Designing beautiful, effective dashboards: Best practices for visual story telling in AI/BI

Summary

AI/BI Dashboards on the Databricks Data and AI platform support Gantt charts, faceting, parameterized text widgets, and theme controls for font size, font family, and corner rounding, giving teams the tools needed to make dashboards look polished and professional.
Effective dashboard design requires clear visual hierarchy, the right chart type for each data story, consistent sans-serif typography, and corporate brand color palettes—all of which can be applied across an entire dashboard in bulk using a single Genie Code prompt.
Responsive mobile design is a first-class concern in AI/BI Dashboards, with specific chart recommendations for small-screen layouts so dashboards deliver the same clarity on a phone as on a large desktop monitor.

Designing beautiful, effective dashboards: Best practices for visual story telling in AI/BI

A great dashboard isn’t just fast and accurate. It’s intuitive, engaging, and easy to act on. Yet many fall short, cluttered with too many visuals, unclear layouts, and designs that don’t translate across devices.
AI/BI Dashboards make it easier to build polished, interactive experiences, but strong design still depends on thoughtful choices around layout, hierarchy, and usability. With rising expectations for mobile access, dashboards need to deliver the same clarity whether viewed on a large screen or in the palm of your hand.
We’ll cover practical design best practices to help you create dashboards that don’t just display information, but make it easy to understand and act on.

Design dashboards with clear visual hierarchy and focus
Choose the right charts and layouts to communicate insights effectively
Design responsive dashboards that work seamlessly across desktop and mobile
Avoid common design mistakes that cause clutter, confusion, and misinterpretation

Talk By: Eason Gao, Senior Product Manager, Databricks ; Mohit Hingorani, Product Designer, Databricks ;

Chapters

FAQs

How do I make my Databricks AI/BI dashboard look professional?

This video recommends focusing on three goals: preventing viewers from drawing the wrong conclusions using trend indicators and conditional color, making dashboards easy to scan at a glance through clear visual hierarchy, and applying consistent typography and brand color palettes. Starting from a well-structured layout and applying these principles step by step produces dashboards that look polished and are easy to act on.

What chart types does AI/BI support?

AI/BI Dashboards support a range of chart types including Gantt charts, standard metrics, dense information-heavy tables, and a variety of interactive visualizations. The key design principle described in this video is choosing the chart type that best communicates the specific data story rather than defaulting to a single format across all metrics.

How do I apply a brand color theme to an AI/BI dashboard?

AI/BI Dashboards support corporate brand theme settings that control color palettes, fonts, and visual style across an entire dashboard at once. Genie Code can update the full palette of a dashboard from a single natural-language prompt, making it fast to apply brand guidelines without manually editing each chart individually.

How do I design a responsive dashboard for mobile in Databricks?

AI/BI Dashboards include responsive mobile design capabilities, and this video provides specific chart recommendations for small-screen layouts. The guidance is to place the most important metrics in the prime real estate at the top of the view and simplify layouts so they render clearly across both desktop and mobile devices.

