WEBVTT 00:00.120 --> 00:00.960 Welcome back. 00:01.280 --> 00:07.520 We generated this table having two columns size and price. 00:07.560 --> 00:12.120 Now we need to introduce something called reshaping data. 00:12.160 --> 00:18.400 Our current data is a combined and concatenated array. 00:18.560 --> 00:21.200 This is the form of our data. 00:21.200 --> 00:26.040 And we use it to make this table and plot. 00:26.080 --> 00:32.480 If we if we need to plot on a graph using Matplot library. 00:32.480 --> 00:42.200 But here we need to send our data to TensorFlow in order to process it, and later for training our 00:42.240 --> 00:42.760 model. 00:42.800 --> 00:47.720 So we're going to do something called reshaping data. 00:47.760 --> 00:59.080 Reshaping is essential because machine learning algorithms expect features to be in a 2D format, where 00:59.120 --> 01:03.880 each row is a sample and each column is a feature. 01:04.040 --> 01:05.840 Please pay attention to this. 01:05.960 --> 01:11.320 Each row is a sample and each column is a feature. 01:11.360 --> 01:17.790 Before we have a 1D array 1.5, 2.02.5. 01:17.830 --> 01:28.510 After we have a 2D array 1.5, 2.0 and 2.5, and the others are filled with other columns. 01:28.510 --> 01:31.590 So we need to transform 1D array. 01:31.790 --> 01:37.510 Or we we need to combine the 1D arrays into two d array. 01:37.750 --> 01:38.270 Okay. 01:38.470 --> 01:42.870 So here we are going to prepare the data for TensorFlow. 01:43.070 --> 01:48.990 Start with x equals to data the data that we created before. 01:49.150 --> 01:52.870 So we are referring to those data. 01:53.030 --> 01:56.990 We need to access from this data the size. 01:57.150 --> 02:07.470 So get the column and pay attention for the name of the column because it's case sensitive.values.re 02:07.710 --> 02:08.390 shape. 02:08.550 --> 02:11.230 This is the function that we're going to use. 02:11.470 --> 02:14.270 And I'll use minus one. 02:14.310 --> 02:18.070 The first parameter and the second parameter is one. 02:18.110 --> 02:21.710 Also we're going to create a variable called y. 02:21.950 --> 02:26.060 And here y equals to Data price. 02:26.100 --> 02:27.900 I'm accessing this column. 02:27.900 --> 02:34.780 Price values dot reshape and again passing minus one and one. 02:34.980 --> 02:39.100 Now let's learn about the reshape function. 02:39.140 --> 02:50.940 The reshape function is very important because we convert the 1D array into a 2D array with one column. 02:50.940 --> 02:54.100 So here we are reshaping our data. 02:54.100 --> 02:55.820 This is a 1D array. 02:55.860 --> 03:00.540 I need to create it as a 2D array with one column. 03:00.700 --> 03:12.620 Okay so this is our reshaping process I'm converting 1D array into 2D array and making the data ready 03:12.660 --> 03:14.060 for TensorFlow. 03:14.180 --> 03:22.020 Again guys, reshaping is essential because machine learning algorithms expect features to be in 2D 03:22.060 --> 03:22.580 format. 03:22.580 --> 03:28.540 So I need to, uh, to pass features into the format. 03:28.540 --> 03:35.260 I don't need to pass arrays with 1D, I need arrays with 2D dimensions. 03:35.300 --> 03:39.450 Okay, so this is how we do reshaping. 03:39.490 --> 03:42.570 I'm converting this array. 03:42.970 --> 03:45.130 I'm accessing the size array. 03:45.370 --> 03:52.450 Getting the values and then reshaping them with one column and 2D array. 03:52.690 --> 03:53.250 Okay. 03:53.530 --> 03:55.610 So this is the magic. 03:55.810 --> 03:59.570 And by the way this is a very important thing. 03:59.810 --> 04:02.810 Then what is minus one. 04:02.850 --> 04:09.090 Minus one means automatically calculates this dimension. 04:09.090 --> 04:14.370 And one makes this dimension have one column. 04:14.410 --> 04:14.970 Okay. 04:15.170 --> 04:20.330 So minus one automatically calculate this dimension. 04:20.530 --> 04:24.690 And one makes this dimension have one column. 04:24.730 --> 04:25.250 Okay. 04:25.490 --> 04:32.330 So take whatever number of rows we have and make it one column. 04:32.610 --> 04:38.730 This is what we are what we told the reshape function. 04:38.770 --> 04:44.130 This is the reshaping process and it works fine.