| 00:02 | Professor Guan yin chang from computer science , give us a guest lecture on |
|
| 00:09 | and why that is important for researchers us. His resource area is |
|
| 00:16 | So um in this presentation we will only get to learn about visualization, |
|
| 00:20 | if you have any questions about PhD towards the end, I think there |
|
| 00:25 | be a few minutes to ask these as well. Professor Groningen will not |
|
| 00:30 | staying for the full An hour and half hour, 20. So we |
|
| 00:34 | to wrap up from 1:50 or So with that let's uh welcome Professor |
|
| 00:40 | . Thank you. All right, you. Thank you for the |
|
| 00:45 | I'm glad to be here today to you a few things about visualizations. |
|
| 00:51 | insurance, I try to condense four , spreads around four lectures of my |
|
| 00:57 | into this one consigns presentation so hopefully get it done properly. So I |
|
| 01:06 | try my best to introduce what is , what is important and then I'll |
|
| 01:11 | about some connective the properties of our perceptions that will be important to know |
|
| 01:19 | we work on visualization and a few that you may feel handy when you |
|
| 01:25 | to create effective flaws and trust for tasks and if I have time, |
|
| 01:32 | will briefly talk about how to properly colors in visualization. So for |
|
| 01:38 | so please bear with me if the are slightly less organized because you know |
|
| 01:45 | try my best to put them into . Mhm. But so we start |
|
| 01:50 | the question why visualization is important and give you a simple a couple of |
|
| 01:58 | . Right. Many times we do is try to convey a message story |
|
| 02:05 | my useful information from data. Here's one example I have. This |
|
| 02:10 | data block contains some integer numbers. can imagine this could be some temperature |
|
| 02:17 | in space. Right. If I you, can you see any patterns |
|
| 02:22 | this representation? Would you be able tell me that anyone may be? |
|
| 02:39 | right down corner is a little bit . Alright. So yes you can |
|
| 02:44 | front left to right and actually upper to lower right. The values increase |
|
| 02:52 | . Right. But in order to out these patterns, what you actually |
|
| 02:57 | , what did you actually do in brain? You march through those numbers |
|
| 03:02 | x one and compel them in order figure out those values those trends. |
|
| 03:07 | doable. But if I give you huge data block with billions of |
|
| 03:14 | do you think you still can do in a finite amount of time? |
|
| 03:18 | not. Right. Okay. Let's it in a different way rather than |
|
| 03:23 | just compare the numbers. How about do some change. Alright. I |
|
| 03:28 | colors to the individual cells. Well values are measured and then I create |
|
| 03:37 | new representation like this. Do you the partners immediately? I guess the |
|
| 03:43 | will be yes. Right. You the patterns immediately. But you don't |
|
| 03:48 | know where it's low wages, But you see there's something and you |
|
| 03:53 | even need to read the actual values those numbers. Right. Okay. |
|
| 03:58 | that's one example demonstrated how effective visualization convey information than the other channels. |
|
| 04:08 | . Here is another example, I have this table with some information |
|
| 04:12 | the percentage of fats within a few of participants. Alright. My first |
|
| 04:20 | to you is how many groups are studying? Can you tell from the |
|
| 04:38 | ? Okay. Mhm. Mhm. know I mean to female, any |
|
| 04:50 | any other answers? Mm hmm. . Or eight. Alright. So |
|
| 04:59 | see how confusing this table for that questions. Right? There were actually |
|
| 05:04 | groups, two groups of males with income, Two groups of females with |
|
| 05:10 | income. And then we try to how the percentage of fat of those |
|
| 05:15 | of people change over time. So first measured them when they're younger, |
|
| 05:21 | they were younger, right? 65 or younger. And then after several |
|
| 05:26 | , measure that again. So that's this. 100%. Right now, |
|
| 05:30 | question will move to, Do you any outliers among these four groups or |
|
| 05:39 | actually change the same over the years these four groups of people? I |
|
| 06:00 | . No, for example, no . I mean like if I tell |
|
| 06:08 | there's one group that behaves differently than other three, which is true. |
|
| 06:16 | that be Female, Right? So is the same group for each |
|
| 06:31 | It's just We measure them in different of their lives. So you're |
|
| 06:39 | Alright, So this group behave So basically all the other groups of |
|
| 06:44 | of fat drop over the time, this particular group of people the percentage |
|
| 06:51 | fat increase over the year. Again, like the previous example, |
|
| 06:56 | can get this information by just spend time to read those numbers and |
|
| 07:02 | But is it effective enough when you try to tell the story to your |
|
| 07:06 | ? Probably not. But if I into visual representation like this. |
|
| 07:13 | so, first the number of it's very clear Before four lines. |
|
| 07:18 | . And the chain of that particular that is different from the other |
|
| 07:22 | It's also easily perceivable. All Rather than just going over the |
|
| 07:27 | you can just easily get this information the chat. All right. All |
|
| 07:32 | . So, these two example demonstrates important visualization can help convey information |
|
| 07:39 | Of course visualization can do more things we'll see. Right, this is |
|
| 07:43 | famous saying. Alright, visualization is about external connection. What is the |
|
| 07:53 | ? All right, let me finish . I'll go back to this |
|
| 07:55 | Visualization is really about external connection. is how resources outside the mind can |
|
| 08:03 | used to boost the connective capabilities of mind. Alright, I have also |
|
| 08:13 | same. So what visualization is actually is try to review changes and patterns |
|
| 08:19 | in the raw data made invisible we have different things around us. |
|
| 08:26 | can feel it, we know they're but we cannot see it without seeing |
|
| 08:30 | . It's really hard to understand. . Visualization can also make Abject concepts |
|
| 08:35 | intuitive to understand like some mathematical concepts limits. What does limit means |
|
| 08:42 | Oh sorry, no chance um means and probably tensions of a curve or |
|
| 08:49 | . What does that mean? without a visual presentation. Those concepts |
|
| 08:53 | not be easy to understand but let go back to this. What does |
|
| 08:58 | mean? Is that it's really an connection. How resources outside the mind |
|
| 09:05 | be used to boost the connected capabilities the mind. So what connection |
|
| 09:11 | what does it do anybody? So means that we try to understand |
|
| 09:27 | We try to learn something this connection . All right, okay, so |
|
| 09:33 | means that visualization is an external boost help us understand things more effectively. |
|
| 09:42 | , and here is another way to what it's visualization. It corresponds to |
|
| 09:50 | means to enable a user insights into data while visual representation right inside into |
