WEBVTT 00:00.800 --> 00:05.340 Hello everybody and welcome back to our visualization session with Matt Platt lip. 00:05.450 --> 00:09.040 So in most cases the graph tells us more than a bunch of numbers. 00:09.050 --> 00:16.040 And for the human eye a graph is typically more intuitive and understandable and therefore also Python 00:16.040 --> 00:20.870 has an extensive library for viz. called met floodlit. 00:21.050 --> 00:24.700 So that actually no limits on customizing grass that's met floodlit. 00:24.800 --> 00:31.890 And yeah you could actually read em Zamir or read you any graph you might think often pacing and and 00:31.930 --> 00:38.760 yeah it's just a matter of time because the functionality of met plot level is more or less unlimited. 00:40.040 --> 00:46.820 So this session should serve as a short intro that enables you to make the most common plots like line 00:46.820 --> 00:53.570 plots scatter plots and his programs and it should also enable you to make the most common and important 00:53.570 --> 00:56.110 customizations to plots. 00:56.150 --> 01:02.810 So first of all we have to impart the plotting framework of bad plot loop called Pi plot so we import 01:02.810 --> 01:06.080 met plots that pi plot SPL T. 01:06.710 --> 01:12.650 So this is a stand up convention to have the revelation P to you. 01:12.650 --> 01:16.430 And if you work and tip in a notebook we should also have. 01:16.460 --> 01:18.770 The magic function map plot lib in line. 01:19.520 --> 01:21.260 So let's execute here. 01:22.420 --> 01:27.530 And then first of all let's create a list with five elements 1 3 5 7 and 9 01:30.170 --> 01:37.100 and we want to create a line plot of these values and therefore we're using the function period T dot 01:37.130 --> 01:45.620 plot for the line plot and we pass our list to a peer to you dot plot and before we want to execute 01:45.620 --> 01:53.030 and show our line plot we uh also showed you as pictured here dot so let's execute the cell here. 01:54.320 --> 01:56.800 So now here we can see our line plot. 01:56.900 --> 02:00.220 So when looking at the y axis we have one here. 02:00.230 --> 02:06.550 Then we have uh three we have five we have seven. 02:07.240 --> 02:14.500 And we have nine we could also change some the last element here too let's say six 02:18.290 --> 02:19.090 so you in the end. 02:19.090 --> 02:20.610 Then we have six. 02:20.830 --> 02:25.890 And by default peer to dot plot starts with the Sera at the x axis. 02:25.900 --> 02:30.420 So this is actually an interpolated line for the points 0 1 02:33.340 --> 02:37.780 1 3 2 5 02:41.350 --> 02:52.060 3 7 and 4 6 and we can also define x axis value so by default Python assumes integer starting from 0. 02:52.480 --> 02:55.310 And here in this case we have 5 elements. 02:55.320 --> 03:00.560 So from 0 to 5 excluding So this is the default case. 03:00.910 --> 03:09.510 So we get here the same graph but we can redefine the x axis so let's say we want to have x values from 03:09.510 --> 03:10.370 five to 10. 03:10.380 --> 03:13.830 We can do this. 03:14.060 --> 03:17.750 So now the Excel values are starting from five to nine. 03:18.650 --> 03:20.870 So on if we cross up to show 03:24.660 --> 03:31.320 then we can still see the plot here but also the FBI description of plot object and therefore it's best 03:31.320 --> 03:35.600 practice here to have the P.A. so before we plot here 03:38.530 --> 03:40.550 and we can also make a get up not to this. 03:40.570 --> 03:45.420 The function P T dot scatter where we define x value. 03:45.440 --> 03:50.800 So from 0 to 4 encoding and the y values is our list. 03:50.800 --> 03:53.230 So let's see what we get here. 03:53.650 --> 03:59.540 So we get your scatters and the first one has an x y you have zero and then y value of 1. 03:59.860 --> 04:08.370 And the second one has an x value of 1 and the y value of 3 and the third one has a mixed value of two 04:08.370 --> 04:09.590 and a y value of five. 04:09.630 --> 04:10.970 And so on. 04:11.050 --> 04:17.470 How why would we have to keep in mind that the amount of x values must match the length of our via values 04:17.490 --> 04:21.990 here Sophia for example X extend range 7. 04:21.990 --> 04:27.810 So with has seven elements of our sexes and only our list L with five elements. 04:27.840 --> 04:34.100 Then we get an error message so X and Y must be the same size 04:38.830 --> 04:39.220 all right. 04:39.220 --> 04:44.740 So let's move on to a histogram and let's assume that we are standing downtown New York at a street 04:44.770 --> 04:47.590 with a speed limit 50 miles per hour. 04:47.920 --> 04:51.260 And we are measuring the speed of 10000 cars. 04:51.610 --> 04:58.240 And for our purposes we assume that the speed is normally distributed was mean 55 and standard deviation 04:58.240 --> 04:58.810 5. 