Full transcript

[00:07] Hey everybody. Welcome to our talk. Hopefully a fun one to round out the event. I know many of us are probably headed to the airport at some point this afternoon, so for today, we're here to talk about designing AI BI
[00:19] dashboards that look really nice. Um Uh you know, we have you know this generic forward-looking statement side that we have to put up, but everything we're talking about today is actually generally available. Um and the team's
[00:32] made a lot of improvements specifically on look and feel in the past year. Uh so we're going to talk through all those mechanisms, try to give you guys as concrete of advice as we can on how to leverage those tools. Um so we can all walk out of here building dashboards
[00:44] that just look more presentable. Um so my name's Ethan. I'm a product manager responsible for dashboards here at Databricks. I'm joined by Mohit. Want to introduce yourself? >> Hey everyone. My name is Mohit. I am the product designer on AI BI dashboards.
[00:58] >> Cool. All right, so plan today is we're going to go through, you know, at a basic level, you know, what makes a great dashboard, and then we'll get into actually specific things like picking
[01:10] charts, designing visual hierarchy, applying applying typography and color, and then each step of the way kind of talk about the new features that we've developed, and give you these tools so you can go out home and try these.
[01:22] Um and to kind of really hone the message, we're going to start from a bad dashboard, and Mohit is actually going to show you guys step-by-step applying those principles, turning it all the way into something that that looks and feels great.
[01:34] Um and when we think about, you know, making dashboards look nice, and I've been on customer calls, you know, with a lot of folks maybe in this room. Um there's been a lot of kind of interest in getting dashboards to look better.
[01:46] And this they kind of fall into three categories. One is hey, I just don't want my viewers to take away the wrong information, right? So you throw them them a number, and it's like, hey, you know, sales were a million last year. Is that good or is
[01:59] that bad? Should I be concerned?" Right? Like there are tools to help you contextualize these numbers that we'll talk about. Um we also want it to be easy to read. You know, the best dashboards, you know, the moment you get on that page, you roughly at a glance can tell if your business is doing well
[02:11] or not well. Right? Like they're they're not meant to be these like really long reports that you read or at least if they are that way, you know, maybe there's a better medium to present that. Um and then finally, just getting them to look nice uh is important, especially
[02:23] if you embed dashboards for your customers. Right? So, ABI dashboards can be embedded externally now um and you want it to match your brand, you want it to match your theme. You want it to feel high quality so that the product that you're selling, right, is putting is in
[02:35] front of folks who who perceive it to be good. Um but even internally, we're finding that, you know, really high quality polished dashboards earn a lot of trust. So, if we've probably all seen dashboards that like don't look super great and you also kind of
[02:48] subconsciously makes you doubt the quality of the analysis. Um so, that that's an important one to consider as well. Okay. So, a tale of two dashboards. Uh today we're going to turn this dashboard on the left, which isn't bad, but but
[03:01] maybe that's what a lot of our dashboards look like today, at least a very little effort. Um and we're going to show you how to turn it to the dashboard on the right. Um and it's going to be more obvious when we go full screen that the combination of colors and typography and layout uh makes that
[03:13] one actually a lot easier to read and also makes the insights uh just quite a bit more powerful. Okay. So, there's kind of like three key areas we're talking about here, right? So, the first one is and and this is a
[03:25] common thing that, you know, a lot of dashboards start with is there's no context with numbers, right? So, you get a single value, you're like, "Hey, total number of properties 158." Is that good? Is that bad? Is that expected? Right?
[03:38] What you can do is now our single values have what we call sparklines, they have uh ups and down percentages, percentage change, and things like that. So, it contextualizes that, "Hey, actually, 158 is awful. Like we saw 78 58 57% drop
[03:52] year over week uh week over week, and we should be really concerned, right? The next one is showing relationships. Um so, conventionally, you know, if you had a couple of different charts, you wanted to There wasn't really a great way to overlay.
[04:04] Um so, here you're seeing kind of overall order volume, um and then you have kind of your average prices. Um and when you overlay that onto the same chart, uh that actually makes it quite a bit easier to see. Um
[04:16] so, that's another technique we'll walk you guys through today. Um another one is adding meaning with color. Um I think we all hate kind of a blank flat table. You know, it's like unless you're doing your own manual sorting, it's difficult
[04:28] to tell if there are any key trends. Um we've added a lot more conditional coloring this year into pivot table and also tables. So, for example, here, you know, your eye immediately goes to the dark greens and the bright reds. Um and depending on your business context, I'm