|
| 09:59 | data can mean the understanding of the behind the data might useful information or |
|
| 10:06 | things. Okay, Alright, so what visualization is trying to do. |
|
| 10:13 | goal is trying to boost our conviction so we can see things that previously |
|
| 10:18 | couldn't. All right. That said saying probably should emphasize more about visual |
|
| 10:27 | resources there are different resources that can us boost our connected process. Speaking |
|
| 10:32 | that another way to boost our condition . So allows us to get into |
|
| 10:37 | data to know more about the data data mining, right? That's another |
|
| 10:44 | of techniques that allows us to get the data and help us understand the |
|
| 10:49 | . Data mining. Right. Can tell me the difference between data mining |
|
| 10:55 | visualization? You can get some hints the two logos I put here I |
|
| 11:01 | a machine on the data mining site a human on the visualization site. |
|
| 11:13 | I guess visualization makes the data more to humans with the mining we are |
|
| 11:23 | large amounts of data that the computer easily process but it's harder for us |
|
| 11:28 | make sense out of it. Thank you very much. Alright so |
|
| 11:33 | mining mostly use machine power to process large amount of data that's beyond our |
|
| 11:39 | to process. Remember the first example show you if I have billions of |
|
| 11:43 | points, how can you ask human process it effectively. Alright, machine |
|
| 11:49 | do it quickly paralleling right. Data is about automatic algorithm to help us |
|
| 11:56 | useful information from the data. While contrast visualization actually utilize human expertise and |
|
| 12:05 | effective representation of the data to help made decision. This decision can be |
|
| 12:11 | simply extract patterns or change or it be bigger like Okay, based on |
|
| 12:16 | I know and what the data showed and make a critical decision. |
|
| 12:21 | Is it that for instance in the diagnosis, is it a humorous or |
|
| 12:26 | ? Right? This you cannot completely on machine, right? You don't |
|
| 12:30 | to put your life on the You want experts, Okay to be |
|
| 12:34 | your side. Right. So these different sets of techniques for camps are |
|
| 12:40 | competing each other. They are actually each other like that. Students |
|
| 12:44 | data mining can help us process large of data where we cannot, but |
|
| 12:49 | extracted information may not be easily So how can we actually see that |
|
| 12:55 | understand it? We use visualization to the information extracted from data mining. |
|
| 13:02 | . On the other hand, visualization need data mining to pre process some |
|
| 13:07 | . So we don't show anything. don't show anything. Okay. We |
|
| 13:11 | show those important things to help experts narrow down what should be focused |
|
| 13:18 | Right? So these two are not each other. They complement each |
|
| 13:23 | Alright. We're important. Alright, go to hear what visualize it can |
|
| 13:29 | do. There was three major functionality tasks that visualization can help us |
|
| 13:37 | It can help us present information or story in the most effective and intuitive |
|
| 13:43 | to the targeting audience, but that's when we actually write our reports or |
|
| 13:50 | a presentation to our boss or We always include charts and graphs that |
|
| 13:56 | easily understandable, not just tables or . Right. Present information in an |
|
| 14:01 | eight second, analyze data to verify falsify hypothesis. Right? This is |
|
| 14:09 | . This task can also be done data mining techniques depending on the |
|
| 14:14 | but visualization can also help with Right? If your verification and falsification |
|
| 14:21 | purely based on some geometric setting in right then this will be easy to |
|
| 14:29 | and process based on the knowledge of experts. Okay, analyze the data |
|
| 14:35 | the biggest task that visualization can help is to explore the data where we |
|
| 14:41 | not know well to look for useful . Okay. We do not know |
|
| 14:46 | is in the data, what can useful or what cannot. So, |
|
| 14:50 | typically is the first choice you probably try to discuss, govern some useful |
|
| 14:57 | or change. So you probably form initial hypothesis. So this go backward |
|
| 15:03 | before that you probably don't even have hypothesis. Now with visual representation, |
|
| 15:08 | may have some hypothesis to say, there's some interesting patterns. Maybe |
|
| 15:12 | Alright. These patterns may be similar something that I have seen and be |
|
| 15:19 | using some established machine learning or data techniques and then I can try some |
|
| 15:25 | mining and machine learning techniques. They express those parts. All right. |
|
| 15:28 | then I verify that. Alright, explore something that's not known to discover |
|
| 15:35 | knowledge and findings. Right. Those the three main tasks that visualization can |
|
| 15:40 | us. Right? So those are basic stuff about visualization. Why visualization |
|
| 15:45 | important because it allows us to get the data effectively using our knowledge and |
|
| 15:52 | the effectiveness of our visual perception channel we'll see next. Right? What |
|
| 15:58 | can do? It can help us , can help us analyze and can |
|
| 16:02 | us explore. Right? So those the basics about visualization. Any |
|
| 16:12 | No. All right now let's move the second topic. The connection perception |
|
| 16:18 | things about ourselves. Right? We to know some of those unique properties |
|
| 16:25 | with our visual perception channel because visualization all relies on our visual perception to |
|
| 16:33 | the generative visual presentation and understand what going on. So we need to |
|
| 16:39 | some properties important properties of our visual channels in order to be able to |
|
| 16:44 | effective visual presentation to present our Right? This shows a very very |
|
| 16:53 | , high level pipeline of visual perception to connection. So first we have |
|
| 17:01 | visual stimulus. Light colors, right? Those are visual stimulus but |
|
| 17:09 | starts with light without light we see . Okay, those visual perception visual |
|
| 17:15 | will be perceived by eyes. eyes is a very very complex |
|
| 17:19 | So we have lens. We have . So the visual stimulus signal will |
|
| 17:24 | through lands and form some upside down at the back of the eyeball. |
|
| 17:28 | which will be the retinal and retinal be connected to some of the |
|
| 17:34 | Throughput channel fibers. New road network you pass the signal to the back |
|
| 17:42 | our brain without the cortex handles the signals. Okay. When we when |
|
| 17:48 | signal reach the back of our the cortex the part of cortex we |
|
| 17:53 | seeing things. We see shapes, , sizes, textural orientation and |
|
| 17:59 | Okay then we add onto those geometry objects. So this is the visual |
|
| 18:06 | process. Right? It's a signal signal capture and processing process. After |
|
| 18:13 | perceive things we see things now we thinking what we are seeing this. |
|
