WEBVTT 00:01.050 --> 00:04.180 Welcome to the second part on arithmetic operations. 00:04.320 --> 00:10.920 And in this video we will add subtract multiply and divide a single scalar value to a column and also 00:10.920 --> 00:12.460 to a whole data frame. 00:12.690 --> 00:17.250 And by doing so we are creating new data and new features for our data analysis. 00:17.790 --> 00:25.410 So again here we are working with our Titanic dataset and we have still imported pandas as PTA and still 00:25.410 --> 00:26.730 we have few on the right hand side. 00:26.730 --> 00:33.000 Our new column number for ladies and we all know that the tragic disaster happened in the year nineteen 00:33.000 --> 00:38.730 hundred and twelve and actually for most of the passengers we know the eights and nineteen hundred and 00:38.730 --> 00:39.120 twelve. 00:39.120 --> 00:45.890 So here we have our H column and at the point in time when the disaster happened for example the passenger 00:46.080 --> 00:53.580 with the index labor to was 26 years old and by combining these two information so we know that the 00:53.580 --> 00:58.590 disaster was in nineteen hundred and twelve and we have for the ages in nineteen hundred and twelve 01:00.020 --> 01:03.660 and therefore we can calculate to the approximate year of birth. 01:03.660 --> 01:11.510 So plus minus one year and we can do this by subtracting the h in nineteen hundred and twelve from nineteen 01:11.510 --> 01:12.500 hundred and twelve. 01:12.500 --> 01:15.420 So let's do this year. 01:16.080 --> 01:21.540 So for example the very first passenger was twenty two years old in nineteen hundred and twelve and 01:21.540 --> 01:28.110 therefore his approximate year of birth was eighteen hundred and ninety plus minus one and next extra 01:28.100 --> 01:30.960 we can do the same with the sub method. 01:31.010 --> 01:36.920 So first of all we take the H column and then from the H column we subtract nineteen hundred and twelve 01:37.870 --> 01:43.720 and by doing the other way around us here we are getting negative fair use and we can turn these negative 01:43.720 --> 01:47.800 values to positive by multiplying all the values by minus one. 01:48.220 --> 01:53.380 So we are actually changing your mind method after the sub method and actually we shouldn't have any 01:53.380 --> 02:01.590 problems with and a values in the H column because here at the beginning of our section so he'll be 02:01.590 --> 02:04.820 replaced all and a radius by the mean. 02:04.870 --> 02:13.030 So this will see are actually not necessary and it's anyway the default setting and let's have a look 02:13.030 --> 02:14.410 here at our new data frame. 02:14.410 --> 02:16.200 So here we created the new column. 02:16.200 --> 02:20.370 Year of birth A.B. F. On the right side. 02:20.410 --> 02:22.940 The approximate year of birth. 02:23.170 --> 02:25.600 All right and also here we have our Fifth Column. 02:25.600 --> 02:31.420 And let's assume that the currency of the far column is the U.S. dollar and that we want to convert 02:31.450 --> 02:33.790 the U.S. dollar into a euro. 02:33.790 --> 02:39.490 And let's assume that the current fixed rate so the currency exchange rate is one point one. 02:39.850 --> 02:42.850 So for each euro we get one point one dollar. 02:43.720 --> 02:49.630 And in order to convert here our dollar farce we have to divide our fiat column by the exchange rate 02:49.630 --> 02:50.520 one point one. 02:51.340 --> 02:52.500 And we can do this. 02:52.510 --> 02:53.040 This the. 02:53.080 --> 02:54.360 This method. 02:54.400 --> 02:59.310 So we have the method for addition we have the sub method for subtraction. 02:59.350 --> 03:01.720 We have the my method for multiplication. 03:01.720 --> 03:07.750 And we have the diff method for division and therefore we are taking here the fact column and we divide 03:07.780 --> 03:15.230 the column by the scale of value f x rayed and here we have to define the scale of value as x rayed 03:15.250 --> 03:17.250 which is the one point one. 03:17.630 --> 03:20.260 And we are creating a new column Euro fair. 03:20.660 --> 03:24.520 And this is exactly the Fair column divided by the exchange rate. 03:24.530 --> 03:26.460 So let's do this. 03:26.750 --> 03:27.550 And that's again. 