WEBVTT 00:00.080 --> 00:01.040 Welcome back. 00:01.080 --> 00:06.240 We learned about Standardscaler and why it's very important. 00:06.480 --> 00:10.320 Now let's run this cell and everything should be fine. 00:10.480 --> 00:17.320 Also, we can check the TensorFlow version by using TensorFlow version. 00:17.480 --> 00:18.760 Let me run again. 00:19.040 --> 00:19.760 There we go. 00:19.800 --> 00:27.360 TensorFlow is already installed and all of those libraries are added and imported correctly. 00:27.680 --> 00:33.000 Now let's move to the second step which is loading and exploring the data. 00:33.280 --> 00:41.360 We'll use the classic auto mpg miles per gallon data set from an online website. 00:41.480 --> 00:43.000 This website is 00:43.360 --> 00:50.840 archived.1.edu/ml/machine 00:50.840 --> 00:52.080 learning databases. 00:52.120 --> 00:55.360 Auto mpg auto mpg data. 00:55.400 --> 00:56.600 This is the link. 00:56.720 --> 01:05.200 Also, you can get it from the resource folder or from the notebook you import from the resources folder. 01:05.200 --> 01:14.160 Under each section, you notice there are a lot of data with the cars, the name of cars, number of 01:14.320 --> 01:18.000 uh, the number, the name of the car, the mileage. 01:18.000 --> 01:22.480 This is then the number that we need to, uh, to predict. 01:22.520 --> 01:24.600 We have number of cylinders. 01:24.640 --> 01:35.360 We have, uh, the displacement, horsepower, weight, acceleration, model year, and the origin. 01:35.360 --> 01:42.480 So one refers for American, two refers for European, and three refers for Japanese. 01:42.520 --> 01:42.920 Okay. 01:42.960 --> 01:53.960 This is our data that we're going to work on and to develop our model that learns all of those data 01:53.960 --> 01:58.120 and predict the mileage and mile per gallon. 01:58.240 --> 01:58.720 Okay. 01:59.160 --> 02:01.360 So in order to load this. 02:01.400 --> 02:03.160 Copy the link. 02:03.200 --> 02:06.720 Load the data by using the URI. 02:07.040 --> 02:16.680 So in the special thing in this application and in this tutorial that we're going to use and load real 02:16.680 --> 02:18.480 data from online website. 02:18.480 --> 02:24.120 So this is real data real statistics and real numbers. 02:24.120 --> 02:27.360 So we're going to deal with those real numbers. 02:27.360 --> 02:29.520 So here this is the URL. 02:29.840 --> 02:32.680 Then we need to get the column name. 02:32.680 --> 02:42.880 So column names mpg miles per gallon cylinders displacement horsepower weight acceleration model year 02:43.080 --> 02:44.560 and origin. 02:44.600 --> 02:54.200 Here we need to read the CSV file from a URL into a pandas data frame with a specific parsing options, 02:54.200 --> 02:58.040 as we did in the previous sections in the previous application. 02:58.240 --> 03:04.160 We collect our data and store it inside the data frame. 03:04.360 --> 03:13.320 So to do that we need to start with row data set equals to PD dot read csv. 03:13.560 --> 03:17.160 Here we have six parameters. 03:17.160 --> 03:19.400 Let me start with the URL. 03:19.840 --> 03:25.800 The path to the CSV file could be a web URL or a local file path. 03:25.960 --> 03:34.400 Also, you can upload your data sample and the data CSV file here in the folders and files for this 03:34.400 --> 03:35.080 notebook. 03:35.280 --> 03:41.080 Also, you can receive the link of your data exactly like we did here. 03:41.080 --> 03:45.120 So I'm passing this URL here in the first parameter. 03:45.320 --> 03:48.400 The second parameter is names. 03:48.760 --> 03:52.200 This provides custom column names. 03:52.200 --> 04:00.280 Instead of using the first row of the file, column names should be a list of strings containing your 04:00.280 --> 04:02.000 desired column names. 04:02.000 --> 04:08.210 So here we have the we have nine columns. 04:08.210 --> 04:13.890 So I should specify nine names of our columns. 04:14.010 --> 04:24.370 The third parameter name, underscore values, treats the question mark as a missing or non-available 04:24.370 --> 04:25.250 value. 04:25.290 --> 04:33.290 Any question mark in the data set will be converted to an Nan not available or not assigned. 04:33.290 --> 04:40.690 So if we scroll down, you notice that in this row we have the question mark. 04:40.690 --> 04:44.570 Also there are a lot of missing. 04:44.970 --> 04:47.250 So if you scroll down here we go. 04:47.490 --> 04:49.890 This is another question mark. 04:49.890 --> 04:57.290 So those question marks will be dealt with as Nan as not available. 04:57.330 --> 05:07.970 The fourth parameter comment backward slash t treats stab characters as comment indicators. 05:08.250 --> 05:11.850 Any text after a tab character will be ignored. 05:11.890 --> 05:13.850 Treated as comment. 05:13.890 --> 05:19.930 The fifth parameter is Sep equals to this empty space. 05:19.970 --> 05:26.730 Uses a single space as delimiter or separator between columns. 05:27.010 --> 05:32.370 The file is space separated rather than comma separated. 05:32.530 --> 05:41.730 So instead of separating the columns by comma, we're going to use a single space as a separator between 05:41.890 --> 05:42.690 columns. 05:42.730 --> 05:50.410 The last parameter is skip initial space equals to true skips any white space that appears immediately 05:50.410 --> 05:52.330 after the delimiter. 05:52.370 --> 05:56.010 Helps clean up extra space in the data. 05:56.050 --> 05:58.250 Okay, now this code. 05:58.570 --> 06:08.170 Use custom column names from column names or array Treat the question mark as missing data. 06:08.210 --> 06:17.450 Ignore that t comment here, for example, as a comment, and parse the values separated by space and 06:17.490 --> 06:22.050 handle the extra space before 6.0 properly. 06:22.050 --> 06:29.770 This is a commonly used for reading formatted data set, where the default CSV settings don't apply 06:29.810 --> 06:30.770 perfectly. 06:30.770 --> 06:37.930 And you can see we use pandas to read the CSV file using the readcsv method. 06:37.970 --> 06:42.610 Now dataset equals to row dataset dot copy. 06:42.810 --> 06:47.330 This line of code creates a copy of the data frame. 06:47.330 --> 06:50.090 What it does creates a deep copy. 06:50.130 --> 06:53.250 Makes a complete independent copy of row data. 06:53.290 --> 07:01.730 Set the new dataset variable points to the separated object in memory, and the changes made to dataset 07:01.770 --> 07:05.610 will not affect the row data set and vice versa. 07:05.650 --> 07:12.650 To print the data set shape we use and to print the data set. 07:12.690 --> 07:14.330 The complete data set. 07:14.370 --> 07:22.490 We use data set if we need to get only for, uh, the first ten numbers, we use the head function and 07:22.490 --> 07:25.650 pass the value of the first columns that we need. 07:25.850 --> 07:28.410 Run the cell and here we go. 07:28.650 --> 07:30.690 This is our data set shape. 07:30.690 --> 07:36.050 We have 389 rows with eight columns. 07:36.170 --> 07:36.690 Okay. 07:36.890 --> 07:44.250 So this is our data set 398 rows times eight columns. 07:44.250 --> 07:52.610 This is a big and a moderate and uh, and real data set that we're going to.