[04:40] sure we can all think of use cases where, you know, that's where we want to draw the eye of our viewer. So, Mohit will show you guys kind of where those controls live in the product as well. So, that was just theory, but for now, I hand it over to Mohit to actually show
[04:52] you guys how that works. >> All right. So, this is the dashboard we just saw in the presentation. And what we're going to start doing is changing some of the settings we have here uh to look a little more presentable. So, the first thing I'm doing is making this
[05:05] uh widget, the text box, a little bigger, and actually updating the header here to maybe heading two or heading one, actually, make it a little more bolder. So, it makes it very clear to the user what they are uh
[05:17] looking at. The same thing I can do for the subheaders as well. Let me Let's try header four. Um the next thing you can see is all these filters in our dashboard sort of looking lost because it's sandwiched between all these charts as well as the
[05:31] uh text box we have above. So, what I can do is actually move some of these filters off to what we call the global filter panel. So, you can cut and paste these over here. And the great thing about global filters is that this will apply to all
[05:44] the pages in your dashboard. So, you don't have to recreate them per page. They will work between all of your pages. So, let's do that real quick. And there you go. So, all your filters
[05:56] have been moved to the left side panel over here. The next thing that we'll touch on is these uh counters. So, like Ethan said, just saying a number doesn't mean much, but let's add more context to this. So,
[06:08] the way you do that is let's start with adding a date here for our data point. And what we want to do is sort of create a trend of base price over time. So, here you go. Now, you have average
[06:22] uh base price over time, uh which is changing by the month. We can keep going further, and maybe let's do a percentage change, so that you can say, "Whoa, something doesn't look right here." And that's because I had picked some
[06:34] base price. Let's change it to average, and this looks a lot more reasonable. So, we have seen 0.02% change in the last uh month.
[06:46] Maybe a This is up to us, really. What if you want to do, let's say, 12 months? Uh so, you can change uh the numbers to whatever you think suits your business case the best.
[06:58] Let's do that again for another number. So, this is for total available properties, and we added a date. Where did my date go?
[07:12] Well, you know the demo effect, right? It happens. Um but, I believe me, we can have sparklines here on all these widgets as well. So, you have your guest capacity here for all our Airbnb uh for our Wonder Breaks properties,
[07:24] and we're going to add uh guest capacity. That's the count, I believe. And you can set it to something like this, then. All right. The next thing I want to show you is how
[07:36] you can use some advanced functionality on your charts um to kind of change the story you're trying to tell. So, this is a very typical uh stacked bar chart. And what I would like to do here is turn
[07:49] this into a grouped bar chart. So, I can do that just by clicking like this. And then, what if I want to move my legend to the bottom? I can do that so by going into color properties, uh
[08:01] legend, and assign it to be at the bottom. There you go. So, this is how you can update uh some of your chart settings. The next thing I'd like to show you is
[08:13] the combination of some of these charts together. So, let's say you had these two two charts, and I want to create some sort of a correlation graph between these two. It's actually really
[08:25] straightforward. I go from bar to a combo, and then what I'd like to do here is add the line on top of this. So,
[08:37] we go into line, and then our base price, I'm going to select average base price, and then switch to a dual axis, since both my axis look very different. And I can simply delete this
[08:50] chart away. So, nice thing about this is now you have two charts into one, and you're telling a more cohesive story. And that actually enables you to do more advanced functionality, such as faceting. So, what if I want to break this chart
[09:04] by property type? Hey, how are my ski resorts doing as compared to summer gateway getaways, or the urban year-round top property types? So, by creating a facet, I was able to break
[09:16] this one chart into four charts and tell slightly different stories, and get more detailed insights. Like, the ski resort is far more the price changes a lot more over time, whereas you have some of the other ones are a lot more stable.
[09:29] The next thing uh we're going to talk about is using color on tables. And And is really critical because often tables can be dense and we want to be able to highlight some insights in these tables.
[09:42] So, let's say I have the base price over here. I want to set a conditional color on it in
[09:54] I click on conditional style. And then, if the price goes over 200, I can highlight those separately. And maybe below 200, I'm fine with it
[10:07] being okay. That's fine. So, this is a great way for you to like emphasize certain numbers. The other thing I want to highlight is the ability to set up banded rows in
