1 00:00:02,780 --> 00:00:07,450 Hey, guys, now let's see how to do the forecasting, the altitude model in Python. 2 00:00:07,990 --> 00:00:08,590 All right. 3 00:00:08,600 --> 00:00:13,760 In the previous lecture, we talk about the simple linear regression and this is what the output which 4 00:00:13,760 --> 00:00:14,270 we got. 5 00:00:14,720 --> 00:00:20,180 And if at all, you can see this is what the output and Peter want is simple, linear. 6 00:00:20,180 --> 00:00:23,650 And also the forecast is also just extension of this line. 7 00:00:24,080 --> 00:00:31,970 And whatever the feeling part, you can see, this part is basically the confidence interval or margin 8 00:00:31,970 --> 00:00:32,390 of error. 9 00:00:33,110 --> 00:00:36,280 Now, let's go back like let's see how to build a. 10 00:00:37,130 --> 00:00:39,600 So this is what the data we're interested in. 11 00:00:39,890 --> 00:00:43,130 So what we will do will take the strain data on that as data. 12 00:00:43,460 --> 00:00:45,650 We already use this with that. 13 00:00:45,890 --> 00:00:52,270 And we will take this train data and the data and for this train data and put the stats data, we will 14 00:00:52,280 --> 00:00:56,100 take the regression or attitude modeling for this. 15 00:00:56,530 --> 00:00:56,930 All right. 16 00:00:57,050 --> 00:00:58,230 So let's get ready. 17 00:00:58,250 --> 00:01:02,320 So click on this text and type the attitude model. 18 00:01:03,740 --> 00:01:07,970 How did you model is basically adding up the seasonalities. 19 00:01:07,970 --> 00:01:09,470 You're just adding up the seasonality. 20 00:01:10,220 --> 00:01:12,920 So let's take this leg up on the train data. 21 00:01:14,300 --> 00:01:16,060 So this is the train. 22 00:01:16,070 --> 00:01:20,600 So we have the train data and we have already the data as we preprocessed. 23 00:01:20,960 --> 00:01:22,490 We've already taken up the time. 24 00:01:22,490 --> 00:01:24,410 We already taken up the month. 25 00:01:24,680 --> 00:01:27,720 We already got the time as well as the month. 26 00:01:28,070 --> 00:01:34,150 So what we need to do here is that we need to create the dummy for this and they've done this. 27 00:01:34,160 --> 00:01:35,180 We need to extract it. 28 00:01:35,540 --> 00:01:42,470 So before that, let's start with the basic visualization so that you can understand much better how 29 00:01:42,470 --> 00:01:46,310 exactly that can all be able to calculate the seasonality here. 30 00:01:47,270 --> 00:01:47,700 All right. 31 00:01:47,930 --> 00:01:51,680 So first step, what we do is that we will draw the scatterplot. 32 00:01:52,430 --> 00:01:58,230 We will draw the scatter drawn between the idea of the school train. 33 00:01:58,610 --> 00:02:00,020 This is my training data. 34 00:02:00,410 --> 00:02:06,200 And I would take the time component is what makes access or so. 35 00:02:07,010 --> 00:02:12,580 And we can take one more thing for both of sales. 36 00:02:13,070 --> 00:02:19,160 So this is basically the scatterplot, which looks like, all right, we got some error. 37 00:02:21,770 --> 00:02:23,150 Yeah, so property. 38 00:02:25,470 --> 00:02:28,380 X and Y, this is the first code train. 39 00:02:29,130 --> 00:02:36,750 All right, so it's a slow train and this is how it looks like, OK, basically like, you know, and 40 00:02:36,930 --> 00:02:39,740 you can take the block. 41 00:02:39,750 --> 00:02:46,700 And if you take the same values on the X axis, these the school year and years on this. 42 00:02:46,710 --> 00:02:56,490 And if you want the blue lines and this is how you get in, so I'll go with the dotted lines. 43 00:02:56,790 --> 00:02:58,020 That will be nice for me. 44 00:02:58,320 --> 00:03:00,930 So this is a just joining of each and every point. 45 00:03:00,960 --> 00:03:02,700 This is how it looks for me. 46 00:03:03,210 --> 00:03:03,540 All right. 47 00:03:03,570 --> 00:03:09,830 Let me increase up the size of being able to go and pick size. 48 00:03:10,320 --> 00:03:12,900 And I take this to the standard 15 to eight. 49 00:03:13,360 --> 00:03:16,030 So this is the way to look at the big block now. 50 00:03:16,080 --> 00:03:17,680 So this is quite true. 51 00:03:18,630 --> 00:03:19,030 All right. 52 00:03:19,050 --> 00:03:23,050 So what I will do now is that let me color it. 53 00:03:23,250 --> 00:03:26,330 So what I'll do instead of being the only the scatterplot, the plane. 54 00:03:26,610 --> 00:03:27,600 So what I'll do. 55 00:03:27,630 --> 00:03:32,230 Let me color according to the month. 56 00:03:32,250 --> 00:03:35,280 Let me color according to the Moniz's, according to the MONDLI. 57 00:03:35,760 --> 00:03:41,040 So to do that I can use the seabourne dot scatterplot. 58 00:03:41,940 --> 00:03:44,270 So we have the bond brought there. 59 00:03:44,280 --> 00:03:50,430 So you have access on the train of EOS or. 60 00:03:51,750 --> 00:03:52,170 All right. 61 00:03:52,170 --> 00:03:55,020 I can take it like it is. 62 00:03:55,020 --> 00:04:06,690 X axis and y axis is sales and the data is equal to the from the school train. 63 00:04:07,050 --> 00:04:14,720 So if I see that the scatterplot and this is how this guy got all right. 