| 18:19 | enter into connection process. It's a process of trying to understand since that |
|
| 18:26 | perceive. Alright, so this connection will be applicable to not just visual |
|
| 18:32 | channel, It can be applied to sense channel of us. Right? |
|
| 18:37 | have hearing, we have taste, have feelings. So anything we feel |
|
| 18:42 | go to the brain. We starting was the thing? I'm what is |
|
| 18:47 | thing I'm touching? What is the that I hear? Well it's from |
|
| 18:52 | . When once we start thinking of question like this we enter into the |
|
| 18:57 | process. We process information we So the same thing happened to the |
|
| 19:01 | perception once we perceive it we start what they are. What are they |
|
| 19:06 | now. All right. In order understand or interpret those things. We |
|
| 19:11 | to assess our long term memory. . So I post here so what |
|
| 19:17 | stored in our long term memory that need to assess to help us understand |
|
| 19:24 | we perceive like visually or through hearing through taste of touching. What is |
|
| 19:31 | ? What are those things that are in our long term memory? Do |
|
| 19:43 | mean senses long term memory since it's senses shortened? We sense it. |
|
| 19:50 | it. Right? When we touch his surface with the extra with withdraw |
|
| 19:56 | finger. Right? That's it. how why do we know it hurts |
|
| 20:05 | ? Of course basic instinct. But understand the things that we see, |
|
| 20:12 | need to assess long term memory. instance if we see an animal we |
|
| 20:19 | thinking what animal it is. Then know okay, it's a cat or |
|
| 20:23 | dark. So we need to assess term memory. What is that? |
|
| 20:29 | would be like experience. Thank Right. It's right here, it's |
|
| 20:35 | knowledge experience we have learned. If you never know what is the |
|
| 20:42 | , what is a cat? Even animal is right there. We will |
|
| 20:45 | be able to recognize it. So connection process. Actually need to |
|
| 20:52 | long term memory. It's a rich of all those entries that register in |
|
| 20:57 | brain. Okay. Alright, so is the entire process of how we |
|
| 21:02 | things and how we process what we . Alright. How we understand what |
|
| 21:06 | see. But to summarize visual perception only these two stage, right? |
|
| 21:12 | then later this if it's effective then can quickly trigger the long term memory |
|
| 21:19 | then help us understand it quickly. ? This is related to the effectiveness |
|
| 21:23 | the visual representation. But again, is the visual perceptions. Alright. |
|
| 21:29 | we perceive things visually. After understanding we perceive things, we need to |
|
| 21:38 | some important properties that may be useful generating visual representation. So next time |
|
| 21:45 | you try to generate some visual you may need to go back some |
|
| 21:49 | these properties to think. Should I this or should I should not. |
|
| 21:55 | , Let's start with two short Okay, let's just follow the instruction |
|
| 22:03 | complete the task. Try to count many times the players wearing white. |
|
| 22:10 | the past people. Only the Alright. How many times did you |
|
| 22:44 | ? 13. 15. Alright. is the correct answer. All |
|
| 22:52 | Next question will be surprising. Did see the gorilla in the video? |
|
| 23:01 | , I did not know. Let's watch the video again. Now |
|
| 23:06 | have the hints. Ah Now you it right? But if you're focus |
|
| 23:19 | fully occupied by the task that is to you, there's certain details in |
|
| 23:27 | visual, you may ignore. All . And let's see another video. |
|
| 23:40 | , let's watch this. Yeah. you see that? Yes, but |
|
| 24:03 | is not the same person. Somebody pay attention. Now after the |
|
| 24:07 | video it's not the same person. not the same person. They post |
|
| 24:25 | classes but and the long hair but have different shirts on. Okay. |
|
| 24:37 | . So that's the the property that want to highlight. Change politeness. |
|
| 24:43 | this can be significant. Right? so they did this. We need |
|
| 24:50 | pay attention to this unique property. huh. That says human needs a |
|
| 24:57 | of tension in order to capture The changes and the details right through |
|
| 25:02 | above example. Especially when the changes over time. Alright. We watched |
|
| 25:07 | video frame by frame but sometimes we shortened memory. Alright? And there |
|
| 25:12 | a lot of information to process, ? We process it sequentially. But |
|
| 25:17 | perceive the information in a parallel We'll see that right? We perceive |
|
| 25:21 | those pixels at the same time. problem at all. But whether they |
|
| 25:27 | make an ink in your mind, a different story. So sometimes they |
|
| 25:32 | okay if you know what things you to you pay attention to right? |
|
| 25:36 | the first video I asked you to attention to how many times the people |
|
| 25:41 | white pass the ball and then you attention to those things and other things |
|
| 25:46 | it's in the same frame you Right? So those are the property |
|
| 25:50 | need to pay attention that said if the data or the story of the |
|
| 25:55 | you want to convey. Need to changes for difference for anomalies in the |
|
| 26:02 | that you need to find a way actually highlight it visually. Okay. |
|
| 26:07 | let the viewer to look through your presentation to find out things that are |
|
| 26:14 | . Alright. Made those different pop . Okay, we'll talk about |
|
| 26:17 | There is some important property that we to make things pop up. So |
|
| 26:22 | following three already showed up right. another property that you should utilizes our |
|
| 26:29 | perception system is good at observing reality in space, reality difference in |
|
| 26:37 | not over time because over time things I said we look at frame by |
|
| 26:42 | , this is short term memory when time passed the framework changed. |
|
| 26:47 | But the relative difference in space we scale it long enough so we notice |
|
| 26:52 | and it's easy to be drawn to boundaries of different regions and objects. |
|
| 26:58 | , so this property people already use to show clusters. Alright, regions |
|
| 27:04 | have different meanings that we actually use colors or different geometric highlight or shading |
|
| 27:11 | highlight them because the boundary between these even they are small. Okay, |
|
| 27:16 | can see that right? And another that this property can lead to or |
|
| 27:22 | should utilize can be showing this example say I have two values. I |
|
| 27:27 | to compel their values which one is of course the two values easy. |
|
| 27:33 | . And you can just compare the of values based on your knowledge. |
|
| 27:39 | if I had many many values and want to compel them ah effectively |
|
| 27:47 | So I tried to use visual So here I use two bars the |
|
| 27:51 | of the bus to represent their actual values. So this is the first |
|
| 28:00 | . Okay. First visual representation. attempt. Can you actually tell which |
|
| 28:05 | is bigger to me? It seems are. Yeah they're the same. |
|
| 28:20 | differences so subtle. All right now do a better job. Okay let's |
|
| 28:24 | okay if A and B correspond to percentage values. So we use 100% |
