Fast abstract ↬

Dribbbleshots simply could be the hotbed of questionable dashboards. Putting visuals, little context, and no analysis: all recipes for mediocrity. Mediocrity gained’t do. We’ll pursue greatness. And in that pursuit, we’ll cowl analysis, decluttering, and knowledge visualization.

Even when it’s a blessing in disguise, discarding parts of your work is not any enjoyable. Tossing out suboptimal elements of our design is usually a daunting process, particularly after you’ve invested hours of labor into it. However make no mistake, it is a bias most designers are liable to. We are able to get too connected to issues we’ve created, regardless of them not offering any actual worth to our customers. Right here lies the distinction between okay and nice dashboard design. The previous is pretty straightforward to attain. The latter isn’t.

I’ve compiled a couple of issues that I’ve realized all through my profession with regard to dashboard design. On this article, we’ll speak about analysis, decluttering, and knowledge visualization, in addition to how these items could make your dashboard design higher.


A dashboard is part of an software that shows world details about the app’s utilization or some other exterior knowledge. They arrive in numerous sorts, from comparatively simplistic one-layered dashboards like this one:

A simple one-layered dashboard

A easy one-layered dashboard (Picture supply: Dribbble) (Giant preview)

Whereas others are extra advanced and multilayered ones just like the one under:

A complex multi-layered dashboard

A fancy multi-layered dashboard (Picture supply: Dribbble) (Giant preview)

The suggestions we are going to present on this article apply to a variety of dashboards, no matter their complexity.

Earlier than diving into our arguments, first, let’s kind out what I imply by “mediocre” and “nice”. These phrases indicate a worth system that isn’t precisely typical, so I owe you an evidence.

Mediocre vs Nice Dashboard Design

The distinction between “mediocre” and “nice”, I consider, lies within the course of. There’s additionally a continuum concerned between these two classes. Within the context of this text, the axes of this continuum are as follows:

  • Sourcing inspiration (copying),
  • Analysis (referred to within the subsequent part as “homework”),
  • Design Validation (substantiating design selections with knowledge),
  • Knowledge Visualization,
  • Shade Palette.
Sourcing inspiration
AnalysisRestricted and fragmentedConstant and purposeful
Design ValidationRestricted to inside stakeholdersInclusive of each inside and end-users
Knowledge VisualizationClosely targeted on aesthetics and tendenciesKnowledge viz options are according to its function; an excellent stability between effectivity and complexity;
Shade paletteArbitrary / trend-basedUtilizable of shade connotations, and alignment with model values

(All through this text, all references to mediocrity are primarily based on these distinctions)

Now that we’ve sorted out the definitions, let’s rapidly give an outline of my arguments. On this article, I recommend that, versus mediocre dashboards, nice ones require:

  • Analysis,
  • A wholesome dose of cluttering & decluttering,
  • A thought-through shade palette.

First, let’s break down analysis, or “homework” as I prefer to name it. How do you even do “homework” as a designer? Is it okay to repeat another person’s work so long as you don’t make it apparent? Or do you begin from scratch each time? — Let’s assume this via.

Extra after bounce! Proceed studying under ↓

Doing Your Personal Homework VS Copying Your Friends

Certain, there’s all the time the choice of copying and tweaking what your rivals already provide. Or, if there are not any shut equivalents, you may simply “Frankenstein” the weather collectively from comparable dashboards. This certainly sounds simpler than “reinventing the wheel”. “*Good artists copy, nice artists steal,*” as Picasso’s quote goes — however, sadly, this strategy is prone to doom your design to mediocrity. Let me elaborate.

A fast disclaimer: if what you’re searching for is to be taught, then by all means copy nice designs and attempt to be taught from them. Nevertheless, in the event you’re engaged on industrial / “actual” tasks, then copying alone doesn’t do.

To begin with, theoretically talking, you would strike gold by copying nice dashboards. There’s a catch, nonetheless. What’s the chance that your supply of inspiration is well-researched, sensible, and most significantly relevant to your specific scenario? Are the dashboards you’ve discovered price stealing? Who is aware of. When you ask me, nonetheless, I wouldn’t guess on it. This ties in effectively with the concepts Austin Kleon outlined in his guide Steal like an artist.

A diagram: is it worth stealing? If yes, then steal it and move on. If not, don’t steal it and move on.

(Picture supply: Amazon) (Giant preview)

Secondly, whenever you’re sourcing inspiration, all you see is the top product: a colourful nice-looking dashboard. Hardly ever can you discover inspiration accompanied by an intensive evaluation of the method, analysis, and decision-making concerned. This results in one thing akin to a cargo cult, i.e. replicating the patterns you see with out understanding why.