04:59.650 --> 05:08.080 So we are creating 10000 the normal distributed random numbers with mean fifty 55 and standard deviation 05:08.080 --> 05:08.610 5. 05:08.620 --> 05:15.250 So we are using here a list comprehension and we are assigning the driver speed in New York. 05:15.250 --> 05:18.090 So actually the length of our list should be ten thousand. 05:18.700 --> 05:18.900 Okay. 05:18.910 --> 05:21.770 First of all we have to import the random module. 05:21.790 --> 05:32.500 Then it should work and we can also check some elements here of our list speed New York. 05:32.610 --> 05:35.270 So let's say let's check. 05:35.290 --> 05:38.620 Uh the first 10 elements. 05:40.620 --> 05:41.130 So we uh. 05:41.130 --> 05:44.470 Fifty four fifty five point four and so on. 05:44.520 --> 05:45.470 Forty eight. 05:47.250 --> 05:51.170 And then we want to create a histogram With our 10000 measurements. 05:51.200 --> 05:58.530 And therefore we are using purity hist and we are passing our list speak New York to beauty list. 05:58.560 --> 06:02.170 And we are defining that we want to have 100 pins. 06:02.340 --> 06:05.120 So let's see what we get here. 06:06.150 --> 06:12.300 So here we have a histogram but our measurements so it's no surprise that s uh we created our measurements 06:12.330 --> 06:15.460 with a normal distributed random numbers generator. 06:15.840 --> 06:21.740 We get here kind of a battleship curve with um the mean at fifty five approximately. 06:22.320 --> 06:29.100 And we can also change the amount of pins here we have 100 pins and uh let's see what happens if we 06:29.100 --> 06:31.530 change it to 10. 06:31.590 --> 06:35.980 So now we have only 10 Ben so that's maybe not the perfect solution here. 06:36.150 --> 06:42.650 So maybe if we increase it to 1000 now this is maybe also not the best solution. 06:42.650 --> 06:47.930 So one hundred bins for 10000 measurements would be a good solution here. 06:48.140 --> 06:50.750 And now what we can do we can do many customizations. 06:50.750 --> 06:53.070 So first of all so let's plot again. 06:53.070 --> 07:00.470 Here we plot our histogram we are passing our list speed New York we are having 100 bins we are setting 07:00.470 --> 07:06.170 the table to a New York data we will see later what this will bring us and also we set out a four to 07:06.170 --> 07:09.290 1 and the color of our eyes again to read 07:12.140 --> 07:17.040 let's draw the graph here and it's essentially the same graph as above. 07:17.080 --> 07:19.690 Uh we only changed the color to red. 07:19.730 --> 07:19.880 Yeah. 07:19.940 --> 07:26.000 And then we can make your customization so first of all we can change in the style we use to see Bond 07:25.990 --> 07:26.390 style. 07:26.420 --> 07:31.940 So yeah there are many styles and Matt plotted and it's more or less a matter of taste but style you 07:31.940 --> 07:32.740 want to use. 07:33.920 --> 07:35.660 So if you change it to seaborne 07:39.210 --> 07:43.640 then it looks a bit different a bit more scientific actually. 07:44.160 --> 07:52.710 And we can also change the size of the figure by using the function Piers dot figure and set the fixed 07:52.710 --> 07:57.970 size to let's say 10 6. 07:58.030 --> 07:59.230 So now it's a bit larger. 07:59.230 --> 08:09.120 We can also change here let's say 212 six and we can also add a title to our graph with some functional 08:09.160 --> 08:15.430 purity job title and then we need to define the title as the strings here measure car speed speed limit 08:15.430 --> 08:16.250 50 miles. 08:16.290 --> 08:23.750 Power and what we can also do we can set the font size so the font size could be 15 year 08:27.680 --> 08:35.840 so it's a bit larger than the excelsior and then we can also introduce labels for the Xs so we can introduce 08:35.870 --> 08:44.770 um and label for the x axis with prototype X label and pass the string speed. 08:44.810 --> 08:46.130 So now here we have speed 08:49.140 --> 08:50.290 and uh the same. 08:50.310 --> 08:59.430 We can do with the VI label and for example we can set up the label occurrences number of occurrences 09:00.570 --> 09:05.370 here and we can also plot the vertical lines. 09:05.500 --> 09:10.040 That's the function peer to peer v lines. 09:10.100 --> 09:17.120 So and the purpose is to draw vertical line at the mean value of our collected our measured data. 09:17.450 --> 09:20.430 So yeah let's see a shift tab. 09:20.660 --> 09:22.580 What we need here. 09:22.640 --> 09:24.250 So we need the x value. 09:24.290 --> 09:27.680 And this is the mean value of our lists speed. 09:27.680 --> 09:28.160 New York. 09:28.160 --> 09:35.550 So we take the sum divided by the length and then we also have to define uh the minimum y value and 09:35.550 --> 09:36.870 the maximum y value. 