|
| 28:30 | the silhouette, the outline and then try to fill up their corresponding |
|
| 28:36 | Now it was slightly better right? on the amount of the white |
|
| 28:42 | the size of the wide region we tell. Okay. Be it |
|
| 28:47 | Alright so this is an improvement. the most effective way that if I |
|
| 28:53 | the property that I just mentioned human really good at figure out the relative |
|
| 28:58 | . I would do this properly ally with respect to a common baseline or |
|
| 29:06 | line. Alright so this is we Weber's law. Okay there is another |
|
| 29:14 | property of our human perception that you be aware of. So this to |
|
| 29:19 | which one seems longer to you. left one. The left one. |
|
| 29:27 | vertical vertically aligned one. But you what I'm going to tell you. |
|
| 29:35 | ? They actually have the same But you field this vertically aligned or |
|
| 29:42 | out table longer. This this is we call vertical dominant in our visual |
|
| 29:50 | . So how can we utilize it our visualization or how should we avoid |
|
| 29:58 | kind of situation when we visualize So I go back to this situation |
|
| 30:05 | ? If I want to compel to right? And the two numbers, |
|
| 30:09 | very, very similar. Alright. I put a vertically and I still |
|
| 30:15 | the bar if I put a vertically be horizontally. Do you still be |
|
| 30:20 | to effectively tell their values difference which is bigger? Probably not. |
|
| 30:25 | Because of the vertical dominant. So said if you want to compel the |
|
| 30:31 | right? And you use certain visual aligns your visual representation in the same |
|
| 30:39 | are either horizontally or vertically or whatever angle but they have to be oriented |
|
| 30:47 | . Do not a random differently. , that's one important thing. |
|
| 30:53 | so here is another thing. You pay attention when you create the visualization |
|
| 30:58 | of those things. I show you far it's really high level. It |
|
| 31:02 | talk about specific visualization techniques, its properties that you should be a weld |
|
| 31:08 | when you produce any time of visual . Alright, let's go to the |
|
| 31:13 | what you see when you see a , depends on what the thing |
|
| 31:17 | So, this is the first what you see the sink earth depends |
|
| 31:23 | what you know about what you are . All right. Any interpretation, |
|
| 31:29 | tried to help us understand these two . It echoed back to the visual |
|
| 31:41 | pipeline that I mentioned in the Yes. So what you see when |
|
| 31:48 | see a thing depends on what the is. So if I have to |
|
| 31:55 | a word for this sentence, the correspond to what fact fact. |
|
| 32:06 | this is fact. What you see sink us depends on what you know |
|
| 32:11 | what you are seeing. This is , opinion. All right. Now |
|
| 32:19 | , can anybody tell me why this important in visual visualization, anyone. |
|
| 32:34 | basically if it is an alien who and looked at these things, they |
|
| 32:39 | basically not find why we would find particular images so unusual. Right. |
|
| 32:47 | huh. Yeah. Yeah. Thank . So that said when we create |
|
| 32:55 | , we had to think about the background of your targeting audience. |
|
| 33:02 | so your visual presentation should be tailored on what your audience knows. Do |
|
| 33:09 | include things they are not familiar If you have to provide sufficient |
|
| 33:15 | Legends or captions to help the audience through your visualization because if you leave |
|
| 33:23 | room for guessing, you know what happen. Right. People have all |
|
| 33:28 | of interpretations. Right. One simple example is you know, have you |
|
| 33:34 | went to the gun to the art and look at some paintings? |
|
| 33:41 | Right. And then, you the the visitors looked at the same |
|
| 33:47 | . We will read different things. right. How many times when you |
|
| 33:50 | look at the painting? And then of the people next to you |
|
| 33:54 | oh yeah, I can see wow, this is cool. And |
|
| 33:58 | you see em I see nothing. ? So, this is really purely |
|
| 34:03 | on the knowledge, right? If never seen this kind of painting style |
|
| 34:07 | you say, oh, this is mess up. Right? So be |
|
| 34:11 | . All right, This is a , very important things. You have |
|
| 34:14 | keep in mind when you prepare your presentation. Right? Now, we |
|
| 34:20 | to the most single important property of visual perception that we utilize a lot |
|
| 34:26 | racial science. Okay, now you back to this as the summary. |
|
| 34:30 | let's see what is pre Attentive. attentive means that we process things very |
|
| 34:37 | in a parallel fashion, right? a large scale task, this pre |
|
| 34:43 | property allows us to process it. is a number that's based on some |
|
| 34:48 | study perception. Study Alright, Alright. I didn't decide the |
|
| 34:53 | All right. So, if you process it within 500 milliseconds, we |
|
| 34:59 | , okay, this is pre We can process it right? If |
|
| 35:03 | individual project approach object, we can it in less than 10 milliseconds attended |
|
| 35:09 | that we need longer time to process and the process is typically sequential. |
|
| 35:17 | so our visual perception actually it's a attentive process Alright because we perceive things |
|
| 35:24 | at once. We see the image at once. All the pixels get |
|
| 35:29 | our brain. You pass through of all those channels through lands Latina and |
|
| 35:35 | visual uh communication channel in our Until we reached the the back of |
|
| 35:45 | break. Right? This is purely parallel process. Pre attentive. |
|
| 35:50 | Alright. And but when we start what we see, this gets into |
|
| 35:55 | let me give you an example right look at this vision right? Without |
|
| 36:01 | too much time. What we can . Is there a lot of |
|
| 36:04 | And then there is some dots that to each other than the others. |
|
| 36:09 | ? This you perceive it in less 500 milliseconds assume that somebody really time |
|
| 36:16 | ? All right. Best pre Now we start thinking right. How |
|
| 36:25 | points in this particular cluster? Those close to each other close to each |
|
| 36:32 | . Now in order to answer this we start counting, counting is a |
|
| 36:38 | process. We count things one by is a sequential. Alright so this |
|
| 36:43 | attentive. Alright and then we'll probably thinking the other points are evenly |
|
| 36:50 | Right to measure distance, We have measure them pair by pair right? |
|
| 36:56 | is sequential. Alright, in our mentally. So now you see the |
|
| 37:01 | . Pre attentive that you barely need process it. You just proceeded |
|
| 37:06 | You notice something right? And then means that you try to go deeper |
|
| 37:12 | understand what you see, what what information there. Right? |
|
| 37:18 | Okay. So visualization mostly utilized the visualization mostly utilized the pre attentive property |
|
| 37:28 | our visual perception channel. Okay, short, so despite all those things |
|
| 37:33 | you can read through it yourself. . So in short, what we |
|
| 37:37 | about is when we create visual we try to make things that we |