[10:24] tables now. And this is critical when you have a lot of dense information. This helps for the viewer to be easily see these numbers um and help their guide their eyes a little bit. And the next thing I want to show you is how to add color to pivot tables. So, as
[10:38] you can see, it's really hard to make sense of all these numbers that are here. I'm going to go over here, go into style, go into color scale, and voila, I already see some colors uh showing up on my
[10:51] uh pivot table. Let's take it a step further. Let's maybe set up a red-yellow-green diverging color gradient. And here you can clearly see that the price was really low over here and it's way higher
[11:04] for Skittles and M&M's. So, it's a quick way for you to like see insights uh without putting in too much effort. Um over to you, Ethan. >> Sounds good. Do you have throw the slides back on?
[11:17] Thank you. All right. So, we looked at charts. Uh the second category that we kind of want to talk about is uh visual hierarchy and layout. Um this is one that's actually pretty straightforward, so we'll go through the
[11:29] slides pretty quickly. Uh but the human eye kind of especially in Western countries where we read left to right generally scans information in a known pattern. So for any kind of like visual
[11:41] heavy page we tend to read in a Z. And for any kind of text heavy document we tend to read in an F. So what that practically means in both of these cases is that your golden real
[11:53] estate is always in the top left corner and your first row is is by far the most critical and the one that your eyes will be drawn to the most. I think we all intuitively know this and we probably design dashboards already this way
[12:05] but it's helpful to kind of think through that means hey you know your most important key in your table for example if you're dividing up by customers like that should be your first column right or if you're going to talk about hey this is my overall sales and these are the contributing KPIs for
[12:17] those sales you want to do it in a row. The first number is top left and then the contributors are around the right right so then there's a lot of kind of continuity as the eye scans through these numbers. The other thing that kind of helps here
[12:29] is anytime you're playing around with multiple widgets on the canvas the relative size actually makes a big difference. So in the left here they have equal sizes. As a viewer you know this is this is that case where you're not exactly sure where you should read
[12:42] first or what is most critical. But if you just break it out a little bit right you know forecast is top right corner not everybody always cares about that but you know this scatter plot is really the key point. Just adjusting it
[12:54] a little bit right going from 2/3 to 1/3 vertically and horizontally can create a lot of emphasis and also guide the eye towards you know from the first thing they should read to the next thing they should read. One more thing that helps and and when
[13:07] we look at a dashboard that makes us as humans feel comfortable a lot of the time it's because it's using white space. So we've developed actually a lot of kind of controls this year to let you adjust margins and padding between
[13:19] charts that we'll talk about but the bottom line here is you know anytime especially if you're trying to convey a lot of the information giving it a little bit of white space really helps in this case. Um and you know, especially if you have
[13:31] consistent gaps, uh that makes the whole dashboard feel calm. Um otherwise, you know, I I feel like we've all landed on pages that are so text-heavy that like we're intuitively stressed just like looking at this dashboard and we're not even really sure why.
[13:43] So, we'll talk through that, too. So, for that, I'll hand it back over to Mohit. >> Great. Let's do this. So, let's say I have a few widgets over here that are all of equal size. Now, a viewer's eye will have a little bit of
[13:55] trouble struggle they'll struggle to figure out which one to look at. What I'm going to do here is kind of play around with uh the tables uh the charts that we have to give it the right amount of
[14:07] importance. So, let's say in my use case, I have a lot of important geospatial data. I can make this map a little bigger to provide more emphasis to it. Maybe not that big.
[14:19] And then, to make space, I just moved my pie chart to be on the right and to be a little more smaller. The next thing I'm going to do is play around with some of these properties here to make my chart my chart a little
[14:32] bigger by moving the legend to the right-hand side. And this is actually a really nice trick to provide negative space by just by moving my chart from the bottom, I ended up activating the space a little more by giving it more
[14:44] negative space here. The next thing I'm going to do is maybe adjust the size of my sparkline and update this to be a little more bigger. So, there you go. Now, your eye will go
[14:57] from here, they'll first see the map, then go to the pie chart, then come down here to understand my area chart, and then over here to the scatter plot.
[15:10] Other thing I want to also emphasize on is internationalization. You should be playing around with the size of the widgets, but keep in mind of your international audience. Let's say, for example, in German language, the text use is usually a little longer. So,
