WEBVTT 00:00.080 --> 00:01.560 -: In this video, you're gonna learn 00:01.560 --> 00:03.870 about something called interactive tables 00:03.870 --> 00:05.460 inside of ChatGPT. 00:05.460 --> 00:07.140 Interactive tables are a new feature 00:07.140 --> 00:09.210 that allow you to dynamically update 00:09.210 --> 00:11.790 and reference tables inside of ChatGPT, 00:11.790 --> 00:13.800 both online on the chat application 00:13.800 --> 00:15.600 and also on the desktop application. 00:15.600 --> 00:17.400 Let's have a look at how this works. 00:17.400 --> 00:19.800 I've got some sample data that we're gonna be using. 00:19.800 --> 00:22.170 So I'm gonna drag in these two tables and see what happens. 00:22.170 --> 00:24.090 So I'm just gonna drag in these two CSVs, 00:24.090 --> 00:28.050 put these in and what happens is ChatGPT, when we hit enter, 00:28.050 --> 00:29.880 it's gonna automatically render out 00:29.880 --> 00:31.590 these interactive tables. 00:31.590 --> 00:33.270 Now, what's interesting about this 00:33.270 --> 00:36.180 is ChatGPT can reference specific columns. 00:36.180 --> 00:38.790 It can even decide to join the tables. 00:38.790 --> 00:39.780 If you have a look at the bottom, 00:39.780 --> 00:42.030 you'll see that ChatGPT is analyzing. 00:42.030 --> 00:43.200 If you click into that, 00:43.200 --> 00:46.380 you'll see that it's loading in the specific data 00:46.380 --> 00:48.750 before it can start doing extra operations. 00:48.750 --> 00:52.260 So first, we're loading the additional employeedata.CSV, 00:52.260 --> 00:55.920 and we're also loading the sample.employeedata.CSV, 00:55.920 --> 00:57.540 and then it's printing the head, 00:57.540 --> 00:59.520 and now it's analyzed the data 00:59.520 --> 01:00.930 and it's giving a little bit of a summary 01:00.930 --> 01:04.200 and it's asking that maybe we wanna merge those data sets. 01:04.200 --> 01:06.960 So it's decided that it wants to merge these data sets, 01:06.960 --> 01:08.310 and you can see that after that, 01:08.310 --> 01:11.430 then it also generates a new interactive table. 01:11.430 --> 01:14.370 Also, you can download these or expand these tables 01:14.370 --> 01:15.990 to hit the entire screen. 01:15.990 --> 01:17.760 When you expand it, then obviously you get 01:17.760 --> 01:20.550 this expanded view and you've also got the ability 01:20.550 --> 01:22.740 to ask questions on the right hand side. 01:22.740 --> 01:24.870 Now I'm gonna go back and collapse it. 01:24.870 --> 01:26.370 Now one thing that can be really cool 01:26.370 --> 01:29.130 is ChatGPT automatically gives you hot actions 01:29.130 --> 01:31.260 or quick actions that allow you to do things. 01:31.260 --> 01:34.440 So for example, we could create a salary distribution chart. 01:34.440 --> 01:37.350 So if I click this, then what happens is ChatGPT 01:37.350 --> 01:41.700 underneath the hood is running Python code on your behalf, 01:41.700 --> 01:43.230 and then it's actually calculating 01:43.230 --> 01:45.060 and showing different types of charts. 01:45.060 --> 01:47.850 So you can see here it's used some adblock lib code 01:47.850 --> 01:49.980 in Python, and if we close this down, 01:49.980 --> 01:52.710 you'll see this actually created a salary distribution. 01:52.710 --> 01:55.530 So other things we could do is we could say 01:55.530 --> 01:56.940 what is the mean age? 01:56.940 --> 01:59.583 So we could say, what is the mean age? 02:00.750 --> 02:04.770 And then ChatGPT will also run additional Python code 02:04.770 --> 02:06.660 on side of that, so you can see it's worked out 02:06.660 --> 02:09.480 that the mean age is 27.8 years. 