|
| 37:43 | the audience to pay attention to pop . Okay, We try to make |
|
| 37:49 | problem. Let's start with an Let's count how many? Three in |
|
| 37:54 | series of numbers. How many? right. In order to find all |
|
| 38:04 | three, you have to go through one by one. Right? This |
|
| 38:07 | attentive or pre attentive process. Now is pre attentive attentive right? Because |
|
| 38:18 | present those information sequentially. Alright, how to make things pop up. |
|
| 38:27 | column. Alright, you immediately see immediately? The street. Of course |
|
| 38:33 | next step will be counted. But see their locations. Alright, three |
|
| 38:38 | up. All right, So there different way to make things pop |
|
| 38:45 | So this just list a few situations few visual properties. Sorry, visual |
|
| 38:53 | that people can utilize to make things up. Right? So this comes |
|
| 38:57 | some visual signs. Okay. Not visualization communities but visualization community utilize some |
|
| 39:03 | these conclusions to help make visualization more . Right. Make the important information |
|
| 39:09 | up. All right. Of course need to know what is important. |
|
| 39:14 | ? So that's a separate issue. by case. Alright. Any questions |
|
| 39:21 | attentive and pre attentive properties of our perception. Alright, nope. All |
|
| 39:34 | . In addition to those uh visual that people can utilize to encode information |
|
| 39:40 | make things pop up. We have other things as shown here that people |
|
| 39:45 | use to encode different types of information data. Right. Data is not |
|
| 39:52 | . Ah singular, Right? We different types of data that describe different |
|
| 40:00 | . Right? We have numeric data probably describe describe quantity information. And |
|
| 40:06 | we have labeled data that describe categorical groups, classes. Right? So |
|
| 40:14 | can we encode this information visually So, here are the options. |
|
| 40:19 | ? There are many of them. the question will be are they all |
|
| 40:23 | effective in different situations? Like if information I want to visually encoded describing |
|
| 40:31 | of information, Which one would be effective? Right, are they or |
|
| 40:36 | effective to answer that? Let's look another example. Alright. So if |
|
| 40:44 | two bars represent two values which one bigger and how big it is. |
|
| 40:47 | one is bigger is? So But which one? So that the |
|
| 40:51 | 1? Okay, so how big much bigger it is compared to the |
|
| 40:57 | bar I guess about five times. . Five. Any option if you |
|
| 41:11 | at my mask cursor 4 4. . four. You can read |
|
| 41:18 | So now let's use the two All right. The right one is |
|
| 41:26 | bigger. But how much bigger? guess. Probably areas four times if |
|
| 41:37 | diameter is double. All right, me quickly show you this. It's |
|
| 41:41 | five times bigger. Alright. It's very effective right? Compared to the |
|
| 41:46 | representation. Alright. So that's because bar representation we used the lens use |
|
| 41:53 | of the geometry to encode numeric value . We use the areas of the |
|
| 41:59 | to encode numeric values apparently lends it's effective than areas. Okay. This |
|
| 42:06 | be very very important when you uh at some later things of trust and |
|
| 42:15 | something. Some type of trust. should avoid music. Alright. That |
|
| 42:21 | I do want to ask one quick going. So if we're just trying |
|
| 42:26 | say one thing is larger than the thing by the certain amount. Most |
|
| 42:31 | the time we would prefer to just numbers, right? Unless it's like |
|
| 42:35 | more complex scenario. Yeah. If just two numbers right? But if |
|
| 42:39 | have many numbers then you still need use so if you just care about |
|
| 42:43 | one is bigger then yes. The can also work right? If you |
|
| 42:46 | want to read the precise values So that says you know the among |
|
| 42:53 | effectiveness of encoding the quantity of data different visual properties, they are not |
|
| 42:59 | effective. Right? So this is trust that I want to share with |
|
| 43:06 | depending on the types of information you your visual perimeters to encode right? |
|
| 43:11 | have different ranking of those attributes to but things are not always absolutes. |
|
| 43:18 | are always exceptions that some of those ranking primitives may not be as effective |
|
| 43:27 | it should be in some situations so have to use it based on your |
|
| 43:32 | situation. But this is the overall based on some studies. Alright, |
|
| 43:38 | that is the visual perceptions and connective for our visualization. Any questions before |
|
| 43:45 | move to the next topic, I we're running out of time. I |
|
| 43:49 | my best. Alright. The next will be about generating affected charts and |
|
| 43:57 | but there is a set of principles gestalt principles that people typically exercise when |
|
| 44:03 | decide whether the plot is effective or . So I will not be able |
|
| 44:07 | go over all of them in So some of them related to our |
|
| 44:12 | perception channels. Some of the property our visual perception channel that we didn't |
|
| 44:18 | . So like the enclosure. So , our eyes tend to connect things |
|
| 44:24 | on our knowledge of a shape. we only provide dash line, straight |
|
| 44:28 | line, we know this is a line. Right? We don't need |
|
| 44:32 | full solid line to realize this is straight line if it's dash like so |
|
| 44:37 | some of those properties still related to uh perception channel but I would jump |
|
| 44:44 | some criteria of determining whether your visual is effective or not. So here |
|
| 44:50 | one famous criteria. It's called graphical . So what is that? It |
|
| 44:56 | that we can the generative visualization can the viewer the greatest number of ideas |
|
| 45:04 | the shortest amount of time. With list in in the smallest space. |
|
| 45:09 | lot of criterias come into here. ? And people can roughly ah reformulate |
|
| 45:20 | criteria into something like this, That's easily ah check about how to |
|
| 45:29 | easily usable. Okay, expressiveness and . Okay. Effect expressiveness in other |
|
| 45:38 | , means that tell the truth. requires the visual representation accurately encodes the |
|
| 45:45 | of the data that needs to be . Try not to distort the |
|
| 45:50 | Try not to include buyers into your . Tell the truth. Alright, |
|
| 45:56 | one criteria, graphical integrity and if I can guarantee you encode the data |
|
| 46:05 | do it effectively, effectiveness with precision and emphasis. Okay. That |
|
| 46:12 | people read the information effectively make things up if needed. Alright, so |
|
| 46:19 | are the two make criterias people utilize they check whether their visual presentation is |
|
| 46:26 | or not out down to specific graphical representation. Right? This is |
|
| 46:33 | set of visual representation, plots and . This is some set of principles |
|
| 46:40 | people practice when they generate plots and . Like like Tre hissed a gram |
|
| 46:49 | and scatter plots, et cetera. . And despite the many principles, |
|