[15:22] you want to make sure that there's space for it so that colleagues can also view the dashboard the same way that you do. Over to you, Eason. >> Cool. Okay, I I think this is the the fun part. Uh typography and color are two
[15:36] areas of investment that we've made this year. Um and specifically uh we've added a lot more font support uh to dashboards. Uh we heard requests like, "Hey, my company actually uses their own custom font. I
[15:48] want to put this in front of my viewers in an embedded dashboard." Or I'm trying to do this very polished internal report. I want it to look like my brand. Um you can now do that, right? So, now you have your sans-serif fonts, your monospace, you have your serifs. Um and
[16:01] you can also upload a local font. Uh so, when you upload that, um any user in your company uh who also has a local font installed will automatically pull those fonts in. Um so, you can distribute things kind of really tailored to your brand.
[16:14] Um kind of the the recommendations we have for this uh is sans-serif fonts are quite good at kind of reducing strain, especially in layouts. So, kind of that's like Arial, DM Sans, Inter, the ones you see on the left. Um monospace
[16:27] is really good for uh tables and numbers. Um and we've also given controls this year so that you can actually pick different fonts for kind of the numbers in a chart versus the title of the chart. Right? So, you can actually uh pick and choose these
[16:40] combinations. Um it also kind of implies tone, right? So, and every time you give a like kind of serif font, um things like, you know, uh Times New Roman, like that conveys
[16:52] kind of authority and it's a little bit more conservative. Uh but something maybe a little bit more fun and a little bit more expressive would be the sans-serif font. So, that's kind of up to personal taste and your own brand guidelines. Uh and so, we don't have to really say that much more about that.
[17:04] Um one common thing that we see uh is there's kind of a desire to create maximum contrast, right? So, you want like completely black text on completely white dashboard. That that creates a lot
[17:16] of strain on your eye, actually. And the one subtle thing that a lot of kind of designer dashboards you see in demos do is that they turn the blacks and make it ever so slightly gray so that it's a little bit less intense on your eye. So
[17:29] this is an example where good contrast is actually that middle tile and not not the bottom one that that sometimes is the default. We're also going to talk through some recommendations on picking colors. So
[17:41] you know, for those of you who actually have a corporate color palette, you know, that makes your job pretty easy. You kind of just put that into a dashboard. You can load it into the workspace as a workspace wide theme and you can just consume them in dashboards. But say, you know, for example,
[17:53] Databricks has our iconic, you know, like our orange. But then the other colors, you know, we're kind of free to play with. There are tools where you can just put that orange in and then I'll generate the rest. We'll show you guys what that looks like.
[18:05] And just to recap, you know, we did a lot of themeing improvements this year. And we'll talk through these in the demo. But things like I want to customize the margin. I want to customize the padding. I want to round the corners of my widgets cuz that's what my application looks like. Things
[18:18] like my my axes colors. I want to just customize that where where hey, like I want a different font like I mentioned for my widget title versus the the widget description, right? These are kind of very specific tweaks. But we
[18:31] really wanted to get these things out the door this year cuz we're just seeing so much more demand in AI and BI. It's actually become like the the presentation layer whereas traditionally we kind of see more kind of engineering internal use cases.
[18:43] With that said, I'll hand it back over to Moheet. >> All right, look guys. Let's do this. So the theme settings are available to you up here in settings and themes. And on the left hand side on the right hand side, you'll see a full menu of
[18:55] options that Ethan just talked about. >> It might help if you command plus the little bit for the folks in the back. >> Okay. Let's try that. >> Maybe one more? >> I think it just did the chart. >> No, it made the menu bigger. Yeah, nice.
[19:07] There you go. Sweet. >> So you can pick a custom theme from here. And also we recently added a bunch of new themes so you can quickly create slightly more stylish looking dashboards. Let's try San Francisco for
[19:19] example. And the great thing about themes in the IBR they come with a light mode and dark mode, which means you can customize each one to your preference. And this is great for our new use cases like dashboards on mobile.
[19:32] If a user has a dark mode set up, you will see the same dashboard in dark mode. So let's go ahead and start editing some of these colors.
[19:47] And I want actually point out one more thing. Uh these themes are also available at the workspace level, so you can set up dashboards to have that workspace theme across your entire workspace. So let's go change the background color
[19:59] in the back to be slightly more towards the yellow. Um I'm going for really earthy tones because Wonder Bricks has uh more of those colors in their palette. I personally
