WEBVTT 00:00.750 --> 00:07.380 In this video we learn how to apply the methods stick and unstick so whenever we have data frames with 00:07.440 --> 00:14.770 a mighty index stick and unstick are actually helpful tools to reorganize the data frame and reverse 00:14.790 --> 00:15.270 in a minute. 00:15.270 --> 00:22.030 What I mean here and first of all we import pandas and we are working with the some of them big state 00:22.060 --> 00:22.390 said 00:25.190 --> 00:30.950 so here we are for the first five rows of our data frame and now let's assume that we want to group 00:30.980 --> 00:33.920 our data frame by the columns country and medal. 00:34.430 --> 00:38.260 And for each group we want to count the amount of matters. 00:38.360 --> 00:44.390 By doing so we are creating a brand new data frame and we are signing the variable medals by country. 00:44.420 --> 00:49.700 So let's do this and let's have a look here at our data frame. 00:51.800 --> 00:52.460 So here we are. 00:52.460 --> 00:57.800 For each country the amount of gold Zuber and bronze medals. 00:57.800 --> 01:04.930 So for example for Algeria we have eight bronze medals five gold medals and two silver medals. 01:04.940 --> 01:10.050 So this is for all editions from eighteen hundred ninety six to 2012. 01:10.130 --> 01:17.060 And apparently here on the left hand side we have a mighty index and the outer index level is actually 01:17.060 --> 01:17.390 here. 01:17.400 --> 01:21.740 Country and the inner index level here we have the medals. 01:21.740 --> 01:29.010 Bronze silver and gold and doesn't see it are actually a pen serious because we only have one column. 01:29.220 --> 01:38.100 So this year the amount of medals and again here we can index or slice our pandas serious with the mighty 01:38.100 --> 01:45.040 index and we can do this with the local operator and we can perform later based indexing. 01:45.060 --> 01:52.290 So for example if you want to slice the outer index and we only want to select the United States so 01:52.290 --> 02:02.180 we can do this here we can see that we have over 1000 bronze medals over 2000 gold medals and one thousand 02:02.180 --> 02:08.870 two hundred seven medals and we've already learned how to simultaneously slice for the outer in next 02:08.870 --> 02:11.060 level and in the next level. 02:11.510 --> 02:18.470 So for example we could slice for the country United States and only the amount of gold medals we can 02:18.470 --> 02:19.860 do this. 02:20.090 --> 02:28.340 And actually here we should create the brackets so we have fewer United States for our outer index and 02:28.340 --> 02:36.960 then for our that index we want to select the gold medals. 02:36.970 --> 02:38.360 So now let's run this out. 02:39.310 --> 02:46.630 And here we only get the amount of gold medals that athletes from the United States won in the Summer 02:46.630 --> 02:47.530 Olympic Games. 02:47.530 --> 02:50.030 So two thousand two hundred thirty five. 02:50.200 --> 02:56.620 And we can also check the shape of our mighty and next panda serious matters by country with the shape 02:56.650 --> 02:58.380 attribute. 02:58.630 --> 03:04.380 So we FBA three hundred forty seven rows and apparently only a one column here. 03:04.690 --> 03:11.130 And now we can apply here the unstick method on our pandas serious medals by country. 03:11.500 --> 03:13.180 And let's simply have a look. 03:13.180 --> 03:14.100 What we get here 03:22.700 --> 03:29.240 and apparently we actually switched here our in that index level of matter and we actually created Tier 03:29.240 --> 03:32.920 3 new columns and created here a data frame. 03:32.930 --> 03:39.820 So now we have a data frame with the index country and we have three columns bronze gold and silver. 03:40.400 --> 03:46.270 So before bronze gold and silver were part of the inner index levels. 03:46.430 --> 03:51.380 And by applying the unstick method we switched the inner index level. 03:51.680 --> 03:54.040 And for each value in our in index level. 03:54.050 --> 03:56.130 So we have gold silver and bronze. 03:56.150 --> 03:58.040 We created a brand new column here. 03:58.040 --> 04:04.190 So we have the bronze column the gold column and the silver column and it's values and our new data 04:04.190 --> 04:07.100 frame here we have the amount of metals. 04:07.100 --> 04:14.670 So for example here for Afghanistan we have two bronze medals and by undertaking the index level we 04:14.670 --> 04:16.710 actually created here two new value. 04:16.710 --> 04:20.820 So Afghanistan gold medals and Afghanistan silver medals. 