| 46:56 | of the of them, I would guidelines and use them as much as |
|
| 47:00 | can. And sometimes those principles may be useful depending on the situations, |
|
| 47:06 | among them, the highlighted ones are applicable. The first one reduce clutter |
|
| 47:15 | data stand out. Use visually prominent elements to present the data all the |
|
| 47:21 | secondary information line, scale lines, line should be put in the |
|
| 47:28 | All right, owning the data should the full attention right. And here |
|
| 47:35 | an important thing you should keep in for visualization less is more Less is |
|
| 47:42 | . If the simple visual presentation can sufficiently conveyed information, do not use |
|
| 47:48 | things. Simple is good. And remove all the unnecessary elements. |
|
| 47:55 | elements. Okay, okay. And understanding provide explanations for each. Trust |
|
| 48:04 | graph should provide a caption to explain this chart or graph is showing what |
|
| 48:12 | can be drawn from the visual If the charts. If the the |
|
| 48:20 | contained multiple charts for instance, for purpose, that you cannot put them |
|
| 48:25 | in one figure because of the overlap those plots properly along their common |
|
| 48:32 | A like just tickles plots. so those are the highlighted principles are |
|
| 48:37 | applicable in most of the situation. other depends on the data and the |
|
| 48:43 | and information you want to highlight. may or may not be applicable. |
|
| 48:47 | right. But one thing you should attention generating effective plots is always an |
|
| 48:54 | process by many times the first iteration plot. It's not the best. |
|
| 49:01 | ? I think you already have experience also many default settings of the plotting |
|
| 49:06 | cannot generate the most effective plots for . The default setting provided by those |
|
| 49:11 | . Many of times you need to the parameters. Alright. Change the |
|
| 49:16 | , changed different visual to generate the effective plus. So plotting is an |
|
| 49:22 | process. Make sure this is in mind next time. All right. |
|
| 49:28 | , those are the things we should next. I spent a couple of |
|
| 49:32 | to talk about things you should not . Since we should avoid. This |
|
| 49:36 | a little surprise. Try not to pie chart. There are different elementary |
|
| 49:41 | type, right. Petra is one the most popular plot type people use |
|
| 49:46 | plot information. Alright, what's the here. Alright, I have full |
|
| 49:54 | . I want to show their market in terms of percentage and I hope |
|
| 50:00 | can intuitively show show me which supplier the most market share. But can |
|
| 50:09 | tell Which one has the most market . Seems like speed or a. |
|
| 50:21 | a or B. But you cannot for sure until I oh I think |
|
| 50:25 | didn't have the number but be looks . Right? But the actual largest |
|
| 50:36 | , the supplier who has large market is A. I will show you |
|
| 50:40 | the next plot. But visually supplier looks bigger. Why there's one property |
|
| 50:53 | mentioned because the like that right one to table Right? three d. |
|
| 51:02 | is always a killer. Don't use D. Effect. This is basically |
|
| 51:05 | the information why you need three This is one thing to a |
|
| 51:09 | I will emphasize that. Again. reason anybody still remember vertical dominant this |
|
| 51:24 | corresponding to supply B. Unfortunately it's vertically while A. Is horizontally. |
|
| 51:33 | see that. Alright. So So actually used the area and the |
|
| 51:43 | of each sector to encode quantity Right? We know these two are |
|
| 51:48 | effective. If your goal is to which one has the largest market |
|
| 51:54 | use this right? We know this the most effective way to encode quantity |
|
| 51:58 | information if you want to compel which is larger. Alright. And there's |
|
| 52:04 | situations that people prefer petra right? Petra gives you a whole if your |
|
| 52:11 | is some percentage within 100% right? split into different sectors then Hi chart |
|
| 52:19 | give you a feeling of whole it's and each one take a chunk of |
|
| 52:24 | but outside of it. Don't use tra. Okay. don't use Petra |
|
| 52:29 | if you have to use Petra do use three D effects. This is |
|
| 52:36 | another example of how 3D effect come the way. The how we can |
|
| 52:41 | read the information. It doesn't It doesn't add anything other than visual |
|
| 52:47 | . Right? We don't need visual . The visualize. The goal of |
|
| 52:50 | is not trying to generate beautiful No no no that's not the goal |
|
| 52:54 | visualization. We're trying to generate effective presentation so the underneath information can be |
|
| 53:03 | conveyed to the targeting audience in an intuitive and precisely accurately. Alright So |
|
| 53:13 | not use three d effects. Secondary Y axis should be avoided or |
|
| 53:19 | as much as you can. So times you need to compare two sets |
|
| 53:24 | data. They have different meanings and data range to associate them. For |
|
| 53:30 | whether they have some correlation but but are defined in different range And have |
|
| 53:35 | units like these. two. One is revenue the units millions in |
|
| 53:40 | of money. The other is number salesforce. How many salesman's? It's |
|
| 53:45 | theaters Right? If we use to one for one data people get confused |
|
| 53:52 | the tendency of people read the value we try to find the nearest reference |
|
| 53:57 | to try to read the value for bars. Right? And then it |
|
| 54:01 | cause confusion. So try not to it the better way is directly at |
|
| 54:08 | values at the data points. If are not many data points or if |
|
| 54:13 | are too many data points that prevent to add those values onto it to |
|
| 54:18 | wide occlusion. Then you put them two plots by aligning them properly. |
|
| 54:24 | you still can compare their chains. . Secondary y axis should be |
|
| 54:30 | Okay. Okay. So I think can still spell a couple of minutes |
|
| 54:38 | with you exercise. Alright, let's to exercise to practice what we just |
|
| 54:45 | about the principles to improve some of charts. So, this first |
|
| 54:50 | Alright, so this is the survey . What survey? We surveys the |
|
| 54:55 | of the preference of music amongst Right? Over the past two |
|
| 55:02 | So, this survey was done in University of Miami. So, students |
|
| 55:07 | the Survey subjects. So they did first survey in 1994. And then |
|
| 55:13 | did it again in 2014, two of how? Right. So they |
|
| 55:18 | to surveyed the favorite music form. ? So, and the question they |
|
| 55:30 | to answer is how this preference changed time. How this change over |
|
| 55:37 | So, the change is the focus . Alright, So this is the |
|
| 55:42 | visual presentation. We use pie All right. And we currently see |
|
| 55:49 | changes. Hard rock music. Got lot of love After two decades. |
|
| 55:57 | ? Apparently have a few more fans the others. But how about the |
|
| 56:02 | john's other music johns like samba raggy , classic? Anything change? It's |
|
| 56:16 | very clear. Of course we can the numbers, right? But this |
|
