WEBVTT 00:00.770 --> 00:03.800 Welcome to the third step of our final project. 00:03.800 --> 00:10.850 And in the last video we rebuilt marriage passport folio and at first glance the past performance looks 00:11.030 --> 00:12.970 pretty terrific here. 00:12.980 --> 00:19.520 However in the past marriage has not closely monitored the performance of our portfolio and based on 00:19.580 --> 00:23.360 the impressive result over the total twenty two years period. 00:23.360 --> 00:30.560 She may be overestimates the performance and in a very first meeting she is set for investment periods 00:30.560 --> 00:32.450 of six years or longer. 00:32.450 --> 00:39.830 I never lost money with my portfolio and the question is here is she right and this is actually the 00:39.830 --> 00:43.560 perfect application case for a return triangle. 00:43.580 --> 00:50.760 We have still imparted our healthcare index and in a next step we change the frequency of our healthcare 00:50.870 --> 00:54.220 index from daily to annual frequency. 00:54.770 --> 01:02.600 And of course we can do the service of the every sample method and we have here and yellow and the for 01:02.600 --> 01:06.760 each and every period for each and every year we want to have the very last. 01:06.800 --> 01:13.730 Trading Prize and by running the sale here we would get a panda series and we can also convert the panda 01:13.730 --> 01:17.620 series to a data frame with the method to frame. 01:17.690 --> 01:25.680 And finally we save the data frame and the available annual so that's at the annual Values for our new 01:25.680 --> 01:33.210 index starting with 100 at the end of nineteen hundred ninety six and over eight hundred here. 01:33.420 --> 01:40.440 And so the column label as health care here and we can change the column label to for example price. 01:40.740 --> 01:43.980 So that's more meaningful here. 01:43.980 --> 01:46.100 And then we add another column called return. 01:46.650 --> 01:53.490 And here we calculate the lock return or the annual lock return of our prices. 01:53.490 --> 01:59.910 And of course we can do this here with the number pi method and Peter lock and we divide each price 01:59.940 --> 02:05.120 by the immediately preceding price of the preceding year. 02:05.160 --> 02:11.550 So let's do this here and then we actually drop the very first row where we have an A values in the 02:11.550 --> 02:14.530 return column and several up here. 02:15.150 --> 02:19.620 So here we have the prices and uh the lock returns. 02:19.770 --> 02:24.720 Starting from nineteen hundred ninety seven and two thousand and eighteen. 02:25.050 --> 02:31.300 And that sum total twenty two years or twenty two periods and we can also determine this uh with uh 02:31.350 --> 02:39.390 the index attribute and then getting the size of our index and we say actually 22 and the variable years 02:41.640 --> 02:49.730 and then the next step we create a list starting with 22 until 1 and we assign the variable windows. 02:49.770 --> 02:56.100 So this is here a list comprehension. 02:56.300 --> 03:01.630 And then finally we create the 22 new columns uh for our annual data frame. 03:01.880 --> 03:06.050 And these columns are actually rolling statistics. 03:06.200 --> 03:13.700 So we have the rolling mean return for Windows starting from twenty two years until only one year and 03:13.760 --> 03:20.980 we can do this here with a for loop and we iterate over the windows list starting with 22. 03:21.350 --> 03:26.960 So we're creating first of all the column with the column label twenty two years. 03:26.960 --> 03:33.170 And here we calculate the rolling mean return with a window of twenty two years. 03:33.350 --> 03:37.280 And that's what we exactly doing for all elements in the windows list. 03:37.340 --> 03:41.610 So let's run the sale here and let's have a look again at our annual data frame. 03:41.660 --> 03:49.520 So still we a fair price and a return as columns and we have a year rolling mean returns twenty two 03:49.520 --> 03:56.620 years until a one year and finally we dropped the columns the price and return and we save for the new 03:56.620 --> 03:58.870 data frame and the rabbit triangle 04:02.330 --> 04:05.640 and now we are finally prepared to plot our return triangle. 04:06.020 --> 04:13.580 And we can do this with a seaborne heat map and we pass our triangle data frame to S.A. that heat map. 04:13.940 --> 04:15.980 And it's always a bit tricky to determine. 04:15.990 --> 04:23.540 We men we max and center but with returns that always make sense to our faith center of 0 percent. 04:23.630 --> 04:26.350 So no positive or negative return. 04:26.360 --> 04:30.230 And we sat here we max to 15 percent. 04:30.230 --> 04:36.050 So we have deep green fourth that's where we have never it's return of 15 percent or more. 04:36.410 --> 04:40.940 And we determined we mean as being minus 10 percent. 04:40.940 --> 04:42.770 So let's create here the plot 04:49.470 --> 04:54.070 so that's the return triangle for our customized index. 04:54.300 --> 04:57.810 And we can see here that is only very little red. 04:58.200 --> 05:04.670 So here we have said during the dot com crisis some reds and also he and the financial crisis. 05:05.250 --> 05:09.870 But the overall performance of our index remains pretty terrific here. 05:10.260 --> 05:12.570 So a lot of green here. 05:12.570 --> 05:20.910 And the final question is do we find any a period with six years or longer that the average return is 05:20.940 --> 05:29.180 negative or in other words do we find any investment period of six years or longer where Mary lost money 05:29.200 --> 05:30.630 with her portfolio. 05:30.780 --> 05:40.470 And obviously if you go here to nine years then we can see an average return of minus 0.01 percent. 05:40.800 --> 05:49.220 And it's actually no surprise that this is the nine year period from 2001 to 2009 including uh the dot 05:49.220 --> 05:52.550 com bubble and also the financial crisis. 05:52.650 --> 05:58.650 But as a summary we can say that Mary was not right and she kind of well estimated the performance of 05:58.650 --> 06:06.050 our portfolio and with this we are finishing with part three and we will continue in part for us. 06:06.060 --> 06:07.680 So hope to see there by.