A plane out of straw positioned on the landscape of a dashboard as a background

Don’t be part of the ranks of Dashboard cargo cultists. 😄 (Dashboard credit score: Supply: Dribbble, BigAlmanac) (Giant preview)

Lastly, one other unlucky aspect impact of merely copying is that it’s not constant. You possibly can’t play the chances and win each time. That’s to not say {that a} “correct” design course of is foolproof. I consider that following the ideas I define on this article will make your designs persistently higher, so you may “strike gold” with a better diploma of reliability and replicability.

Simply put, following a proper process and having a command of design principles, as opposed to mindlessly copying, gives you a better chance of coming up with usable designs with fewer iterations and useability testing sessions.

So how do you strike a stability between “copying” and “doing your individual homework”? Right here’s what I believe. On high of “stealing” (ahem.. searching for inspiration), additionally discuss to your customers.

I 👏 can’t 👏 stress 👏 this 👏sufficient.

I guess we’re all (myself included) bored with listening to this mantra. “Speaking to customers” is like exercising or maintaining a healthy diet — everybody is aware of they need to do extra of it, however few really do it.

“Folks ignore design that ignores folks”.

— Frank Chimero

Is there a worse strategy to ignore customers than excluding them from the dialog altogether?

Alright, speaking to customers is essential. That a lot is obvious. What isn’t self-explanatory although is how precisely this communication ought to happen — let’s go over a few actions to look at a few methods.

Consumer interviews:

Listed here are a couple of questions price trying into when designing a dashboard.

  • What data do customers want probably the most?
  • What’s the function of this dashboard?
  • What do customers seek the advice of this dashboard for?
  • How do they go about searching for this data at the moment?

On this piece, Stelian Suboti, a UI & UX designer with greater than 7 years of expertise, the writer claims that:

Truly valuable insight can come out of a short user research phase with just five users — and it will save an enormous amount of time down the line.

Doing analysis (consumer interviews on this case) is step one of the design pondering course of (empathy), in addition to the UX course of on the whole.

research of the user interview

(Picture supply: Adam Fard Studio) (Giant preview)


“Card sorting by its very nature, is a technique you’d use whenever you wish to uncover classes, teams, or interrelationships. Use this methodology whenever you wish to understand how customers reply to visible cues and also you wish to seize the similarities and variations.”

— Assume Design

As we’ve simply realized, card sorting helps us perceive how customers group and categorize data. In sensible phrases, card sorting appears one thing like this 👇

card-sorting with different color stickers

(Picture supply: Medium) (Giant preview)

You typically wish to break up your dashboard into “atoms”, bits of data of the smallest measurement, in order that the members can set up the relationships amongst these parts. That is how you would construct the knowledge hierarchy throughout the dashboard that’s according to customers’ psychological fashions.

Usability Testing

Usability testing is arguably a very powerful and insightful analysis exercise you may conduct with regard to dashboard design. Enjoying by ear and skipping this step altogether simply could be the largest cardinal sin of UX design.

Right here’s what Nielsen Norman Group, the world’s main product design authority, has to say about usability testing:

“Even one of the best UX designers can’t design an ideal — and even adequate — consumer expertise with out iterative design pushed by observations of actual customers and of their interactions with the design.”

In sensible phrases, you’d wish to conduct usability testing classes after card sorting and consumer interviews. The latter two actions blended with your individual assumptions ought to end in early dashboard variations. These tough concepts will then should be examined and iterated on from sketches, and low-fi wireframes all the best way to a high-fidelity prototype.

Now that we mentioned the analysis strategies, we are able to safely transfer on to my second argument, which is all about decluttering. How are you aware in case your dashboard is cluttered? Does the “much less is extra” precept work each time? — These are the matters we’re tackling subsequent.

To Declutter Or Not To Declutter

In a traditional sense, decluttering is almost synonymous with simplifying. Furthermore, the observe of “maintain it easy silly” (KISS) has change into considerably of a design truism. The less complicated the higher, proper? I want it was that straightforward (no pun meant).

A cluttered design, by definition, overwhelms its customers. That’s to not say that an app can’t have a studying curve. Adobe merchandise are in all probability a textbook definition of what “overwhelming a first-time consumer” means.

Let’s check out the next dashboard:

An ostensibly complex dashboard with financial information

(Giant preview)

This is a wonderful instance of what could be thought-about “cluttered”. A more in-depth have a look at the dashboard will let you know that it has to do with inventory (index) worth motion, buying and selling quantity, financial occasions, and so forth. When you assume that this appears intimidating — you’re proper. For a median consumer, it in all probability is. I’d wager, although, that the folks certified to make use of this sort of software program really profit from this complexity — it permits them to work effectively. Obscuring a lot of the knowledge offered on this dashboard will solely end in pointless clicks and extreme friction.

