Save Looker-on Us State with airSlate SignNow

Fax signature block looker-on

we're here to talk about the block's directory and the blocks that comprise the blocks directory as you heard this morning we we unveiled a bunch of new types of blocks which only sort of diversifies the offerings and really increases the amount of cool stuff that you can do with blocks in the block's directory how many people have checked out the blocks directory in their life ever people how many people are actually currently using a block or blocks on their looker instance there are people in the back alright okay we're okay so that number means that the vast majority of you here are not getting as much value out of looker as you could be I'm sorry to be the one to break this to you but it is true blocks are a big deal they're really at the core of why looker is powerful and and why looker can help you accomplish your goals the thing about blocks is you know looker uses look ml and is a sort of open platform and what that means is we can take we at looker who deal with hundreds of customers can see the things that they are doing over and over we can reduce the effort because instead of making you do that thing over and over we can just give you what someone else has told us they needed to do with your salesforce data or to get a median we can improve the reusability because we're giving you code that you can then reuse throughout your your ecosystem we're really it's all about accelerating your time to value so that you know you're not we're not giving you a look or instance that's that's clean that's empty of content but rather one that is full of best practices that we've learned from working with all these customers the thing about blocks is they're not a black box it's not we give you the blocks code and you are then stuck with it it's just code just look I'm out right so you pull it out of github usually and then you can go hog-wild with it right so it's not oh you have Salesforce data and your Salesforce data we assume looks like this and so we will transform it this way and if that's wrong too bad it's you have Salesforce data most people do we know that your Salesforce data probably looks mostly like everybody else's but a little bit different because I'll be honest everybody's Salesforce data looks a little bit different we can give you the Salesforce block that matches where however you're pulling that data into your data warehouse or your database get you 90% of the way there and then you do the 10% extra customization right but now you've saved 90% of that effort and you know as we've said on stage this morning as we said in in the deep dive this morning you all influenced what blocks we choose to build and you directly influence it from here a join because if you have not already you should go to the product feedback wall and say put your sticker next to the blocks that you would like to see us build right the you absolutely influence it because we don't want to put blocks in the blocks directory that no one's using we want to say oh we keep seeing these customers run up against this same problem let's take that package it up and just make it available so that everybody can take advantage so that's what blocks are all about so the blocks directory looks different as of this morning because there are more types so we have analytic blocks we have source blocks we have data blocks we have data tools we have viz blocks and we have embedded blocks and we have three great experts here with me today we're gonna join me in talking through what each of these block types is and how it can help you get more value out of looker sound good yeah everybody with me are you excited yet all right we've got I got two woes so that's good that's a start by the end hopefully we'll be up to seven or eight so okay so the first type of block that I'm gonna go over as a brand new one rolled out this morning viz blocks so the idea here is custom visualizations have always been something that you could do on Locker except you had to have two things one you had to be hosting your own look or instance on premise which most customers don't and two you had to have enough d3 and JavaScript know-how so right the visualization to plug into your looker lots of customers do have that lots of customers are hosting on-premise and they're already using custom visualizations but way more customers are not getting the value out of this right and we know that there are more visualizations more specialized types of visualizations for special types of data that people would like to be able to access and put in their looker instance so what we've done is we've taken the ones that we knew were the most in demand we've built those and we've made it so that they're available on any looker instance and it's not now we are making these extra you know 10 visualizations available and that will be it until we add more to the library you still can put any kind of custom visualization on your looker but now you're not restricted to having to host it yourself so if you have the if you have somebody who can write you a d3 visualization or a JavaScript based visualization that does exactly what you need that's great you can put it on your looker as of looker 5 you know that this I think really opens up the ability to give people exactly the kind of visualization to effectively communicate all the value in in the data that they're bringing right because data is kind of useless to most people right you show them a table yeah you show them a visualization humans are very visual animals it connects with people that get a lot more out of it so let me show you what some of these visualizations in the in the library look like so we we built a reorder roll matrix so this is something that that folks use a lot when they're comparing different kinds of event streams you know they want to see ok if someone does this first what do they do second they want to see how that sort of heat map works we give you a collapsible tree because there are lots of sort of nested types of data that this is really useful for exploring hierarchical sets of data that this is useful for exploring we give you a sunburst because as jamie said this morning you know if you're a product person or a marketing manager and you're trying to understand how are people moving through my website okay they go to the homepage then they go to the product page and then they go to the shopping cart and then they fall out sunburst is a great visualization for understanding that we give you a Sankey diagram another way to visualize flows through sessions we give you a chord diagram so you can visualize flows or you know airline flights and connections between destinations all kinds of new data types that really are not well covered by the visualizations that come shipped with every looker now are covered by these and finally we give you a tree map so that you can you know see again those hierarchical relationships so these all ship to