64 00:04:14,730 --> 00:04:22,260 So and I can still increase the size X is equal to something like three hundred, I can increase the 65 00:04:22,260 --> 00:04:25,110 size of my silicones like Bego Cycos. 66 00:04:25,110 --> 00:04:31,200 I can get with this and medium level up so I can get to Endre. 67 00:04:31,620 --> 00:04:36,950 And so this is, this is, this is what I can get is what I can really get from this. 68 00:04:37,320 --> 00:04:44,640 So what I can do, I will, I will basically color with you is basically the coloring font and you can 69 00:04:44,640 --> 00:04:46,500 color this using the. 70 00:04:52,600 --> 00:04:53,540 Month them. 71 00:04:53,560 --> 00:04:58,790 So this is the month name I can use this month, name and color recording. 72 00:04:59,510 --> 00:04:59,890 Yes. 73 00:05:00,190 --> 00:05:02,360 Now you can see we have the coloring. 74 00:05:02,870 --> 00:05:08,320 This is the January, February, March, April, May, June, July, August. 75 00:05:08,680 --> 00:05:11,640 And you can see the plot is being repeated. 76 00:05:11,640 --> 00:05:20,260 The plot has been repeated and it's clearly visible, like how exactly the data has been changing. 77 00:05:20,500 --> 00:05:23,100 So according to the colors, we have it later. 78 00:05:23,320 --> 00:05:29,440 And if at all you want this month name you could all on this month name to be displayed. 79 00:05:29,740 --> 00:05:37,060 And we can also do that using the power to text be able to detect is an excellent way of doing that. 80 00:05:37,330 --> 00:05:45,160 What you do what we need to do here is that we need to give the exposition and we need to give the Y 81 00:05:45,160 --> 00:05:50,070 position and the text and we need to do the Y position. 82 00:05:50,080 --> 00:05:52,540 And finally, we need to do the text here. 83 00:05:52,990 --> 00:05:53,980 So this is the text. 84 00:05:54,060 --> 00:06:01,830 So the deal follows train and it's a month we can get it, but we have to do the step by step if we're 85 00:06:01,900 --> 00:06:03,280 not put on the spot. 86 00:06:03,310 --> 00:06:10,420 So what I'll do for I in range or for I in range of. 87 00:06:11,520 --> 00:06:16,360 You can take it from Lentol Daftness Catrine. 88 00:06:16,770 --> 00:06:24,390 So this is where to put all these values, I can iterate this and for each and every part I'll take 89 00:06:24,390 --> 00:06:25,100 it to I. 90 00:06:25,410 --> 00:06:26,910 And this is also I. 91 00:06:27,750 --> 00:06:31,680 This is also I and this is also my eye. 92 00:06:31,920 --> 00:06:34,090 So you can get all the violencia. 93 00:06:34,380 --> 00:06:42,400 So what I mean to underscore, for example, predict like zero you'll get the zero to value. 94 00:06:42,420 --> 00:06:45,410 So this is the zero is nothing like a two thousand. 95 00:06:45,440 --> 00:06:52,260 And so those information you can get if at all, you want the month of detail, you will get the January 96 00:06:52,560 --> 00:06:56,980 and February one and so on. 97 00:06:57,360 --> 00:07:00,570 So they can do like an excellent way to execute that. 98 00:07:01,140 --> 00:07:07,190 And you can see that, you know, this is a January, February, March and May, June, July. 99 00:07:07,200 --> 00:07:13,470 And you can easily plant it and you can see the behavior, how exactly the data has been there. 100 00:07:13,470 --> 00:07:14,470 And you can see that. 101 00:07:14,880 --> 00:07:15,340 All right. 102 00:07:15,360 --> 00:07:17,130 So we can also do one more thing. 103 00:07:17,280 --> 00:07:18,690 That is a one special. 104 00:07:18,690 --> 00:07:22,640 May I use it for, like, you know, setting up these options? 105 00:07:22,650 --> 00:07:25,140 It is like coloring. 106 00:07:25,260 --> 00:07:27,480 I use this this time on the set. 107 00:07:27,490 --> 00:07:29,730 And one is also the color map. 108 00:07:30,120 --> 00:07:34,560 This is nice color map where we can segregated, sort of like the rainbow colors. 109 00:07:34,890 --> 00:07:39,330 It's a solid line, solid color to each and every month. 110 00:07:39,340 --> 00:07:41,340 It's really the coolest part. 111 00:07:42,030 --> 00:07:43,860 But the problem is that you like. 112 00:07:44,370 --> 00:07:46,230 But it's OK then it is. 113 00:07:46,560 --> 00:07:52,170 The color is being repeated but still we can understand much better visualization than the previous 114 00:07:52,170 --> 00:07:52,440 one. 115 00:07:53,040 --> 00:07:58,080 But anyway, but one thing we can see is that we can see the seasonality part here is 12. 116 00:07:58,320 --> 00:08:04,180 So roughly, roughly every year that's the cycle has been repeated and roughly making up like every 117 00:08:04,180 --> 00:08:06,600 year the cycle has been repeated. 118 00:08:06,960 --> 00:08:15,510 So in this case, my seasonality is to OK, so now the season from this observations are right on seasonalities 119 00:08:15,540 --> 00:08:16,110 to all. 120 00:08:19,080 --> 00:08:20,470 Seasonalities, Twala. 121 00:08:20,730 --> 00:08:33,870 And we need to create, create and minus one dumbness, so which is basically 11 double's 11 columns. 122 00:08:34,220 --> 00:08:38,350 OK, so I need to create 11 columns here.