03:27.550 --> 03:29.540 Have a look here at our new data frame. 03:30.590 --> 03:34.020 So here we have added the euro Fair column. 03:34.130 --> 03:40.690 So you have the very first passenger paid to seven point two five dollar and by dividing the dollar 03:40.690 --> 03:43.140 fare by the dollar euro exchange rate. 03:43.150 --> 03:48.790 We are getting here a fair in euro terms on euro currency. 03:48.790 --> 03:54.280 All right so in the last session and in this session you created some new columns for our Titanic data 03:54.280 --> 03:54.790 frame. 03:55.060 --> 04:01.000 And now it's the time to drop some columns so that we can proceed our further analysis with a smaller 04:01.030 --> 04:07.810 and leaner data frame and actually want to drop the columns S.P. in part because here we created the 04:07.810 --> 04:10.040 new column number for latest. 04:10.330 --> 04:17.380 So we no longer need the columns the best P in parts and we can also delete the deck column because 04:17.380 --> 04:23.320 here we have over 600 missing values and also a year of birth and the euro Fair column. 04:23.320 --> 04:29.080 These are interesting and nice to know but for the time being we are also deleting these columns here 04:29.410 --> 04:30.370 and we can do this. 04:30.370 --> 04:36.400 This is the drop method and we have to set the in-place parameter to true as always. 04:36.520 --> 04:43.120 And then let's have a look here at our new data frame and here we only have two columns survived the 04:43.150 --> 04:48.100 peak last sex age fair embarked and the number of total relatives. 04:48.100 --> 04:52.390 All right now let's go back to our state's data frame and let's have a look here. 04:53.080 --> 04:58.470 So here we have the daily say it's for the say it's man Stephen Mike Andy and Paul. 04:59.440 --> 05:05.360 Now let's assume that the states men have daily fixed costs of five units. 05:05.410 --> 05:08.740 So in each day the states men have some states and revenue. 05:08.770 --> 05:13.080 But on the other hand side they have fixed costs of five units. 05:13.150 --> 05:20.940 So let's define you the variable fixed costs and all we want to calculate and determine the net savings. 05:20.950 --> 05:22.130 So the states. 05:22.180 --> 05:27.220 After subtracting the fixed costs and therefore we have to select here the columns. 05:27.220 --> 05:33.450 Monday to Friday so we all rows and all columns except the last two columns here. 05:33.460 --> 05:37.450 So by doing so we're creating a slice of our say instead of frame. 05:37.450 --> 05:39.620 Let's have a look here. 05:39.670 --> 05:42.960 So we have we have the slice of the say it's. 05:43.150 --> 05:48.040 And from each dailies say we now want to subtract the fixed costs of five. 05:48.310 --> 05:54.430 And we can do this with the sub method and we are passing the scale of value of fixed costs. 05:54.430 --> 05:58.030 And we have here for Stephen on Thursday and in value. 05:58.150 --> 05:59.310 We have to define that. 05:59.350 --> 06:03.940 We want to fill all in the end values with the value 0. 06:03.970 --> 06:05.320 So let's do this here 06:09.230 --> 06:16.000 and here we can see for example for Stephen on Thursday the net say after abstracting the fixed costs 06:16.000 --> 06:23.270 of 5 for sexually minus five because we treat it here the any end value as being zero. 06:23.300 --> 06:30.340 And before here we had the percentage bonus for say it's men and they actually had the different bonus. 06:30.340 --> 06:34.440 So Stephen had 12 percent Mike 15 and 10. 06:34.640 --> 06:38.350 And let's assume that all of these states men have the same percentage bonus. 06:38.360 --> 06:41.440 So 10 percent on the on the sales. 06:41.750 --> 06:48.250 And now again for each daily sale we can calculate the dollar bonus for the salesmen. 06:48.260 --> 06:51.590 So first of all we are slicing our sales data frame 06:54.390 --> 07:00.360 and then we are applying the model method and multiplying here each element by the percentage bonus 07:00.410 --> 07:01.270 0.01. 07:01.710 --> 07:05.520 And again here we want to treat all any end values as being zero 07:08.760 --> 07:12.090 and here we have further daily bonuses for the salesman. 