the blocks directory ASAP they will be available in Locker 5 we hope that you'll use them and we hope that you'll give us feedback about what other visualizations types you you need and I should say all of these are fully interactive fully customizable just like any other visualization that you'd use in looker this is not a static visualization it works just the same as any other visualization and we hope that you will find new visualization types that would be useful to you and if you build them bring them to us if you're willing to share them we'd love to take a look and if we think they'd be useful to other customers we'd love to put them in the blocks directory as well so that's that's viz blocks so now let me talk quickly about data tools so data tools are really about curating user experiences to something that's more curated than an explorer page which to some more novice users can be pretty intimidating right you just have a field list and then a blank panel and you have to figure out where to go from there but they're more flexible more interactive than a dashboard where you only have you know very limited set of filters so data tools are really built on parameters which is an feature that we've rolled out in in 4.22 which is rolling out right now so that you can curate these experiences focused on a specific task you can give somebody a tool for solving the problem that they're solving every day we're shipping some pre-built tools in the blocks directory but again fully customizable and we'll be shipping more you can vote on what data tools you would like to see us build and we'd love we can't wait to see what you build using parameters but just to let you know what we're shipping right now we're shipping a web - also that if you have you know somebody who manages your website and wants to see sessions and pageviews you know and they want to see oh I want this rolled up to the hourly level know the daily level and I want to change the filters that's something that just works that's gonna be right there and look or for them same thing with cohort analytics where you know you can do as Jamie showed this morning you know you can define how you wanted to find a cohort and then see how they compare to other cohorts over time and we've got more in the works so what this looks like here's you know a web analytics tool you can see those sort of core metrics at the bottom you've got your page views sessions you know at the top oh I want to you know change the filter and make that weekly or I want to you know look at it monthly and see page views versus pages per session all that just works okay so now I'm gonna hand it over to har see who's gonna tell you a little bit more about what she talked about this morning about data blocks so we officially launched data blocks this morning they're publicly available datasets that we at looker have cleaned modeled maintained so you grab these datasets from liquor you grab a model that we also provide and you combine it with your data to start getting insights we are launching with these four data blocks I'll go into detail on each of the data blocks some ways that you can start to combine these data blocks with your own data and then we'll talk about how you can start getting access to these data blocks so the first is demographic data from the American Community Survey so this data is at a very granular level you're looking at information at the state county zip tract and then a block group level which is just very granular and in five times more granular than just looking at this information in a zip code and this this dataset it contains population information age gender race income and so many other metrics that are collected by the American Community Survey so we took this data set we combined it with one of our retail data sets to see how purchasing patterns for different brands were affected by income and so we looked at two particular brands Levi's Ralph Lauren's and just by looking at the data we were able to see that most people no matter their income tended to purchase leave ice but for Ralph Lauren that didn't tend to be the case 85% of Ralph Lauren's user base tended to be within the 50,000 to 150,000 income bracket so we just joined in the ACS data that we had with our retail data set to figure out how income a particular metric from that set data set was affecting brand usage we also looked at this information population information at a tract level and so in San Francisco say we wanted to open up a new store and use store front targeting a certain population in this case we looked at users over the age of 65 how exactly would we go about determining where to place stores to attract that demographic so in the background there we're looking at percentage of the population that's over the age of 65 so the areas that are darker are more red or areas where you have a significant amount of the population over 65 and so the area is the circles that we're looking at are where we currently have stores and so it's easy to tell just by joining in this dataset that we probably want to add another store over there and then in the bottom left hand corner where we see a lot of yellow which tells us that we have a lot of users in our demographic that we're trying to target and so the use cases here are endless and the data is so granular that your your options are really endless the next data set is the weather data set I talked a little bit about this earlier today but we worked with Google to get access to weather data across the US and this data dates all the way back to 1920 and we are updating it and maintaining it on a daily basis so that you can get access to weather patterns in the past week in the past day and we've taken this data and we've made it accessible by zip code so that you can easily join in by zip code to any any date and can kind of get weather patterns for that day the precipitation the rainfall snowfall and and just other other characteristics so we worked with SeatGeek to look at how weather patterns affected their festival listings and how it affected prices and how it compared to other types of events that they tended to host you can join this data set into any e-commerce data set to even look at how weather patterns affect purchasing patterns to maybe see how for example how many raincoats are bought compared to the weather for that particular area the third data block relates to financial indicators so this is updated on a monthly quarterly basis and contains indicators like unemployment debt inflation and interest rates and so we combined this again with a ecommerce data set to look at how revenue can be adjusted for inflation and we saw that the actual revenue when looked at compared to today's dollars were significantly different so how can you start accessing this data you can go to one of our customer hosted buckets you go to GCS or s3 and grab the raw data in its raw form or you can get a connection to a low core database you can connect to