You said to draw the large screen, but do not know the data, and then you want to 🍾🈳🤏🏻 🈳, then I ask you to change to late night?

Hello everyone, here is EasyV digital Twin visual designer – Yuan Bao leg length 1.8m)

He fills in the blanks at 🙊, which he wrote about “data research” in recent days.

🍴 comes first

▶︎ Because I am a designer, this is an article about data research experience from the perspective of designers. If you are not professional enough, please comment and exchange

▶︎ easy-going and amiable 🙃, 🚫 serious

Unless separately stated, the illustrations are from personal production or EasyV platformClick the link to jump tool oh ~ free

1. Why do 🍱 do data research

The theory:

▶︎ Visualization = Plate loading, data = dish

▶︎ There are several dishes, including dry and soups, long and short, and if you don’t know what a dish looks like, you won’t be able to pack it well

▶︎ Avoid “creating something out of thin air” and “thinking 🐘” to do design, otherwise when receiving data, you will have to change many 😮💨, and customers will think you are not professional

▶︎ The client is satisfied with your “ideal effect drawing”, but due to data reasons, the landing effect is poor, and they will feel “cheated”. The emotion of raising first and suppressing later will affect the delivery

Good! Here are some vivid examples of potholes that aren’t designed based on real data 🌰🌰🌰🌰🌰🌰🌰🌰🌰 👇🏼 :

1, many dishes, but you prepared a small plate 💥

The space you reserved is full of data!

🌰 The most common is this 👇🏼

If we know the data before design, we can make specific adjustments:

▶︎ Adjust the text direction of the X-axis labels to avoid overlapping

▶︎ sort (When there are few columns, we can quickly compare the size with the naked eye, but when there are too many columns, it is not so easy to judge)

Then you get a slightly better visualization 👇🏼

🌰 does not truncate the data 👇🏼

Source: Project screenshot

🌰 Does not deal with the index name Mrs. Mrs. Too long 👇🏼

Photo source: taken at the project site

🌰 grows by more than 100%, but the maximum on the right axis is stuck at 100👇🏼

Source: Project screenshot

🌰 left too few digits for the flop to block the icon👇🏼 on the right

Source: Project screenshot

2, You prepare a big plate, but the dish is only a little 🥣

You think there’s a lot of data and it’s full, but it’s dry.

🌰 a recent trend chart that turns out to have only two years of data. If I had known this, I would not have used the line chart 👇🏼

🌰 do you think the data is the full bloom of the province 👇🏼

But after receiving the data, found very “bleak” 👇🏼

If we had known that this was the case, in terms of design, we could add a little halo to the northwest to make the map more balanced and more prominent 👇🏼

Of course, it’s possible to zoom in on the map to the Southeast, but it’s not conducive to getting a “big picture” view 👇🏼

3. You prepare the plate, but the dish is out 🤲🏻

When you finish your design, the client says you can’t get the data.

Reasons for the customer to decide to remove a metric, including but not limited to 👇🏼

▶︎ requires a cross-department count, but the person in that department will not give it

▶︎ This data shows that our business is very poor

▶︎ This data has a field in the database, but it has not been collected actually

▶︎ The calculation rules for this data are too complicated to be explained clearly when asked

▶︎ The points and routes on our map are drawn by hand without latitude and longitude data

▶ ︎

(Cross it out: I just don’t want it 🙃)

▶︎ …

As a designer, what can you do?

Therefore, before starting the design, try to determine the difficulty of taking numbers and data quality to reduce the probability of customers cutting data.

4, the dish is not good-looking, but you should be good-looking 🧚♀️

Translation: Data beautification.

🌰 For example, the data are so different that except for series 4, the trend of the other three categories is almost invisible: 👇🏼

In this case, we should not force a small gap (such as increasing the minimum value of the Y axis), which will cause “distortion” of data observation! But we can solve it through interaction 👇🏼

Because I often suffer this disadvantage, I now prefer the diagram component with the “click on legend to hide series” function to restore the design.

This pit, for the “unique” bar chart, can also be applied to 👇🏼

🌰 and this, revenues are rising, growth rates are falling, clients say it looks like performance is falling, I:?? 👇 🏼

First of all, I think it is a matter of reading the chart (reading comprehension) and not to conclude that “a decline is a poor performance” without setting aside the meaning of the indicator itself. The month-on-month growth rate fell, but it was still positive, so it was just “slowing down.”

🥚🦁️, it is understandable that customers do not want to be perceived by the audience as poor performance at first sight for the large visual screen of display class! Therefore, it is suggested that the growth rate in the case of decline, do not show the growth rate, directly replace the column data, enhance the feeling that the business is rising 👇🏼

If Party A wants to show the growth rate, it can customize the growth rate based on the previous picture and display it between the two pillars 👇🏼

🌰 Some negative data, in fact, less data represents good business, but customers often feel “space” 👇🏼

Faced with this perception feedback, we can offer the following three solutions:

▶︎ Empty what empty? That is good business, please respect the facts, so do not change!