Now let’s attempt to check out one thing seemingly decluttered and “clear”. By the best way, this specific dashboard, designed by Bhojendra Rauniyar, gained within the minimal dashboard problem by UpLabs. Congrats to Bhojendra!

An ostensibly simple but poorly usable dashboard

(Picture supply: UpLabs) (Giant preview)

Earlier than I begin critiquing the dashboard above, right here’s a disclaimer. To begin with, it is a dribbble shot, so we are able to’t be sure whether or not it’s an remoted piece of design or part of a product. Secondly, I don’t have entry to the designer to ask for his rationale behind sure design selections. Thirdly, I have no idea whether or not this design has been examined and the complete context of its utilization. As such, all I’ve left is to take a position and assume. I hope, for academic functions, that may do.

Let’s begin with the chart. It compares shopper scores with earnings. I’m left to marvel although what values outline these two classes? The values need to be the identical since each curves are on the identical graph. So will we measure earnings in common score or score in {dollars}? Moreover, there are not any labels for the dimensions, so I can’t even make certain of the models of measurement. What would in all probability be extra helpful is having separate graphs for key metrics with clear labels, i.e. to introduce “additional complexity”.

Moreover, the stats alone usually are not very helpful. Is incomes “5k” in “1.8k” hours an excellent consequence? How does that evaluate to my friends? What are my dynamics?

That’s to not say that the dashboard I discussed first is ideal and the second is unhealthy. You might argue that each of those require additional work. My level is that decluttering for the sake of decluttering is a poor design maxim.

Cluttered apps are unnecessarily overwhelming and laborious to navigate. How are you aware if an app is all the above? Presto! That’s proper, you discuss to the customers and have usability testing classes with them. Apart from usability testings, there are a couple of different practices that may assist be sure that your interface isn’t cluttered:

  • Clear data hierarchy;
  • Using modals or panels;
  • On-hover interactions;
  • Two to a few colours;
  • White house, loads of it.

Now that we went via analysis and cluttering, we are able to transfer on to knowledge visualization. Knowledge visualization is irreplaceable for a lot of causes. It permits you to:

  • See patterns,
  • Examine knowledge,
  • Articulate the knowledge visually,
  • Monitor knowledge dynamics.

Does this ring a bell? Precisely — that is just about all the pieces {that a} dashboard ought to do. It’s no coincidence that the majority dashboards characteristic graphs and charts. Due to this fact, being competent at knowledge visualization instantly interprets into being a reliable dashboard designer. Let’s zoom in on that.

Step Up Your Knowledge Visualization Recreation

There are a number of kinds of charts obtainable, thus we have to select properly. The sunburst chart does look superior, however is it as clear and clear because the pie chart? Maybe. That actually does rely in your customers. Forms of charts apart, there are additionally colours, their semantics and so many different issues that add to the complexity of knowledge visualization in dashboard design.

An infographic with 30 different kinds of charts

(Picture supply: The Unstated Pitch) (Giant preview)

On this part, I’ll do my finest to stipulate a few of the knowledge visualization finest practices and their software in dashboard design. Let’s begin with colours.

This part is essentially primarily based on the analysis of Claus O. Wilke, Professor in Molecular Evolution at The College of Texas at Austin, and the writer of Fundamentals of Knowledge Visualization.

Shade In Dashboard Design

Let’s deal with the problem with numbers first. What number of colours must you use in UI design and dashboard design particularly? — Frankly, there’s no magic variety of colours that works it doesn’t matter what. Nevertheless, if I have been to suggest you a quantity, it could be 5. That doesn’t embrace shades if the depth of the colour represents a worth or semantic colours in the event you want them (e.g. crimson for error messages or inexperienced for fulfillment messages).

Why 5? Effectively, there’s a 6:3:1 “golden” rule of visible design. These numbers characterize the proportion amongst three model colours: fundamental, secondary, and accent. Nevertheless, it is best to keep in mind that on high of those three colours, you’re probably going to wish some variation of white and black for textual content and its background.

An example of three color palettes

(Picture credit score: QED42) (Giant preview)

One other factor to make use of colours for is to tell apart amongst completely different entities. Listed here are a couple of widespread shade palettes.

An example of three monochromatic pallets

(Picture supply: Oreiily) (Giant preview)

Use monochromatic shade palettes if exhibiting data inside a single class. That manner you may introduce one other dimension to your chart that’s straightforward to learn. For instance, the extra intense the colours, the upper the worth and vice versa.