02:09.480 --> 02:11.340 We can click into the view the analysis, 02:11.340 --> 02:14.970 and it's used a panda's package on the merged data column 02:14.970 --> 02:17.670 in square brackets and it's got the dot mean, right? 02:17.670 --> 02:21.180 So underneath the hood, when you upload CSV files directly 02:21.180 --> 02:24.660 to ChatGPT, you are actually getting some Python operations 02:24.660 --> 02:26.370 that are running in a sandbox environment 02:26.370 --> 02:28.560 to do those types of calculations for you. 02:28.560 --> 02:30.450 Now the great thing about these interactive tables 02:30.450 --> 02:32.490 is you can easily download these as charts, 02:32.490 --> 02:34.860 so you'll see if I download this as an output chart, 02:34.860 --> 02:36.660 then I get a really nice output chart 02:36.660 --> 02:38.010 that I could share with my boss. 02:38.010 --> 02:40.230 So we could also get it to generate lots and lots 02:40.230 --> 02:41.610 of different types of charts. 02:41.610 --> 02:44.040 One other thing we might wanna do is, if you see here, 02:44.040 --> 02:47.070 we've got the emails, we have the age, 02:47.070 --> 02:49.050 and we've got all of the employees. 02:49.050 --> 02:51.240 One thing we might want to get is, for example, 02:51.240 --> 02:55.923 the mean salary, so let's say what is the mean salary? 03:00.600 --> 03:01.920 All right, so the mean salary 03:01.920 --> 03:04.020 of all the employees is $68,000. 03:04.020 --> 03:05.760 Let's go and have a look and just spot check. 03:05.760 --> 03:07.290 That does look about right. 03:07.290 --> 03:08.670 The other thing I want to show you 03:08.670 --> 03:12.780 is you can also specifically reference different columns. 03:12.780 --> 03:15.570 So for example, we could say, if we click on this column, 03:15.570 --> 03:17.400 notice how in the bottom, 03:17.400 --> 03:21.000 you see that ChatGPT has the hire date column selected. 03:21.000 --> 03:25.350 We could then say filter by this column for 2018. 03:26.730 --> 03:29.640 And then because we've referenced this specific column, 03:29.640 --> 03:32.100 it knows that we're probably 03:32.100 --> 03:34.500 after filtering the data by the hire date. 03:34.500 --> 03:36.990 So you can see now if we clicked on the operation, 03:36.990 --> 03:39.270 it's filtering by the date of 2018, 03:39.270 --> 03:41.820 and then we get a different table that's rendered out. 03:41.820 --> 03:45.480 And so that's a great way to use these interactive tables. 03:45.480 --> 03:47.760 You can reference specific columns, 03:47.760 --> 03:51.120 and you can then use that in your ChatGPT at the bottom, 03:51.120 --> 03:52.860 and you can also reference multiple columns 03:52.860 --> 03:54.540 by holding the shift character as well. 03:54.540 --> 03:56.460 So if you hold shift, you can see 03:56.460 --> 03:59.280 that now ChatGPT is referencing three columns 03:59.280 --> 04:00.810 or four columns or five columns, 04:00.810 --> 04:02.940 so that's a really great way for you to limit 04:02.940 --> 04:05.310 the amount of data that you specifically want 04:05.310 --> 04:08.490 or just telling ChatGPT to focus on specific columns. 04:08.490 --> 04:10.230 So a couple of interesting points here. 04:10.230 --> 04:12.510 You get into those nice hot actions at the bottom, 04:12.510 --> 04:13.860 you can reference columns, 04:13.860 --> 04:16.050 you can download the interactive tables, 04:16.050 --> 04:18.240 and you can even join data sets together, 04:18.240 --> 04:20.160 and underneath the hood, everything's running 04:20.160 --> 04:22.080 using a Python sandbox environment 04:22.080 --> 04:23.640 so that it can scale to thousands 04:23.640 --> 04:25.690 or hundreds of thousands of rows for you.