(Need some courage with Leong)

▶︎ can be reversed into words like “overdue rate,” “yield rate” (😂, but there is no such antisense) and “pass rate,” so that the data is full when the business is good

▶︎ Change the graphic to a color with a warning for a smoother logic: less red = less bad = good business 👇🏼

We should see, we designers can do the “beautification”, not to dry “down to rise, less change, forced to pull small gap” this violation of the fact of the SAO operation.

But with respect to what the data really looks like:

▶︎ let the audience see all the data clearly

▶︎ Help customers better achieve their presentation goals

▶︎ Present data in a more logical way to promote understanding

5, other

🌰 Pictures (pictures are also a kind of data) are not all square in size and are deformed when filled into containers 👇🏼

Source: Screenshot of the project (the head picture in the picture is simulated data)

🌰 A large chain of 0️ on the Y axis is really 👇🏼

🌰 is the most annoying, most of the X-axis label length can be displayed in 2 lines, there is a 🙃 is extremely long, also did not cause occlusion, but it looks uncomfortable 👇🏼

⬆️ In this case (i.e. when the text length gap is large), it is recommended to go straight to the horizontal bar chart and start with 👇🏼

In short, in the case of insufficient understanding of data, the visualization system designed by us cannot well undertake and show the appearance of real data, which will appear very “lame” and “unprofessional”!

2. What aspects of 🔨 survey data

After demonstrating the importance of doing data research before design, I tried to summarize the following features of data that should be paid attention to in advance 👇🏼

1. Whether the customer can provide data

▶︎ Can you provide data for all these indicators to be shown?

(Provide data interface)

Check the details of the data interface with the data developer

Check whether data formats are standardized

▶︎ Can you provide the latitude and longitude data on the map

🌰 the longitude and latitude of companies across the country

🌰 is the longitude and latitude of waterways

Spending…

Because the customer cannot provide the longitude and latitude of the point, it is necessary to manually “create the point”.

2. Data meaning and text length

▶︎ What is the meaning of data

(Don’t finish the design, they ask you, you can’t say what the indicator means)

▶︎ Relationship between data

▶︎ Indicator name/Length of text data

At least a few words

The longest word

Check data relationships with colleagues

3. Quantity and variety of data

▶︎ How many data are there in total

🌰 Bar chart of revenue by district: There are 28 districts and counties

🌰 Node health status: There are more than 120 nodes

The map at 🌰 shows that there are only three branches

Spending…

▶︎ How many categories are there

(We need to use the number of preset colors/styles)

🌰 : There are 15 types of goods

(For too many categories, the general operation is: take the first 5, then classified as “other”, we preset 6 colors)

The 🌰 alert has four levels: moderate, minor, severe and very severe

Spending…

You prestyle all the states in the component and then match them against the state values

4. Numerical range

▶︎ is usually a number of digits

▶︎ What is the maximum value

Is 🌰 more than 100 percent

▶︎ What is the minimum value

Is 🌰 at least negative

▶︎ Specifies whether the password can be empty

How to deal with empty space: Hide or placeholders

▶︎ Whether the interval is set by someone

🌰>80 percent is on alert

Spending…

How to set the maximum/minimum value of the Y axis in Keynote: If you leave this parameter blank, set it to “Auto round” mode

5. Data format

▶︎ What is the unit

🌰 The yuan is still the dollar

🌰 or won

Spending…

▶︎ Decimal or percentage

▶︎ Save a few decimal places

Whether to keep a decimal number ending in 0

▶︎ Data format with latitude and longitude

Based on what coordinate system

The format of dot, line, and plane data

Spending…

Check data units with colleagues

6, other

▶︎ Monitor video streams

What is the scale of the picture

What is the format of video streaming

▶︎ Image data

Source: Static or URL

Size ratio

▶︎ Data style specification

🌰 Is red increasing green decreasing or green increasing Red decreasing

The color of 🌰 should be blue instead of green

Spending…

▶︎ there are many more bits and pieces, but I am too old to remember 🕳

3. Data survey in 📍 real work

In real work, we are often unable to obtain sufficient data research support due to various conditions, including but not limited to:

▶︎ No data analyst to help you

▶︎ Your own research is too passive

▶︎ The contact person of Party A also has to ask different people, and the research cycle is too long

▶︎ …

Under such conditions, we have to design first, and many problems can only be found after connecting with real data.

Read a lot of articles, still not good designer’s life

🥚🦁️, keep your mind on it. My experience tells me that when you start to pay attention to the data, you will automatically predict the possible problems mentioned before, and then do some preparatory processing in advance. When it happens, you will not panic (but still will be annoyed), and you will have a kind of “I knew it would happen!” The sense of stability hahaha.

Four, 🙊 finally with three words, want to let everyone give me 18 thumbs up

▶︎ Most of the time, we are not “designing data”, but “designing containers for data” 🥣

▶︎ Visual design based on real data will make it easier to feel the significance and value of data and more rewarding ➕

▶︎ I wish you less pit, less map, less overtime ❤️

Whisper BB: The Nuggets text editor is really good!