An example of three monochromatic pallets

(Picture supply: Claus Wilke) (Giant preview)

Right here’s an instance:

A map of Texas indicating the county income

(Picture supply: Wilke Lab) (Giant preview)

What you can even do is decide two colours to characterize the other ends of a spectrum. That manner you may simply inform aside values that belong to completely different extremes. Right here’s an instance:

The map of the United States indicating the change in average age

(Picture supply: Dallas Information) (Giant preview)

Listed here are a couple of shade palettes for this strategy you may select:

An example of diverging color pallets

(Picture supply: Higher Figures) (Giant preview)

Beware Of Colours Semantics

Now that we went via the approaches you may take with utilizing colours, it must also be famous that colours elicit unconscious reactions. These reactions are also known as “shade semantics”.

When evaluating 2 teams, utilizing colours like crimson and inexperienced (due to the semantics related to these colours) would possibly result in misrepresentation. Purple, for example, is usually related to hazard, failure, and poor efficiency.

Check out the graphs under, which illustrate the maths efficiency of two courses. We would robotically assume that college students on the graph on the left in Class A are performing poorly, as a consequence of the truth that the values are represented with crimson. Purple, particularly together with inexperienced, usually elicits such an interpretation.

Two scatter plot charts are compared side by side. The only difference between the charts is the colors used. The left chart uses blue and green which bear no connotation, while the one on the right uses red and green that imply a connotation.

(Picture supply: Adam Fard Studio) (Giant preview)

Nevertheless, there are instances the place this semantic distinction can be utilized to our benefit like within the instance under, i.e. utilizing crimson to show efficiency that’s under common, and inexperienced for above. Additionally, notice that it’s extra pure for folks to instinctively affiliate values on high as better than these under. The graph under illustrates the final level.

An example of two bar charts. The first chart doesn’t utilize the color semantics, while the other one does.

(Picture supply: Adam Fard Studio) (Giant preview)

As you may see, inexperienced represents college students who carry out above common, whereas crimson is reserved for individuals who underperform. Semantic shade cues make this chart much more readable than it could in any other case be with one other shade scheme.

On Which Charts To Use

Knowledge visualization can induce a good quantity of friction when performed incorrectly. Be sure that to decide on probably the most appropriate sort of chart to ship the correct of knowledge. Beneath, you’ll discover a kind of complete listing of charts you should use to characterize the knowledge featured in your dashboard, relying on the roles they fulfill.

NB: Most charts serve a couple of function. As an illustration, a pie chart is used to check values, present composition, and knowledge distribution. As such, I’ll point out these charts a number of instances relying on the use case. Nevertheless, I’ll solely give a quick rationalization for every chart as soon as.


  • Pie
    Sometimes used to characterize fractions of an entire. Pie charts work nice at any time when you should evaluate a comparatively low variety of segments which might be comparable by way of their measurement. Conversely, having too many segments or tiny segments makes the chart too laborious to learn. A serious con of pie charts is their lack of ability to point out adjustments over time.

An example of a pie chart

(Picture supply: exceljet.web) (Giant preview)
  • Stacked Bar
    Particularly helpful when evaluating classes. Identical to the pie charts, the extra advanced they get (extra sequence & classes), the more durable they’re to learn. A bonus that stacked bar charts have over pie charts is the power to see adjustments over time. One other factor to look out for is that when you’ve sequence that adjust of their worth, it will get more durable to check them visually. For instance, check out the instance under, it’s not apparent whether or not the orange sequence for “Eyebrow pencil” is bigger or smaller than that of “Basis”.

An example of a stacked bar chart

(Picture supply: (Giant preview)
  • Mekko
    Good for visualizing the variations of sure classes inside a number of dimensions. Identical to the remainder of the charts we’ve gone over up to now, they’re straightforward to learn however get more and more more durable to understand as soon as the classes and dimensions begin to pile up. Moreover, it is best to keep away from utilizing Mekko charts if the variations amongst parts are too drastic.

An example of a Mekko chart

(Picture supply: (Giant preview)
  • Stacked Column
    A good way to point out comparisons between classes. You might need observed that this chart sort bears a robust resemblance to a stacked bar chart. Although there are variations in naming, it is best to know that horizontal bars often go from highest to lowest worth (or vice versa), whereas vertical ones indicate one other order rationale.