| 56:20 | you to do what attentively compelled the to notice the change. Right? |
|
| 56:29 | any better idea of plotting this information than using pie charts? Remember the |
|
| 56:37 | here is the change of the preference time. We don't care about the |
|
| 56:47 | percentage it takes right? Each music takes. We just care about the |
|
| 56:54 | over time, bob jobs post. Batra can only show one year. |
|
| 57:06 | need to show the change over Remember the second example, I show |
|
| 57:11 | about the change your percentage of That will be the chart. I |
|
| 57:17 | use. All right. I don't about the percentage the precise percentage of |
|
| 57:27 | particular music young that is preferenced by students, right? I don't care |
|
| 57:34 | the exact percentage. I just care the chain over time. So, |
|
| 57:39 | this plot will be more effective, ? You can clearly show we get |
|
| 57:42 | lot of new fans for heart. walk. Simba drop a little bit |
|
| 57:49 | hip hop. We gained a couple friends, new friends. Sorry, |
|
| 57:55 | friends, but not a lot. , But country and classic jobs. |
|
| 58:00 | right over the year. So less and fewer people like to listen to |
|
| 58:05 | and classic. But this information can more effectively converted using this truck than |
|
| 58:12 | truck. Right. Okay. So one exercise next. Right? This |
|
| 58:19 | another survey data we try to So the So the question is in |
|
| 58:26 | , what attributes are the most important you In selecting a service provider? |
|
| 58:33 | this is a service company and there's couple competitors that this company try to |
|
| 58:40 | out what part of the things that should improve in order to attract more |
|
| 58:45 | . Right? So there were 77 they are looking at. They try |
|
| 58:51 | decide which one they should focus on improve their business. This is the |
|
| 58:56 | results and they conclude that demonstrating effectiveness the most important consideration for customers to |
|
| 59:03 | a provider. Right? So they to visualize their results. Alright. |
|
| 59:10 | any issues you see and how you it. Anyone. Okay, So |
|
| 59:41 | of the time limits. Let me show you show you the improve with |
|
| 59:46 | presentation. All right, So this much better. All right, tell |
|
| 59:52 | what has been improved. We changed collapse of the most important notices and |
|
| 60:02 | but stopped. Alright. We exercise adding emphasis to emphasize the information that |
|
| 60:11 | want the reader to pay attention Alright, previously, they're all having |
|
| 60:17 | scent emphasis. That means there's no in your vision. All right. |
|
| 60:23 | your story has always a seam You to complain. The scene is here |
|
| 60:30 | is the most important consideration. You want to highlight that conclusion. |
|
| 60:35 | ? This one highlights it anything How about the alignment of the |
|
| 60:46 | Which one is easy to read? have to say the new one, |
|
| 60:52 | . Why cliff was left aligned? . Left justified A line. That's |
|
| 61:01 | human reading habit. We tend to from left to right, not central |
|
| 61:07 | . Try not to use central Okay. If you want people to |
|
| 61:11 | your information left justified align. So course this is a lie because this |
|
| 61:18 | right justified line because they are close the bars, right? You want |
|
| 61:21 | align them closer to the bars so the exception. But most of the |
|
| 61:27 | you should align them to the left another changes. You see all those |
|
| 61:33 | key elements. They're gone. That's also our human reading habits. |
|
| 61:40 | tend to read things are that are horizontally more effectively than those informational elements |
|
| 61:47 | in non horizontal orientation. Right? instance in the previous representation, this |
|
| 61:53 | fortified degree oriented. You have to your head to read it right? |
|
| 61:58 | first situation some of the before the , the default setting of some plotting |
|
| 62:04 | . We vertically aligned those labels. bad. That's really bad. |
|
| 62:09 | Try not to do that. All . Always oriental labels text horizontally as |
|
| 62:16 | as you can. And also we rid of non necessary visual elements to |
|
| 62:21 | . So we get rid of the vertical axis. This exercise. One |
|
| 62:26 | the visual property We can connect those even they are not physically connected. |
|
| 62:32 | human perception can align them following the part. Right? And we also |
|
| 62:37 | rid of those this reference line. don't need it because we have the |
|
| 62:43 | . All right. So those are improvement. And of course we keep |
|
| 62:48 | . Sometimes. Wide spaces, not in your visual representation whitespace just like |
|
| 62:55 | punctuations in sentences without them. It's , really exhausted to read or listen |
|
| 63:04 | right. The same thing happened for . So leave out some white space |
|
| 63:09 | it's needed. Okay. All So any questions when I play this |
|
| 63:14 | animation. I just wanted to make comment about both of them that do |
|
| 63:20 | very good job is sorting by the of the bars. Because if you |
|
| 63:26 | have a lot of experience, you just let the default graphing software to |
|
| 63:30 | . Depending on the order of the points to display. I think it's |
|
| 63:34 | better. Do you agree with Going great. So it's much better |
|
| 63:39 | sort the bots based on their otherwise, you know, it's really |
|
| 63:44 | to read. Especially when the two are too close and they are located |
|
| 63:48 | far away. Right? So this another example to demonstrate how you can |
|
| 63:58 | clutter your visual representation to make an representation of the data cleaner. Any |
|
| 64:07 | questions. And then to emphasize something not critical. Right? Those uh |
|
| 64:20 | lie those numbers. You know you just emphasize them or even just put |
|
| 64:25 | actual values on the bars. If don't need to show all of |
|
| 64:29 | One question I had going. If just had those bars and which we |
|
| 64:35 | want to emphasize any particular bar. you want us to use black on |
|
| 64:41 | or different colors or? Great. you have any suggestions there? If |
|
| 64:45 | don't want to emphasize anything same color which color would you like us to |
|
| 64:51 | ? Depending on If everything is equally , Everybody is equally important. You |
|
| 64:56 | still use visually prominent color depending on background. So most of the time |
|
| 65:01 | background is white you should use darker color for your chance for your |
|
| 65:07 | . But if you do want to things like this particular example, all |
|
| 65:12 | other bars that are not need to paid attention to should use a very |
|
| 65:17 | color like the Great. All Like this one because you emphasize this |
|
| 65:28 | only. This one got color the stone. But if they are equally |
|
| 65:33 | , which means that there's no then all of them should be color |
|
| 65:37 | color using the same color. Try use many cards. Try not to |
|
| 65:46 | many colors in the plot. It help unless they have different groups. |
|
| 65:51 | bar trust belong to different groups or clusters. Then each cluster should get |