[20:11] recommend not going too far from the white in your main widgets to maintain a good amount of contrast and keep the dashboard looking clean. Um typically, I like setting the border to be the same color as the background
[20:25] so that I don't have an additional color over here. So those are some of the simple colors we can set up. Selection color is interesting. Selection color essentially decides this color that you see over here or in those drop-downs over here as
[20:37] well. So let's go edit that to be maybe a little more orange. Next up are some of the spacing parameters. And if I want to go for
[20:51] a high padding, so as you can see like all your text visualizations moved inside the widget a little more. I personally prefer going for low margins and high padding. And if you're doing that, maybe set the corner radius to
[21:03] zero so you get these nice sharp edges. So it feels like a nice tight grid. You can also mess around with title alignment. So, some companies prefer to have it on the left. Some, based on your brand guidelines, it might be in the
[21:15] center. I personally prefer title alignment to be the left on the left. Next up are going to be some of the settings we have for our fonts. So, the way you have to think
[21:28] about this is you have a default font, which let's say at our company we have DM Sans. We can do that and then also change the case and the sizing around, but let's not do that just yet.
[21:42] Let's go into title and I want to update the size of these titles of each of our widgets. We can do that by changing the size and setting it to maybe uppercase to bring it a little get a little more punch into it. The one thing I want you to be
[21:55] mindful of when setting this to uppercase is that you want to keep the titles to be shorter because if they're super long, it kind of stresses the eye out a little bit. Um the other thing I want to highlight is
[22:08] how you should be thinking about fonts. I typically recommend not using more than two fonts in your dashboard and playing around with color and weight properties instead because A, every time you load a dashboard we have to load all
[22:21] those fonts and that might slow down the performance a little bit. Um but second, for consistency. Uh I want you to really be intentional about when you think about these things. So, example, for field values, I would
[22:33] recommend using a DM mono or any monospace font. The nice thing about monospace fonts is each of your characters are the same width. So, in your tables, all your numbers will line up perfectly and look really clean. So,
[22:47] let's do that. So, I updated this to DM Mono and my field title um to match it, I'm also going to update that to DM Mono. So, the how does this look? Let's go check it out. So, now Now can see you
[23:00] have DM Mono of installed on all the numbers here and here as well as the field titles here. And you can make this bold. So this is what we call the field title. Let's set this to uppercase.
[23:13] Great. And then maybe make this medium cuz why not? Great. So as you can see your your using typography is really your secret weapon. Set it to your brand fonts, play around
[23:25] with these settings to give your dashboard a unique identity. Um the next thing I'm going to talk about is access colors. So sometimes especially for if you have to present, you probably don't want access uh
[23:39] colors to be visible just to give your dashboard a clean look. So what I'm trying to do is get rid of these lines that I'm seeing over here. So I can simply set the access color. Actually, it'll be the grid color over here to be the same as this and voila, all
[23:51] your lines here have disappeared to get that clean, professional look. The next thing I want to talk about are colors and this is really important. If you don't want to spend too much time designing a color palette, I highly
[24:04] recommend using Color Brewer. This has been around for ages and is one of the gold standards for colors in data visualization. And the best way to use it is to first decide the number of colors you want to use
[24:16] for your palette. I recommend maybe having like seven to nine colors at least. Um and then deciding what the feature is what what's the purpose of your dashboard. Typically, you want a qualitative palette where each of the continuous
[24:28] color each of the adjacent colors are very different from each other to make sure that they look uh separate uh in your chart. And yeah, feel free to play around. You got sequential, diverging, uh
[24:41] qualitative. And here's the magic of using uh Genie code actually. So I'm going to copy these hex codes over here. Go back to my dashboard. Open up Genie code.
[24:53] Sorry if there are too many panels. We're zoomed in right now. And say, "Update cool color palette to this."
[25:06] And that's it. Let's give it a second. Uh Genie code will think and upload all of our hex codes into our theme over here. So, instead of having to go and copy and paste each of these hex codes, we can
[25:19] leverage Genie code to quickly update the entire palette. >> Yeah, this this was super annoying to do uh manually before. >> Did it do it though?
[25:32] Yeah. Well, you know, the demo effect. I promise you it did it right before we tried it. >>
[25:45] change our default gradient to something more interesting if you wish. And this gradient will be applied across your dashboard then as the default gradient. The nice thing about having these colors