07:12.090 --> 07:15.480 So for example Mike receives for his saves. 07:15.480 --> 07:18.670 On Wednesday seven point four bonus. 07:18.690 --> 07:19.070 All right. 07:19.080 --> 07:21.650 So now let's again check our daily sales. 07:21.960 --> 07:28.320 So we are slicing our sales data frame except the last two columns with the percentage bonus and the 07:28.320 --> 07:35.880 total weekly bonus here and that's assume that the bonus metric is not as simple as here. 07:35.880 --> 07:40.890 So by just multiplying a percentage bonus of 10 percent to I'd say it's data. 07:41.340 --> 07:44.770 But let's assume now that for each 10 say it is. 07:44.820 --> 07:47.310 So we have a lot size of 10. 07:47.310 --> 07:50.940 The salesman gets a bonus of one point two five. 07:51.240 --> 07:57.210 So for example if one salesman has on one day say it's off line you get nothing. 07:57.720 --> 08:01.440 And if the FSA is off 10 you get to one point two five. 08:01.530 --> 08:04.080 And if he has to they say it's off 15. 08:04.080 --> 08:05.660 You also get one point two five. 08:05.700 --> 08:13.260 And also with the 19 and then the case that a salesman has a daily sales of 20 then he get two times 08:13.260 --> 08:15.150 one point two five. 08:15.150 --> 08:19.000 So this is the the more complicated metric here. 08:19.020 --> 08:23.030 And for doing this we are having the very head for method flawed. 08:23.070 --> 08:30.410 So first of all here let's again slice here for our daily sales and then we are applying the flautist 08:30.510 --> 08:33.990 method and the flautist method actually divides. 08:33.990 --> 08:39.180 Here are the elements of for example 35 for by the lot size 10. 08:39.180 --> 08:41.740 So this is a three point four. 08:42.030 --> 08:46.090 And then it's around set down to the next integer so three. 08:46.110 --> 08:50.620 So we have here for the Monday sales of Stephen the 34. 08:50.640 --> 08:54.660 We have three completed lots with the lot size 10. 08:54.750 --> 08:57.780 And for example on Tuesday we have two completed lots. 08:57.810 --> 09:00.990 And on Wednesday one completed lot and so on. 09:00.990 --> 09:02.220 So we are using here. 09:02.220 --> 09:03.710 The flawed this method. 09:03.720 --> 09:04.770 So let's do this 09:09.460 --> 09:12.840 and first of all of course you have to define what size equals 10. 09:12.840 --> 09:15.580 And bonus pay a lot this one point two five. 09:17.050 --> 09:21.140 And here we can see the amount of completed lots for each daily say it. 09:21.160 --> 09:27.250 So here we have a three for Stephen on Monday two for Tuesday and one for Wednesday. 09:28.090 --> 09:34.130 And as we passed a zero to the pyramid the full value we have on Thursday also zero. 09:34.130 --> 09:40.840 And now we can multiply the daily lots by the bonus pair lot by changing the model method 09:45.490 --> 09:52.690 so he Steven had on Monday three completed lots and he gets a bonus of one point two five for each lot 09:53.020 --> 09:57.350 and therefore his bonus for Monday is three point seventy five. 09:57.490 --> 10:02.680 And we have done this here for all elements in our data frame and now we can also calculate that the 10:02.680 --> 10:05.150 total weekly bonus for let's salesman. 10:05.290 --> 10:13.220 And we can do this by changing or adding the sum method and we want to take this sum for all rows. 10:13.330 --> 10:15.700 So here for Stephen for Mike for India 10:19.350 --> 10:23.450 and again here we are getting the total weekly bonus for all faucets man. 10:23.490 --> 10:30.000 So these are some examples of Eric Medtech operations but pan now serious and also data frames with 10:30.000 --> 10:33.900 a single scalar value and actually no limitations. 10:33.900 --> 10:37.720 So you can make this as complicated as you like. 10:37.830 --> 10:44.370 And the good thing in Python here is that we can apply here multiple operations so flawed of multiplication 10:44.370 --> 10:47.670 and some actually only here within one line of code. 10:47.700 --> 10:49.510 So this is quite comfortable. 10:49.530 --> 10:55.380 And with this we are finished now with this session and I hope to see you also on the next video by.