An example of a Stacked Bar chart

(Picture supply: amcharts) (Giant preview)
  • Space
    A extremely widespread manner of representing quantitative knowledge. Usually used to point out the rise or lower of assorted knowledge sequence over time. The foremost con of such a visualization is that it’s laborious to attach a particular level of the graph to a worth on each axes. And, as all the time, the extra data you cram in, the much less readable it turns into. Lastly, if the graphs overlap, then one space will go over one other and vice versa. Due to this fact, space charts are good for serving to see the large image with out an excessive amount of regard for being extraordinarily exact.
    You may additionally discover a similarity between space and line charts. The previous ought to be used when there’s an emphasis on a part-to-whole relationship.

An example of an Area Chart

(Picture supply: (Giant preview)
  • Waterfall
    This type of chart is usually used within the monetary trade to show the motion of worth and its incremental path in direction of an endpoint. These charts, not being all too widespread, may not be probably the most intuitive to learn, so they need to higher be reserved for professionals who cope with waterfall charts usually.

An example of a Waterfall Chart

(Picture supply: Fusioncharts) (Giant preview)
  • Line
    A good selection whenever you’re trying to current a sequence of values. These values (additionally referred to as markers) are usually related by straight line segments. The road chart mimics the disadvantages of the realm chart, apart from the colour overlapping drawback.

An example of a Line Chart

(Picture supply: Excel straightforward) (Giant preview)

Showcase the composition:

  • Pie,
  • Stacked Bar,
  • Mekko,
  • Stacked Column,
  • Waterfall.

Distribution of Knowledge:

  • Scatter Plot
    A good way to emphasise the relationships between one or a number of numeric variables and/or their distribution throughout two axes. The foremost drawback with utilizing scatter plots for UX functions, it’s that such a chart is significantly more durable to learn than all the different ones we’ve mentioned up to now. As such, scatter plots ought to be reserved for knowledgeable customers solely.

An example of a Scatter Plot

(Picture supply: (Giant preview)
  • Mekko
  • Column
  • Bubble
    Generally used to current monetary knowledge. Bubble charts are just like scatter plots, nonetheless, they supply a extra in-depth understanding of the information, since they’ve 3 axes.

An example of a Bubble Chart

(Giant preview)

Correlation and relationship between values:

  • Scatter Plot,
  • Bubble,
  • Line.

If we have been to summarize the charts above in a matrix, it could look one thing like this.

An infographic with the types of charts. For the sake of comparison, you can use pie, mekko, area, waterfall, line, bar, and column charts. To compare, you can use pie, mekko, waterfall, bar, column charts. To show distribution pie, mekko, waterfall, scatter plot, bar, column, bubble. To show relationship, scatter plot, line, bubble. To show change, you can use area, waterfall, line, and column.

(Giant preview)

It’s only becoming that we summarized the kinds of charts and their utilization as an infographic. 😄

Be suggested although, that I’ve indicated what these charts are typically used for. Theoretically, in the event you wished to, you would do a bubble chart, the place every bubble is a pie chart. Voila, your bubble chart can now additionally present the composition of every bubble. In fact, it’s simpler mentioned than performed, as a result of often, bubbles differ of their measurement drastically. Anyhow, you get the thought: the matrix is a basic guideline, however you may get extra inventive with charts if you wish to at your individual discretion.

Balancing Between The Complexity And Effectivity

So right here’s a factor. As you might need observed, a lot of charts can be utilized for comparable functions. That doesn’t imply, nonetheless, that these charts are equally good for a sure process.

Usually, when you slim down your selection to a couple charts that appear to work finest, you then wish to ensure that they’re the least advanced they could possibly be. Beneath, I’ve outlined a few of the charts we’ve talked about by the diploma of their complexity and effectivity.

An infographic outlining the efficiency and complexity of the charts is mentioned in this article. The more efficient and less complex a chart is, the more effective it is.

This ‘chart for charts’ (very meta, I do know) ought to be taken with a grain of salt. The effectivity axis implies that you just’re utilizing the chart correctly. The complexity axis is a illustration of my expertise. (Giant preview)

In a nutshell, you wish to maximize effectivity and reduce complexity. That is the top-left quadrant. These charts are straightforward sufficient to be understood by center faculty college students. You must keep on with these for the “shopper” persona, i.e. a median Joe. The extra you progress towards the appropriate, the extra causes it is best to need to consider your customers can be comfy with studying them.

What’s Subsequent?

In case you’d prefer to proceed studying about dashboards and the facets of their design, listed here are some sources my crew and I’ve discovered useful (to an rising diploma of complexity):



Scientific papers

Smashing Editorial
(ah, vf, il, yk)

#Analysis #Decluttering #Knowledge #Smashing #Journal

Leave a Reply

Your email address will not be published. Required fields are marked *