make this request and looker is going to go and get those numbers to get those answers directly from from liquor so you have the freedom as a user to explore this information and to enrich it and add it so I can add this also to my table and pivot my stores for example that's gonna allow me to view this data from a from a different angle now I'll show you one last thing about about when this is you know almost unintelligible but maybe we can reduce to the stores are more specific to me all security here is apply at the same time so if I'm only allowed to see Philadelphia I wouldn't only be seen at Philadelphia same report sent to somebody Phoenix they will only be looking at Phoenix information and lastly if I am back to being that that that person in my son who's in store and I want to be tracked and I'm going to be alerted to these automatically it's up to you sort of make available the destination of those of those produce those integrations for scheduling if our I could be again sending text messages automatically if something happened I could be enriching data in Google sheets I could be starting new machine learning algorithms of feeding into them I could be working in marketing identifying users and and sending them to certain campaigns I'm listing I could be tracking ROI on AdWords and going and posing them automatically if those certain thresholds up are taken and the beauty is that yes you can do that as a one-off and send it but you can also have looker sort of monitor it for you through a shilling and kind of on the you'll go and say as this happened as something new happened is there a new data here that I need to send a certain threshold being taken etc so you can start automating automating those workflows all of this exists all of this is possible thanks to the model so what I'll do is I'll just try to hit some of those same point I was just presenting but from the from the looking outside all right so it's not working that say you were given an instance of looker the first thing you would do is actually go and connect that to your database right whether it's bigquery or another one but get that connection going straight forward JDBC connection nothing to complicate it there based on that you can then go and create a model so this this is the demo instance is a few different ones here some of these you can see marketplace these are some of the example of the blocks that can be automatically imported and will help with for example than this data or maybe big Google bigquery performance right or if you have Bing ads or little marketing etc all of these are will be imported automatically for you you just need to get them from the marketplace in this case retail is something that was you know developed internally and we can have a look at what this model look like so a few things to know from from the get-go all right so a you can already see all of this is code based all right looker really leverages sequel and so it makes sense that to be able to do and the best way sequel is code let's use code as well you will be defining all the logic that we need in these files and then essentially making it available to to the users through the interface as well as through the API to manage all of this we are working with git right so let's say I need some work earlier that I forgot to undo so I'm just gonna go in and commit my changes reset back to production actually my my code here is not on latest version so I can just go and pull that I've made them very easily for users for developers to kind of collaborate well together each of your own branch or your own sandbox you can do all the testing that you want without messing up anything up front then if things get complicated you can also create more branches and manage that it like you would any workflows but while we look at exactly here on the on the in the middle right well to explain what's in the middle let's start on the left side we a connected liquor to a database we then created this model what's happened well quite simply you the developer is are able to select the datasets select the tables that you want to work with and looker will automatically create a representation of those tables so all of this here is metadata right we haven't actually moved any of the data it's just description of what's there in this case description is the table and then for each of the columns they exist each of the fields will create automatically dimension photos and that does include message fields for bigquery so you have all the information like this and a trace exists it automatically delete email but you're going to need to enrich that right there's going to be other definition kpi is business logic etc we want to include that in here that can be something very simple like say how do I calculate mind about transaction or what's the definition of a percentage of customer transaction we're defining all these elements and making them available to the to the business user let's watch for a couple of example here cuz I think it's kind of important to explain what's happening the key thing or the first key thing to understand is this purple writing so I said multiple times I look I use a sequel and if you look at line 104 this is kind of you can recognize a bit of sequel there right now if as a sequel function there's a min down there in 111 that's also a secret function whenever you are describing a field declaring a field whether it's a measure or dimension looker you can write C code directly here but the great thing and what this purple writing is is a reference everything you do is referenceable so you only need to declare elements once you do that once in one place and then wherever you need it you can just go in code for it so you have that single definition single source of truth and that's how we're starting to build up governance right you have the governance of the field for example to calculate my person customer transaction I'm using my number of customer transaction that's fine right here it could be defined in a different file you could be defined somewhere else in the model you can still make use of it you can still call for it this calculation here is a count distinct reusing the field called transaction ID which is simply one of the columns or in this case a dimension referencing the column and to do this calculation we kind of 1 exclude orders that don't have a customer ID so kind of like clean up the data a little bit before running this calculation that's the great thing right I declare something once I can go every use everywhere well I find even better is what you need to do if you want to change something right maybe there's other values you need to filter out right maybe there's a different way of calculating transactions because the tables are changed or maybe there's a new