WEBVTT 00:00.550 --> 00:06.050 We have already learned how to down sample time series data with the read sample method and whenever 00:06.050 --> 00:11.810 we down sample data from higher to lower frequency we have to chain an aggregation method. 00:11.810 --> 00:18.830 So first of all we import pandas and then again we are working with the stocks data set and we select 00:18.890 --> 00:20.370 the close column. 00:20.780 --> 00:26.690 So that's the close or the stocks data frame and apparently here we can see on the left hand side that 00:26.690 --> 00:32.140 we have daily data and let's assume that you want to down sample to annual frequency. 00:32.480 --> 00:40.100 And the question is always how to aggregate all timestamps of a year so never it's we have 252 timestamps 00:40.130 --> 00:41.300 and prices. 00:41.300 --> 00:47.190 So for each trading day and actually we have quite some options how to aggregate them to one price. 00:47.670 --> 00:49.480 So let's have a look here. 00:49.630 --> 00:53.810 And first of all we could take the very first price of each year. 00:54.050 --> 00:56.420 So to say the open price. 00:56.420 --> 01:00.420 So we use the urine sample method on the stocks data frame. 01:00.590 --> 01:07.790 We select annual frequency and as a matter of taste you pass the period to the kind parameter and then 01:07.790 --> 01:08.490 we change here. 01:08.510 --> 01:14.890 The first method so let's have a look at the year 2010. 01:14.890 --> 01:16.700 For Apple we fear the price. 01:16.780 --> 01:18.740 Thirty point five seven. 01:18.970 --> 01:22.110 And this is the very first price in the year. 01:22.110 --> 01:29.920 So on the fourth of January then next we could also take the annual maximum price by changing the max 01:29.920 --> 01:30.490 method 01:34.250 --> 01:43.530 so here for example for the year 2010 the maximum price was forty six point four nine U.S. dollars and 01:43.530 --> 01:51.340 we can also take the annual minimum with the main method and here we are for Apple for example a twenty 01:51.340 --> 01:54.980 seven point forty three in the year 2010. 01:55.090 --> 01:59.040 And finally we can also take the very last price in the year. 01:59.440 --> 02:02.670 So typically it's on the thirty first of December 02:05.980 --> 02:13.250 and obviously here are the very last trading price in 2010 for Apple was forty six point 0 8 so this 02:13.250 --> 02:16.520 for aggregation methods are commonly used in finance. 02:16.730 --> 02:18.640 So let's have a look again here. 02:18.650 --> 02:25.990 We have uh the very first price which is also in the finance industry called the Open price. 02:26.120 --> 02:29.240 Then we have the maximum price soar so high price. 02:29.390 --> 02:32.900 Then we have the minimum price which is the low price. 02:32.960 --> 02:38.650 And finally the last price which is also called to the close price. 02:38.660 --> 02:44.630 So essentially we have you open high low close and finally we can also take the average where you are 02:44.630 --> 02:46.310 with the mean method 02:50.110 --> 02:54.970 so these are essentially four or five commonly used aggregation methods. 02:55.100 --> 03:01.870 And now let's assume that you want to have more than one aggregation in a data frame and we can actually 03:01.870 --> 03:10.750 do this by chaining here the ACT method and then the parent uses that we can pass a list with all aggregation 03:10.750 --> 03:11.690 methods. 03:11.770 --> 03:18.260 So let's assume that we are interested in the open price and the close prices for the year. 03:18.430 --> 03:25.830 Then we can pass here first and last for the method the first and last so let's have a look here. 03:27.660 --> 03:33.300 And here we have actually a multi indexed in the columns and on the outer and next level we have the 03:33.300 --> 03:39.540 stock take us and actually for each and every stock we have uh the first price for the open price and 03:39.540 --> 03:42.360 the last price the close price. 03:42.360 --> 03:49.150 So for example the very first trading price in 2010 for Apple was thirty point five seven. 03:49.260 --> 03:53.430 And over the year had actually moved to forty six point 0 8. 03:53.430 --> 03:57.360 And we can further customize some aggregation with so the act method. 03:57.510 --> 04:04.290 So let's assume that for the Apple stock we are only interested that in open prices for Boeing we are 04:04.290 --> 04:11.280 interested in the closed prices and uh for this we want to have the minimum values and the maximum values 04:11.760 --> 04:18.210 then we can pass the array mapper dictionary to the ACT method and actually as the kicker we need to 04:18.210 --> 04:19.550 pass the column label. 04:19.590 --> 04:27.450 So here the stock take us and s value is the desired aggregation function some methods. 04:27.450 --> 04:28.490 So let's have a look here. 04:31.250 --> 04:38.240 So here we f for APA the first price of the year following the last price and the for this near the 04:38.240 --> 04:40.090 minimum at the maximum price. 04:40.190 --> 04:47.020 And also here we have a multi index in the columns so thanks for watching and I'll see you also in the 04:47.020 --> 04:48.190 next video by.