WEBVTT 00:00.990 --> 00:06.210 In this video we learn how to create two dimensional scatter plots so scatter plots are specifically 00:06.210 --> 00:10.530 useful to detect the relationships between two numerical features. 00:10.770 --> 00:16.680 So for example in our Titanic dataset we have the numerical features aids and fair. 00:16.920 --> 00:22.710 And by creating a set up plot for these two features we might get a first visual impression whether 00:22.710 --> 00:26.400 there is a relationship between these two factors. 00:26.400 --> 00:32.970 So for example our hypothesis might be that older passengers have more money and therefore pay a higher 00:32.970 --> 00:40.430 ticket price so we will see this in a minute but first of all let's import pandas and we import met 00:40.430 --> 00:51.600 plot lip and we are using the Seabourn style and then we are working with the Titanic dataset and now 00:51.630 --> 00:58.200 if you want to create a scatter plot for the Titanic data frame we can call here the plot method and 00:58.200 --> 01:03.320 instead of passing here a line to the kind parameter we can pass scatter 01:06.390 --> 01:08.480 and now let's press shift type here. 01:08.550 --> 01:11.810 So here we have the parameters X and Y. 01:12.030 --> 01:19.070 And here we have to determine which columns should be on the x axis and its column on the y axis. 01:19.340 --> 01:26.880 And here and our example we want to have the H column on the x axis so we pass here the H column to 01:26.880 --> 01:35.260 the parameter X and within quotation marks here and on the y axis we want to have the column 01:37.980 --> 01:41.570 and we also set the figure size to 15 8. 01:41.880 --> 01:44.700 And now we can run any of the code 01:47.550 --> 01:53.730 and here we can see the scatter plot for the H and the fact column and actually use a single dot to 01:53.870 --> 01:55.380 set us a passenger. 01:55.380 --> 01:57.790 So for example here we have a peasant. 01:57.870 --> 02:06.360 That is the 64 year old and the peasants are paid actually about two hundred sixty or seventy a ticket 02:06.360 --> 02:07.270 price. 02:07.380 --> 02:12.900 And so we have five peasants this year where we have an on missing values in the fair and each column 02:13.290 --> 02:18.980 one dot or one scatter a year and apparently both a very first inspection. 02:19.350 --> 02:26.730 It's quite hard to detect yet any relationship So our hypothesis was that the older the peasants the 02:26.730 --> 02:28.380 higher the ticket price. 02:28.410 --> 02:34.830 So if our hypothesis holds true then we would actually expect to get us from the bottom left to the 02:34.860 --> 02:35.620 top right. 02:36.090 --> 02:42.300 But this is actually hard to confirm visually and in this case you would need to perform some more advanced 02:42.420 --> 02:49.560 statistical analysis so let's move on with our plot and actually we can determine the color of the US 02:49.590 --> 03:00.200 and we can do this with the parameter C and we can pass flight sample red and that's also a Pyramid 03:00.220 --> 03:06.180 where we can change the type of the dots are the shadows here and this is called marker 03:08.880 --> 03:12.720 and the default setting is that that year we have a kind of a round dot. 03:12.750 --> 03:17.070 So the default setting here is this kind here. 03:18.390 --> 03:29.420 So we can see nothing has changed but we can also change the marker for example to the OR to uppercase 03:29.430 --> 03:38.050 X or to a lower case x and then we get here actually. 03:38.060 --> 03:43.820 Different scatters and also we can change the size of the scatter scene with the paired meter. 03:43.830 --> 03:48.900 As far for size and t we can pass in. 03:48.970 --> 03:51.180 So for example let's try to find. 03:52.280 --> 03:58.500 And we can also increase the size by increasing the integers for example to 50. 04:00.170 --> 04:03.700 And now we have four x three here a larger scatters. 04:03.740 --> 04:05.590 So now let's go back here to 20th 04:10.510 --> 04:17.110 and that's actually one more future where we can transform our two dimensional scatter plot into a three 04:17.110 --> 04:24.280 dimensional scatter plot and we can do this by passing here to the color perimeter another column of 04:24.280 --> 04:25.390 the data frame. 04:25.390 --> 04:28.120 And if you pass here these left column 04:32.690 --> 04:39.110 then we get here for the dots of the scatters the different colors depending whether the corresponding 04:39.110 --> 04:42.080 passenger survived or not survived. 04:42.080 --> 04:48.440 So for example the peasant so here has a white color and white color XY means here a zero in these survival 04:48.470 --> 04:52.430 column which means that this passenger did not survive. 04:52.430 --> 04:56.460 So in the end we can say that here we have a three dimensional graph. 04:56.510 --> 04:59.560 So the first dimension is here the x axis AIDS. 04:59.990 --> 05:04.480 Then we have the bi axis fair and we have the color. 05:04.490 --> 05:08.400 It gives us information about the the passengers survived or not. 05:08.900 --> 05:16.250 And we can also change the color style here of the survive column and this the pyramid color map. 05:20.480 --> 05:26.260 And if we a look at the map plug the page then here we can see some color style. 05:26.270 --> 05:32.560 So we have a video this my inferno Mama and some more here. 05:32.690 --> 05:40.120 So let's get to see a video of this and we possibly read this here as a string 05:42.900 --> 05:44.190 and it's running the sir. 05:45.150 --> 05:49.160 And here we can see that we have successfully changed the color map. 05:49.530 --> 05:56.180 So now yellow stands for survived and the dark color for not survived. 05:56.190 --> 06:00.920 So again in this last column we only have two different values zero and one. 06:00.930 --> 06:06.340 And therefore we have only dots scatters that are dark or yellow. 06:06.360 --> 06:10.470 But let's try also here the column peak class 06:16.170 --> 06:21.510 and here we can see now that we have dots scatters with three different colors. 06:21.510 --> 06:24.400 So we have for the third class the yellow. 06:24.450 --> 06:27.440 Then we have for the second class green. 06:27.630 --> 06:32.210 And for the first class we have here kind of a dark color. 06:32.370 --> 06:38.230 And here we can see visually that the first class hasn't just paid to the highest ticket prices. 06:38.250 --> 06:42.110 So we have fear the dark shadows or dark plots here at the top. 06:42.240 --> 06:48.240 Then we have the second class with the green dots here in the middle and then at the bottom we have 06:48.240 --> 06:51.760 actually the yellow dots for the third class. 06:52.140 --> 06:57.580 And actually the relationship between H and P classes the less intuitive. 06:57.810 --> 07:04.680 However one could to conclude actually that the very young passengers here are more likely to be in 07:04.680 --> 07:13.120 the second or third class and the older passengers are actually more likely to be first class passengers. 07:13.140 --> 07:13.560 All right. 07:13.560 --> 07:18.150 This was the video on scatter plots and I hope to see you also in the next video by.