one of our big query instances or snowflake instances and get access to that data and once you've get gotten access to the data you can go into our blocks directory and grab our our model for that particular dataset and start to extend that model and start creating your own custom metrics for that data so to get started you can get access to that data in GCS or s3 or connect to one of our liquor databases you can import the look ml model enable model extensions essentially extend and build off of what we've already built for you and then start creating your own custom metrics the neck type of block I'm going to talk about or are embedded blocks embedded blocks are a way to grant access to data to your customers and your users just outside of the liquor interface you're bringing the data to the people where they want it in their normal workflow and embedded blocks come in a variety of different ways the first is just public embedding so you can take any report any dashboard from liquor embed it onto a public website for anyone to come and play around with that data and we're powering the visualizations behind that you can also embed into an internal site like Salesforce or confluence and you can take visualizations from liquor put it in Salesforce or one of those sites to just be a part of the users workflow the third type of embedding consists of building your own custom portal for your customers to allow them to authenticate in and with that type of embedding you have complete customizability over how you showcase data from liquor and the first way which we cover in our blocks is just being able to use our modeling layer to pull up metrics from liquor so here we're using our liquor API to pull in metrics like loyal customers total revenue number purchased and we are then surfacing this data on another site and making that data available and so our blocks walk you through how to make these API calls and how to render render the data in your own site you can also embed liquor visualizations so here we've taken an entire dashboard and embedded it onto a customer site and we our blocks again walk you through that whole process of how to get a dashboard from liquor and how to embed that iframe into your application and our blocks also again walk you through how to save and schedule and how to create an explore interface outside of liquor in your own portal and they walk you through how to set up a full use case for your customers and with that I'm going to pass it on to you Dylan thanks Aarthi cool so I'm gonna talk about analytic blocks and analytic blocks are blocks that are designed just for very specific pieces of analysis so things like cohorting retention period over period specialization stuff a lot of you guys probably have in your instance this is what analytic blocks were designed for and there are pretty much just pieces of look of metal in sequel blocks so really easy to package up and deploy it what's really interesting about analytic blocks is they were actually one of the inspirations for our entire blocks program so a looker across our entire customer base it's really rare that we see a new type of analysis odds are one of our other customers has done this before so naturally us analysts would constantly reference old design patterns old sequel queries whatever we're getting started with new customers so we had a mechanism for sharing all that kind of stuff internally and it was really great because me as a new analyst I could start if a customer requested retention I could know that you know not only can I find that pattern I can be sure it was tested it's gonna be the most i'm pattern and it's gonna be the most effective for the type of analysis that i'm trying to perform here three points about analytic blocks first again it's best practice design pattern so whatever is most performant it should be the most suitable for the type of analysis you're doing again tested and proven with a lot of other customers as well so we know it is the best best way to perform that analysis lastly we have dialects specific analytic blocks so these are things that utilize user-defined functions so one example we have is weighted median in bigquery and we built that using user-defined function we have all sorts of those for various dialects across our databases and what's great about these is we have a like write-ups and instructions and live demo instances for all of these so I can go ahead and show you a live demo really quickly so here we have in our discourse page discourse is our community forum on liquor we also have all these blocks listed again in our blocks directory so you could navigate to them either way for each of these we have a write-up that includes all sorts of information show some visualizations of the analysis it talks about how to get up and running with this analysis how to implement the block the types of data it works with the actual block code itself and then people go and comment and add all sorts of questions and stuff which we respond to but what's really great is we actually have a live link to this individual report that we're looking at here so you can go in demo and play with this data yourself so I'll save you the load time sweet so here is just a full embedded Explorer experience it's the same Explorer you get what any other time you're using looker this is publicly available so anyone can come in and just start playing with them stuff on their own you can substitute fields just like you would any other normal looker experience so say I want to look at my total amount spend and instead of my percent active now instead of seeing percent I'm saying how much each of my user cohorts spent over the last 12 months and so again I can go and cut and slice this further as much as I wanted to retention say I wanted to look at this by our source you know how we acquired those cut which medium is the best for retaining them I could add some fields I'll spare you all the work but essentially just add some filters and fields and we can see that oh it looks like our customers that are coming in through search are constantly being better retained than those coming through display for whatever reason that might be so we have about 30 of these analytic blocks open so far our customers and partners are constantly contributing to these and there's again there's full write-ups on our discourse and web pages so I encourage you to go check all these out and with that I'll hand it off to Aaron who's gonna talk about source blocks okay so I'm going to present on source blocks so source blocks are look Amil templates for expected data sources so think sources with common tables common fields that we can model in advance to provide quick time to value so it includes establishing table relationships key metrics as well as dashboards that we think they'll be useful for all of our customers utilizing both looker and that source product so these are created not only by looker but also in conjunction with our with our