|
| 65:58 | unique color that's good questions. Any questions dr Chen I have two |
|
| 66:12 | Yes. Go ahead on the What course is this from? You |
|
| 66:15 | you were teaching this India course visualization And my second question is given a |
|
| 66:23 | . How do we decide which type visualization to use? Good question. |
|
| 66:31 | very good questions, depends depends on you want to show and of course |
|
| 66:36 | nature of the data. So if have a tabular data right? The |
|
| 66:43 | entries it's stored in a table and entry has multiple attributes then depends on |
|
| 66:50 | attributes and what type of this particular values in this particular attributes they |
|
| 66:56 | You choose the proper plotting type for if you want to plot numeric value |
|
| 67:03 | data entries are organized based on some order. For instance time you measure |
|
| 67:10 | over time then you can use line right? Like the temperature change over |
|
| 67:16 | . So use line charts. But the data entries do not have intrinsic |
|
| 67:23 | not order based on time. It simply labeled like this french fried potato |
|
| 67:28 | that use by trust. If the are numeric. Alright. And if |
|
| 67:36 | is only true. If you only to plot one attribute. If you |
|
| 67:40 | to plot to excuse to to show correlation then you can either do the |
|
| 67:48 | live close or you do the just plots like the example I show and |
|
| 67:53 | said try to avoid to set to access right? You can superimpose the |
|
| 68:01 | live plots together to see whether they similar change or you can do scatter |
|
| 68:07 | . So it all depends on what you want to review from the |
|
| 68:12 | But in the beginning, most of time you start with some basic common |
|
| 68:19 | types. Okay. To check those attributes to find whether this attribute is |
|
| 68:26 | or not. Should I further explored or not? Sometimes quickly future. |
|
| 68:30 | this is not interesting and move on the other attributes. So you only |
|
| 68:35 | time to the attributes that after the exploration you decide this is important. |
|
| 68:44 | you. Alright, let's thank Professor for spending his time with us. |
|
| 68:49 | if you want to learn more about . He teaches an excellent course. |
|
| 68:54 | you. So I think everyone is cars to take because you learn a |
|
| 68:58 | because visualize system is a really important regardless of what you do in your |
|
| 69:03 | . Thank you Professor. Yeah, probably have more slides about colors but |
|
| 69:08 | was just showed it with dr normally . And then he will share that |
|
| 69:13 | colors is also very important. Thank . Thank you. So we're gonna |
|
| 69:20 | spend a few more minutes with the If you guys have any anything you |
|
| 69:26 | to discuss from the from what you today, you can ask those questions |
|
| 69:32 | if you want to discuss anything, know, we can stick around for |
|
| 69:35 | few more minutes. Mm hmm. the anomaly. What are some of |
|
| 69:45 | libraries that you use for plotting? , that's a good question. And |
|
| 69:48 | should have also asked, yeah, Glenn ng Chen that question. These |
|
| 69:56 | , a lot of the libraries have very good. Even let's say Microsoft |
|
| 70:00 | allows you to produce very good graphs days. I personally use a Mac |
|
| 70:06 | live. It's a python library and I, you know, you guys |
|
| 70:11 | doing a lot of data science, already had a lot of visualizes on |
|
| 70:16 | of it. But remember we're mostly about research papers and there, all |
|
| 70:21 | need is basic graphs. And A for you to customize basically various elements |
|
| 70:28 | that graph. So these days, example, if we're having this conversation |
|
| 70:32 | years ago um, you know, , Excel would probably have been a |
|
| 70:37 | poor choice, especially the defaults. you know, things have gotten very |
|
| 70:40 | now. So, um, so you can use pretty much any |
|
| 70:46 | they are reasonably good. I personally that part live in a new |
|
| 70:51 | That's another tool that I've used over time. I can let's hear from |
|
| 70:55 | , you know what tools you might using beyond the Excel method labor and |
|
| 71:09 | just do a quick survey. So many of you have written research |
|
| 71:15 | or reports in the last, let's one year. You know, what |
|
| 71:18 | do you use? Maybe, you , everyone can basically mention your tool |
|
| 71:23 | , on the chat. Probably it's efficient to do this kind of |
|
| 71:26 | So if you could just type up tool that he used for plotting |
|
| 71:33 | because you know, it's kind of to know what other people are using |
|
| 71:37 | in case. You know, we to check them out and we're not |
|
| 71:40 | about diagrams, we're talking about We got one answer saying anything |
|
| 71:51 | Anyone has used everyone used math Okay. The crazy part for using |
|
| 72:13 | , did anyone actually use Excel? have used Excel. Yeah. |
|
| 72:22 | um, you know, one of issues with something like Excel is for |
|
| 72:26 | , partly discipline, etcetera. You very easily export to pdf and include |
|
| 72:30 | in the atlantic document. And we're most of your, I'm going to |
|
| 72:36 | latic's or have used that trick for research publications. It becomes very easy |
|
| 72:43 | , if you change the data, becomes very easy to rerun the |
|
| 72:46 | generate the pdf, it goes to right fight and you just compile |
|
| 72:49 | you know, paper when you're If you use something like Excel, |
|
| 72:54 | a little bit more clicking around opening exporting, you know, there's a |
|
| 72:58 | bit of overhead there. So that's to keep in mind. But you |
|
| 73:02 | export even Excel charts as Pdf, is what you should be doing, |
|
| 73:08 | your paper. You should not be exporting the Jpeg or something like |
|
| 73:11 | You want the vector format that's uh and nice. And Excel does allow |
|
| 73:16 | to explore that but it's a little clunky. Especially if you're changing the |
|
| 73:21 | . But if it's in the data have to you have to somehow import |
|
| 73:24 | data re plot and then maybe, know, manually exported, move the |
|
| 73:30 | around. It's a it's a you , it's quite involved. But if |
|
| 73:33 | using python my plot live or juicy organized plot, it's um it's a |
|
| 73:39 | easier because you can have the read the program, plot the |
|
| 73:43 | export it and put it in the folder paper. They kind of want |
|
| 73:48 | efficient pipeline when you're close to writing paper but when you're in an exploratory |
|
| 73:55 | , something like that. Yeah. you know, Seaborn is I believe |
|
| 74:00 | some something that's built on top of part. Is that correct? My |
|
| 74:05 | is I think it uses my part . Yeah, but it allows a |
|
| 74:10 | higher level control, you know, the hole in the charging process. |
|
| 74:16 | if you're trying to build some complex especially multi chart interactive um visualization. |
|
| 74:26 | you would probably use something like It's also very popular, which is |
|
| 74:32 | more like a data science type Alright, so let's wrap up today's |
|
| 74:40 | I'm gonna discuss. Stop |
|