column etc if something changes like how I want to calculate these changes I only have one place to modify all right I only have one place to modify this measure it's going to flow through the rest of the model update everywhere else are you it automatically and it's also gonna good update you dashboards your API calls your emails etc so everything stays in same through through this mode that was let's kind of first thing you have views and we're going to be declaring a different dimensions and measures in here we also have a model file where we'll be declaring simply there's a relationship between the between the data size between the tables how do these things come together right and the combination of both is what's going to be essentially controlling the interface that we're providing to people and to the API now let's have a look at a few things there was let's talk about this bigquery machine learning model for example alright so in one of the views here the one called stock forecasting we are actually a creating inputs for our model and we're then using determine in the same way that you would use it directly in the query and that will be able to leverage it because the beautiful thing is that big caramel is gonna generate those tables for us so then we can go and use them and combine them to with the rest of our our data so what are we talking about here right we need to do a stock forecasting and we're gonna do that using big KML we're gonna do that using linear regression we're trying to predict the real quantity and then there's a few different features we can that's right a few different parameters that we can define now the other bit is very important is going to be the input for you know creating and training this model right that's simply defined above here all these columns that you're seeing are coming from any of the definitions that you've already done in your model right so you're sort of ensuring that the input of your models is consistent with the business definition is consistent with the rest of what your company is doing so that everyone picks the same data language and your folks are gonna be you know accurate that's up to you but at least correct in the way that the data it's using so we're just selecting all these fields these are going to be all the features that we're going to be leveraging it gonna be using Animoto and then bigquery does all the hard work for us first generate the model train it second predict their predict the results right based on that we can essentially receive a table we have all these dimensions a measure that can be declared and we can then join it back to the rest of the information and now we have side by side a forecast and the actuals and what's in my study and what's in myself the cetera so that's the next nice example of that we were looking at store names and App Engine so let's have a quick search here that now will look like that's quite simply declared in looker good that's the dashboards probably not the right way to not try place to look at that but since it's the same retail and see there's one here exposed to Salesforce Explorer my data source was directly the other one that I wanted to look at is in the product all right I wanted to call my category manager again it does require you yes it does require you to have define the app on the App Engine side so you know in your Google console but then from looker being able to very easily parameterize that being very easily enter information pass it call it you kind of make that and enrich the availability inside of looker there's a few example of essentially the world that you doing in bigquery can be leveraged and well the world that you doing in DCP can be leveraged to looker look the looker model is gonna provide you with governance and then this governance can be leveraged by your users very simply can be leveraged by our API whether you want to show this internally like we've seen so far or even externally through for example portals and and such leveraging API calls for data you know making API calls using JSON in your pro tools in your custom application or even bringing the visualization directly from looker via dashboard in and single visualizations etc so kind of looker house governance layer looker can cover you data and then you can use the data in whichever format where and when you want it kind of goes back to that first slide I was showing you right that that information of what looker is brilliant I think thank you very much for your time I'm gonna pass this back to Imran here to go to the QA see if we ever any catching questions on there on spider Sebastian thank you very much for a really informative and interesting presentation and demo so yes we do have questions and I'm going to switch over to slider to look at them let's start with the ones that's been up voting you have 11 questions how integrated were bound is Google to looker do you need to grant permissions to look her to be able to use it nice doubled up with another question of what is the cost of using occur I think we should break them up so one is how close the integrated is Google that is looker to Google so the looker has always been clad as gnostic right and Google has a multi cloud strategy so in that respect we're going to continue to be able to serve all the clouds that doesn't mean that we're also not gonna kind of take on the opportunity to integrate with as many services and have much of the value as GCP can deliver right so we will sort of be able to integrate for the views for the customers are within the GCP product within the GCP client but it's not a pre requirement at all to be using looker you don't have to be a DCP customer to be or you don't have to be in the DCP cloud to be able to use looker now do we need to grant permissions to look her to be able to use it you will need to allow looker to query your database right so looker needs to be able to select data your database to be able to access that information I'm feeling that I didn't actually show you the sequel which is probably a bit of a faux pas there but literally every single report I showed you behind it there is a single query that lucre is generating based on your model so that sequel query needs to be about to run on your database Thank You Sebastian just going to run through these this interest of time if that's okay the cost of using look I know this isn't always transparent and and yeah fixed there's a there's a few different parameters there so there's the domain focus is sort of there's this a listing on a DCP marketplace that you're welcome to look at so you can purchase it through there but we also do offer sort of like volume discounts and those sort of things if you are if you want to have a discussion talking with with our ex CEO talking with one of our as a