09:36.870 --> 09:46.760 So here we can set it between 0 and 400 and then we can define that the colors of be red and the lines 09:46.780 --> 09:55.150 should be here Darst and the labels would be a mean New York so let's see what we got here 09:58.230 --> 10:05.270 right here you can see a we can see the test line but as our histogram is in transparent Yeah. 10:05.300 --> 10:10.760 We cannot see our dashed line here and therefore we have to change and the out file for our histogram 10:10.760 --> 10:15.650 so alpha equals 1 means there's no transparency of our histogram. 10:15.650 --> 10:24.240 So if you change it to let's say 0 point 5 0 is a gram here is getting transparent and then we are seeing 10:24.240 --> 10:26.240 the vertical line here. 10:26.430 --> 10:29.980 Next we can set the range of taxes. 10:30.120 --> 10:41.420 So here with POTUS access and let's see what we need here so we have to set X-Men X Max Wyman and Y 10:41.420 --> 10:41.840 Max. 10:41.840 --> 10:53.510 So why should be between let's say 0 and 400 and X should be the 3 and yeah 30 and maybe 90 so let's 10:53.510 --> 10:54.140 run here. 11:02.630 --> 11:06.190 And then we can also set the ticks off our axis. 11:06.230 --> 11:14.510 So to say here we have ticks at the 30 40 50 and if you want to take set let's say 30 35 45 then we 11:14.510 --> 11:17.280 have to change that here. 11:17.420 --> 11:26.360 So from 30 to 91 we want to have ticks at every fifth value. 11:26.970 --> 11:33.310 So here we are and the same we can do with them the VI label. 11:34.550 --> 11:37.970 So it makes sense to have some ticks at every 50s. 11:37.990 --> 11:41.970 No so we already have it here 11:49.500 --> 11:52.790 and then we can define if you want to have the grid here or not. 11:52.790 --> 12:03.570 And by default with them the Seabourn style it's enabled but we can also disable it with the force 12:10.110 --> 12:11.770 let's make it for Reagan. 12:15.990 --> 12:20.470 And before we head to our histogram we had the label the new york data. 12:20.490 --> 12:25.260 And then our vertical line the label mean New York. 12:25.260 --> 12:29.840 And to show the labels we have to call the function purity legend. 12:30.270 --> 12:35.740 And in this beauty legend we can define the location of our legend. 12:36.030 --> 12:41.820 So let's enable it here and let's press should tap 12:48.580 --> 12:51.180 so he allocation string so it can pass best. 12:51.190 --> 12:54.630 Or alternatively zero or upper right. 12:54.760 --> 13:03.710 This one here I have center right center right is seven and so on and we also define the font size equals 13:03.760 --> 13:04.840 13. 13:04.910 --> 13:06.470 So let's see what we get here. 13:08.540 --> 13:14.420 So now we have here our legend and we can also make many more plots within one graph. 13:14.450 --> 13:21.690 So let's assume we are going to Boston and also measure 10000 times them the speed in a street with 13:21.690 --> 13:23.240 the speed limit 50. 13:23.990 --> 13:31.220 So let's assume the Boston people are driving faster and on average Uh 60 miles per hour and then 50 13:31.220 --> 13:32.510 miles per hour zone. 13:32.510 --> 13:35.280 And the standard deviation eight. 13:35.450 --> 13:41.480 So let's create other random numbers and start in the list speed Boston. 13:43.010 --> 13:47.210 And then we can also plot the histogram here from Boston. 13:48.350 --> 13:52.280 So Patty Hearst we passed speed Boston. 13:52.310 --> 13:54.840 We set pins to a 100. 13:55.200 --> 13:57.220 Set the label to Boston data. 13:58.050 --> 13:59.840 Alpha equals 0 point 5. 13:59.850 --> 14:01.900 And the color to blue. 14:01.920 --> 14:07.970 And let's see what happens. 14:08.000 --> 14:11.450 Here we have both a histogram some New York and from Boston. 14:11.450 --> 14:12.200 And one graph. 14:12.220 --> 14:19.070 And it makes it very easy to compare and we can also enable here the vertical line at the mean of the 14:19.070 --> 14:20.190 Boston data. 14:22.360 --> 14:22.490 Yeah. 14:22.830 --> 14:28.070 And here we can see on average I mean the Boston people are driving faster than the New York people. 14:28.910 --> 14:33.080 And also with the variability of the Boston data is a bit higher. 14:33.080 --> 14:39.140 So it's a bit more diverse here because of the standard deviation of our Boston data was higher with 14:39.170 --> 14:40.610 eight. 14:40.650 --> 14:41.120 All right. 14:41.130 --> 14:44.490 We are finished now with our shot introduction to Matt Black live. 14:44.490 --> 14:49.980 And if you want to go deeper into details I recommend you to go on the documentation side 14:54.630 --> 14:57.800 so here you can find many examples and tutorials. 14:57.810 --> 15:06.820 And if we have a look at the examples so you can see here the options are actually unlimited. 15:08.770 --> 15:12.310 So there are hundreds or thousands of examples here 15:15.740 --> 15:20.370 so have fun if you like and I hope you enjoyed the session and outtake take by.