04:20.820 --> 04:26.970 And we have yet any and or missing values because here in our original and stick to pen the seriously 04:27.090 --> 04:33.540 mighty index we do not test for the group Afghanistan gold and Afghanistan silver and by UN sticking 04:33.540 --> 04:39.810 to the index level medals we're actually creating the groups Afghanistan gold and Afghanistan Silver 04:40.740 --> 04:41.510 and the same. 04:41.510 --> 04:43.530 We actually have here for all countries 04:47.060 --> 04:52.850 and we can also handle here the any end values but we will do this later about first of all let's go 04:52.850 --> 05:02.460 you're inside the brackets of the unstick method and that's press shift tab and this pyramid level. 05:02.460 --> 05:07.490 And here we can define which level off our mighty index we want to unstick. 05:07.530 --> 05:10.720 And by default to the inner index levels unstick. 05:10.740 --> 05:15.240 So here in this example the metal a column here the metal index 05:18.930 --> 05:25.400 so level equals minus 1 means we are unstuck in the index level at position minus 1. 05:25.410 --> 05:26.390 So an easy words. 05:26.400 --> 05:30.170 Actually the index level at the most the right hand side here. 05:31.080 --> 05:33.420 But uh we could also change this law. 05:33.420 --> 05:35.010 Let's do this here. 05:35.010 --> 05:37.620 So the default setting is minus 1. 05:38.580 --> 05:40.670 And by doing so we are sticking here. 05:40.680 --> 05:41.890 The medal and next. 05:42.000 --> 05:46.580 But we could also say that we want to unstick here the country index. 05:46.590 --> 05:51.940 And if we have a look at buff so the country indexes at index position minus 2. 05:52.050 --> 05:53.600 So medalists at minus 1. 05:53.600 --> 06:00.380 So the most right index and consequently country is at an exposition of minus two 06:05.240 --> 06:10.420 and now we are getting a different data frame and on the left hand side US index we have our medals 06:10.430 --> 06:17.360 bronze gold and silver and we actually transformed the country index as being the column C of our data 06:17.360 --> 06:17.750 frame. 06:17.750 --> 06:20.970 So far each country here we are getting a column. 06:20.970 --> 06:25.120 So for example the Afghanistan column or the Algeria column. 06:25.400 --> 06:28.070 So however this is here not really a good structure. 06:28.070 --> 06:31.270 So let's go back to a level a minus one here. 06:32.600 --> 06:40.480 So here we have our three columns bronze gold and silver and we still can see here the missing values 06:40.480 --> 06:43.720 for the groups that we created here with the unstuck method. 06:43.720 --> 06:49.450 So for example gold and silver medals for Afghanistan and to this another parameter here. 06:49.450 --> 06:57.190 So let's have a look here called full value and here we can define the value that should replace them 06:57.220 --> 06:58.640 and ends. 06:58.840 --> 07:01.150 And here in our example it should be zero. 07:01.150 --> 07:05.540 So there are no gold medals and no silver medals for Afghanistan. 07:05.650 --> 07:14.040 And I've already done this here so we are undertaking our medals by country pen that serious and we 07:14.040 --> 07:21.960 are undertaking it by the medals index or by the index and we are filling any and values by the value 07:21.950 --> 07:22.560 0. 07:22.890 --> 07:27.570 And then we are reassigning or overriding our variable medals by country. 07:27.600 --> 07:35.830 So let's do this here and let's have a look here at our brand new medals by country data frame. 07:35.830 --> 07:41.980 So yeah the first five rows and on the left hand side we have Afghanistan with two bronze medals zero 07:42.010 --> 07:44.980 gold medals and zero silver medals. 07:44.980 --> 07:48.550 And we can also check here the shape of our new data frame. 07:48.550 --> 07:54.190 So here we have one hundred forty seven rows centre three columns and this is actually one of the main 07:54.190 --> 07:56.740 benefits of the unstick method. 07:56.770 --> 08:02.300 So it changes the shape of our data frame our pandas serious from long format. 08:02.950 --> 08:09.310 So here we can see we have a long format that means we have many rows and only few columns so here we 08:09.310 --> 08:11.390 have 347 rows. 08:11.500 --> 08:15.660 So far each combination of country and medal we have a row. 08:16.090 --> 08:22.230 And by using the unstick method we are transforming the format from long to white slow. 08:22.230 --> 08:26.240 Now here we have a white format. 08:26.350 --> 08:31.870 So actually we have less rows and more columns compared to our long formats here 08:38.030 --> 08:42.830 and this is extremely useful in particular for data presentation purposes. 08:42.830 --> 08:49.150 So the right format is in many cases the more comfortable format with an improved readability. 08:49.160 --> 08:55.770 So this is quite here easy to read and understand and we can also rearrange here the order of our column. 08:55.770 --> 09:02.330 So first of all you want def code to then serve and then bronze and we can do this by selecting the 09:02.330 --> 09:08.500 columns gold silver and bronze in the order that we want to have and then be reassigned over right here 09:08.560 --> 09:11.080 our existing data frame medals by country. 