partners who have key knowledge about their sources and then oftentimes with feedback from our existing customer base who's using these blocks to in their own their own business so just to give you an overview of our currently available source box I've grouped them into three categories so the first category is web analytics and personalization so this is really where our source block program started we realized that analyzing event data was pretty challenging and that our partners had common data structures that a lot of our customers were using so we worked with segment with heap with snowplow on developing some of these blocks to make web analytics just more accessible and easy to start from in looker we then took that same concept and applied it to business tools I know this is a broad category but really just tools that our customers were using to drive their business that they also wanted to analyze the data from and looker and through our ETL partners as well as just scripting that data and you know having it in their data warehouse themselves we were able to work with them work with our own use of our own data sources like Salesforce like Zendesk to build blocks for those tools as well and then ultimately you can customize those to join that data to other data sources as well and then finally we're seeing a lot of demand for marketing and advertising data so partners like Google have made this really easy to get sources like Adwords and double-click into bigquery additional partners like and in privado make it really easy to get other sources in your data warehouse as well and then we've just created a template to make it easy to look at overall campaign metrics like impressions like clicks conversions and then you can ultimately join that to other sources and look at things like cost so just to provide an example of a source block and I figured I would cover segments and who's one of our sponsors at this event and also one of the first partnerships I worked on at looker and one of the first blocks that I helped develop so if you're unfamiliar with what segments does they help you track your data from one place and then send it to their library of integrations so tools that you use to run your business ultimately they decided to make this data available to their customers in a data warehouse and from that we were able to create a web analytics block to help you get started with segments so some things that the block helps you do is map your user IDs to one universal user IDs so you can track your users across visits as well as across the their lifetime on your on your application you can take these mapped user IDs and you can create the notion of a session so session izing events into a visit you know with a period of inactivity marking the end of that session so you can look and see what they've done within the confines of the visit to your application and then finally taking those mapped users taking those sessions that you've created and liquor and then using them for more advanced analysis like creating funnels like doing cohort analysis and things like that so what has this block led to so in addition to making it really easy to analyze web analytics data you know we've worked with segments to make sure that our mutual customers are having a great experience and looker across all their products so as they made their warehouse is the data available and other sources like post crafts like bigquery we've adapted the block to support those as well additionally they've launched data sources which includes SAS applications like Salesforce stripe Google Adwords Facebook Ads we have blocks for all of those that are either available on our directory or will be launched soon and then finally segment was one of our first partners and to participate in our newly launched action hub so you can actually take the analysis that you've completed and through the blocks so the web analytics data your sources data and act on it by actually sending the results of those analyses back to looker so as an example you could create a list of users in looker and then send that back to segment and and then through segment send it to Marketo to launch a campaign or something like that based on analysis you completed in looker so really completing the look though we've there with that I'll pass it back to Daniel thanks Aaron so I used to be a customer I was a very early customer of lookers and back then blocks wasn't a thing I feel like I'm like we had to walk uphill both ways in the snow and we had to build our own ETL pipeline because none of our great partners existed and we had to build our own sort of look ml to interpret that data and all that I'm probably not supposed to insult the audience but you would be crazy to do that now if you are using one of these sources if you are trying to accomplish one of these analyses if you are trying to use one of these visualizations you can get 90% of the way there you may be able to get a hundred percent of the way there by grabbing one of these blocks they're free they're available they will spark ideas for you that you may not even thought were possible because you know I know the thing about look ml is you know it's based on sequel it's a wrapper around sequel right and so lots of people think well sequel is great for some things and terrible at other things you wouldn't build a recommender system in sequel and that's true you would be crazy to build a recommender system in sequel but you actually can do it in look ml because look ml let's look or write the sequel for you and while it's not sequel that a human would want to write it's totally fine for looking to write it it's a pretty concise local male model so it can say oh this song is like this song and I will recommend this one you can do that and look em out you can do retention analysis you can do kaplan-meier survival analysis in look ml we have blocks for all of these things they will simplify your analytics because they will let you do things in really concise nice ways because you get to benefit from all of the knowledge that all of the other people here have gained over their years of looking using looker they will let you reuse this stuff and benefit actually not just from what the current state of the block is but what future states of the block are because as we update that you benefit from those updates you can merge those in and of course they'll accelerate your analytics right we had to write we had to figure out how to session eyes our data when I was a customer that was a pain it took a while to figure out how could we do that in a performant way you don't have to do that now you can just grab the code and go so I really encourage everybody to go check out the blocks directory it is a huge resource I want to see everybody using it I think you will all get so much more value from looker and the people that you serve at your companies will get so much more value from looker if you take advantage of this stuff [Applause]