[25:58] set up properly is when you're actually working with your dashboard, once you've set up your theme, you can do some really cool things. Like, for example, in this number over here, I would like to make it
[26:10] bold as well as apply a conditional style over here. So, if, let's say, my value is greater than target over here, let's make it green. Or red if you want.
[26:23] So, this allows you to provide more context to a user if a number is good or bad, for example. I'll take a pause here. Over to you, Ethan. >> Yeah, sounds good. Um So, the last thing we wanted to plug
[26:37] is um you might be wondering how dashboards get translated into mobile. Um we know kind of conventionally some other tools, you know, you might have had to create a whole separate layout for mobile to be available for that dashboard.
[26:49] Um we've actually taken the path of just generating for you the mobile version um as soon as your dashboard is published into Genie 1. Right, so what that means is if your dashboard is laid out in a grid, for example, we go row major,
[27:02] meaning we go from the top left corner, scan right, next row, scan to the right. And we place each of those widgets into a single file column that's automatically generated for you in the mobile app. Um so we'll show how that works in a second, but the kind of the
[27:14] key things to consider there are first of all, we don't kind of adjust for you the height of your widget. Um so whatever height it appears in your dashboard becomes that in the mobile app. So kind of use that as a guideline
[27:26] for kind of thinking through how you want to size your things. Uh but then also tables in the mobile app also scroll horizontally, right? So you might have a table that's pretty wide. How does that work in mobile? Well, you can scroll through and you basically have a
[27:38] sliding window that you can go through. Um so that also means putting the most critical important things to the left so that folks don't have to scroll all the way to the right, right? To to see that. Um and this is also an example where if you know something is going to be used a
[27:50] lot of mobile, uh counters with a single with sparklines that we talked about in the earlier session earlier in the session, uh tends to work better than kind of doing like very fine bar charts
[28:02] and line charts. Um so that that's another one to consider as well. Um there's actually a really nice way to test how mobile looks. Uh so Moheb will show you that. >> So once you have created your dashboard, you can publish it.
[28:18] And the way you can test this is simply by reducing the width of your screen. And as you can see, all the charts will update. I got to go smaller. And this is the mobile view. So you can see your pages move down here, all your global
[28:30] filters available here. And this is how you should expect your dashboard to show up in your mobile. >> Right. Oh. Cool. Okay, so just to wrap, um
[28:42] couple couple of like super practical things, right? So always use sparklines and conditional coloring where you can. You can ask Genie code to do this for you. You don't have to go through and click everything yourself. Um make sure always leverage size
[28:54] differences. Um so, some panels should be more important than others in a dashboard. If that's the case, right, make the more important one bigger. Play around with ratios. Um using a color palette generator, if
[29:06] you're not already, can really help you save a lot of time. And And again, I can attest that putting those of hex codes into Genie code does work like 95% of the time. So, um if and if not, you can always prompt again. Um testing layouts uh is helpful here um
[29:21] for mobile. Um and hopefully in the session you see that, you know, we've added probably 20, 30 knobs in the past year for different, you know, adjustability based on feedback from from folks in this room. Um I remember last year presenting at
[29:33] Summit some a lot of people actually asked me like, "Hey, are we going to get more like visual adjustable things?" Um hopefully you've seen that that has come. Um there's kind of two more things that we're working on that's helpful to know. Uh one is we've doubled the density of
[29:45] the grid this year, and we don't plan to stop at this density. So, we're also looking at offering like even more configuration in terms of like how many rows, how many and like exploring a more absolute layout uh approach um similar
[29:57] to kind of what you might have had in Tableau. Um and then the other thing is uh we're working this quarter on actually helping you map specific regions or specific field values to colors and then enforcing that consistently across your
[30:09] entire kind of set. Um so, for example, if you're graphing countries like Canada's always red, France is always blue, right? And no matter if those slices appear in bar charts or line charts or pie charts, right, you want to make sure that Canada's always red so
[30:22] you can just use color as reference. So, that's another thing we're building. Um so, hopefully that gives you kind of a set of kind of things to play with when you get home.

Learn more about the Databricks Data and AI platform.

The information provided herein is for general informational purposes only and may not reflect the most current product capabilities or configurations.