representative then they can adjust pricing based on on what value you're getting out of the tool like it's passed you know there will be a link to download and try to occur for free as towards the end of the session the next question Sebastian and I think it's probably called a popular one actually is where does data studio fit around Lucca yeah that's so that's a good question because GCP does have these two tools now right there we have well Google has these two tools right there's a studio has been there for a while and now we have they have a co-ed locker so there's a sort of the right tool for the Royal job I think is what we are pushing is looker is enterprise scales right so you need scalability you need security you need governance you need to be able to deliver data in multiple places right we're not just talking about dashboard here we're talking about a data platform that can be powering custom application that can be you know sending your email or integrating with you other application that you already have and integrate with those so it's kind of a low a lot of different features their large simply not part of the data studio wheelhouse and the studio on the other hand is going to be great for sort of like the person that is a bit technical that just one probably might be working in marketing once you get some answers from the adverts some answer from bigquery data may be combined out with some CSV files and mock up this dashboard and share it with a few with a few people so we do see this sort of coexist in terms of data studio and and looker in in several companies thanks fast you know I think that that answer will be really really informative for a lot of GCP users so the next question which has just picked up voted is regarding the Magic Quadrant Luka is currently in the challenger's quadrant completely lost from the recent report what gap or step is required to move into leaders so we we've been in there for two years now we moved from niche to challenger this year which is a great progression I think historically you've seen a lot of the current leaders I've done I've gone through that path so we definitely also this is very encouraging I don't know if I have a specific answer here right there is one big thing you know if you simply look at the axes here the way where we we have the the vertical one is execution and then the horizontal one is completeness of vision completeness of vision and so let's see what let's see how looker and GCP can go then can do together to complete that vision right and kind of foster more innovation and also enrich the capabilities of looker and then it's kind of up to you to the analyst to to make their mind all those things thanks pass you know I've got the Conn report openly in front of me actually and the two to the cautions are narrowness of product vision and geographic or geographic presence and I think the fact that is now part of Google is going to really address those in space in terms of broadness of vision and and geographic coverage so it would be interesting to see where Luca learns and next year's Gartner we're short on time so we'll try and get through as many of these as we can I suspect all of them the next question is do the reports respect permissions on the back end tables and views in the database what does look a require a separate authorization making so either right you can't get two options each has their advantages disadvantages you can either have looker access all the data Luka has the permission to access all the data and then apply the security within looker that would still be integrated with you user management so that's it they're still linked they're just not linked directly to the database permissions or you can have depending on the database but on the query you can have both and can pass through the users and then leverage the permission of the user from the database are things so you have options there great Thank You Sebastian this will take one loop on a question and apologies everyone II couldn't make it has been quite a few those of your questions so yeah Ark us are the customers you have looked as part of the data size environments or is it primarily for bi so definitely user customers are using liquor in identify in as part of their data science right it's not a standalone tool for data science but looker will be very valuable in those environment to provide data and provide governance into into the actual math learning right wherever which ever tool you are using for your models you'll be Luca will be able to provide information in the govern way and save you all that data prep or that data cleansing time and then at the other end of spectrum once the data is in process once you have made your prediction Luca will be great to actually present that back to the business users right so looker is kind of help it on both ends on the data prep and then on the data delivery for the things and you guys can work forth imagine they're in the tool bag that's your still you know the best possible with Thank You Sebastian and I think that's probably all the time we have questions if there are any further questions then you are very welcome to reach out to me on the look side of things and I can pass some sebastian when in fact we probably just answer them ourselves in terms of actually like sigh next steps to close so firstly from a community perspective from a user perspective you're very welcome to present a talk please contact me if you have anything interesting to propose in terms of and unique experiences or they'd like to share or anything that you think would be of interest to the wider GCP data community you can reach out to me on LinkedIn with this bit me or via the meetup organizer anyone feeling a bit nervous or a big bit don't worry we'll support you throughout the prep I really encourage it it's a great experience if you're interested in look a great tool we certainly drink the kool-aid here I strongly suggest trying it here's the link where you can get a free download look we know that you've come from the user group and you know opening up some additional benefits in terms of support from from agile exciting and finally we will be presenting a master class and looker and this is something I am Sebastian if it brings availability will be will be giving its online it's free and everyone is welcome whether using for work or in fact you're not working you just want to upskill you're very very welcome come along would make it happen Sebastian thank you very much it's been a great first session and we hope to hear everybody to have you everyone on again for the next one on date of all and the 27th of May I I I think and don't bring my trusty hand in again subscribe have a great evening everybody take care bye bye thank you bye Sebastian bravely [Music]