09:11.150 --> 09:20.960 So let's do this and let's check again here the first five rows so yeah we have gold silver and bronze. 09:20.960 --> 09:28.400 And now we can sort our new data frame and we want to sort our new data frame by the code to serve and 09:28.400 --> 09:32.390 the bronze column and therefore we are passing here a list. 09:32.390 --> 09:38.930 Gold silver and bronze to the perimeter by and with the ascending power meter we can define the order 09:38.960 --> 09:45.150 for us three columns here and of course here for our purposes we want to have a descending order. 09:45.170 --> 09:52.300 So starting from the highest values and by passing year more than one column to the by power meter the 09:52.310 --> 09:54.770 sequences here actually are important. 09:54.770 --> 09:57.480 So the first hot hierarchy is actually gold. 09:57.540 --> 10:01.650 Then the second hierarchy serve and the third hierarchy bronze. 10:01.730 --> 10:07.820 So only if two countries suffer the same amount of gold medals then these countries are sorted by the 10:07.820 --> 10:09.440 silver column. 10:09.860 --> 10:16.040 And only if two or more countries have the same amount of gold and silver medals then these countries 10:16.040 --> 10:18.460 are sorted by the bronze medals. 10:18.590 --> 10:24.950 And last but not least if you want to override our data frame and safe our sorting we have to set the 10:24.960 --> 10:29.570 in-place perimeter to true. 10:29.580 --> 10:31.800 So now let's again check here. 10:31.800 --> 10:39.420 Our sorted data frame medals by country in the first 10 rows and no surprise we have here in the first 10:39.420 --> 10:44.430 row United States with the two thousand two hundred gold medals. 10:44.430 --> 10:53.040 Then we have the former Soviet Union and Great Britain Italy Germany and so on and now we can also visualise 10:53.070 --> 11:01.880 our results and therefore we import met plot lip as peyote and we are using the Seabourn style. 11:03.270 --> 11:07.920 And we actually want to create a pop plot for the ten most successful countries here. 11:07.950 --> 11:13.890 So we are plotting methods by country the first 10 rows and we are creating a bomb plot and we are using 11:13.890 --> 11:21.790 the fixed size twelve eight and as X Labor we have country X Y labor matters and the title is medals 11:21.810 --> 11:22.700 per country. 11:22.800 --> 11:24.960 So let's try this out here 11:31.010 --> 11:35.030 and here we have a bomb plot of the medals of the United States. 11:35.060 --> 11:38.100 So in blue we have the gold medals over 2000. 11:38.120 --> 11:43.370 Then we have the silver medals and green and the bronze medals and red. 11:43.610 --> 11:49.130 And we actually have this for the ten most successful countries here. 11:49.150 --> 11:55.510 All right so now last but not least I also want to introduce the step method so let's again have a look 11:55.510 --> 11:58.180 here at our data frame medals by country. 11:59.260 --> 12:00.660 So let's cross here out. 12:00.700 --> 12:02.860 And this is our data frame. 12:03.670 --> 12:08.950 And by using the stick method we are actually reversing our unstick method. 12:09.490 --> 12:16.360 So far with that the unstuck method we switched the inner index medal with gold silver bronze from being 12:16.360 --> 12:19.030 an inner index level to columns. 12:19.030 --> 12:23.520 And now with these test method we can recreate the inner index level medal. 12:23.530 --> 12:24.650 So let's do this year 12:32.100 --> 12:32.550 and again. 12:32.550 --> 12:34.320 Here we have our mighty index. 12:34.320 --> 12:39.870 So we have country s at the outer and next level and we have medals on the index level. 12:40.230 --> 12:42.910 And again you know we have a pen that serious. 12:42.990 --> 12:45.610 The values here the amount of medals. 12:45.870 --> 12:49.120 And now again we can unstick here our mighty index. 12:49.140 --> 12:50.400 So let's do this 12:53.800 --> 12:57.100 and again we are getting our data frame in the right format. 12:57.100 --> 13:02.640 So here we have the format in the index every better as the columns here of our new data frame. 13:02.650 --> 13:04.680 So go to the bronze. 13:04.690 --> 13:05.170 All right. 13:05.170 --> 13:11.290 So this was then an interaction on the unstick and the stick method and they are very helpful when we 13:11.290 --> 13:15.640 want to change the format of a data frame with a mighty index. 13:15.640 --> 13:22.290 And in particular love group by operations where we in many cases create multi indexes the methods stack 13:22.290 --> 13:26.700 and unstick very frequently used methods here. 13:26.700 --> 13:30.960 All right we are finished now with this video and I hope to see in the next one by.