WEBVTT 00:00.000 --> 00:03.139 -: So attached to this video will be a link to a CSV file. 00:03.139 --> 00:05.310 Make sure to download that file. 00:05.310 --> 00:07.470 It's called students.csv. 00:07.470 --> 00:08.640 If we have a look at this, 00:08.640 --> 00:10.510 and I'm gonna go and open this up, 00:10.510 --> 00:12.240 you can see we have some data about students. 00:12.240 --> 00:14.700 So for example, the first name, the last name, 00:14.700 --> 00:17.670 the gender of a student, the age, the grade level, 00:17.670 --> 00:20.580 and some other things such as what type of degree they took, 00:20.580 --> 00:21.810 their attendance rates, 00:21.810 --> 00:24.360 their extracurricular activities, et cetera. 00:24.360 --> 00:26.640 We're gonna use ChatGPT to analyze this dataset 00:26.640 --> 00:28.200 rather than having to use Excel. 00:28.200 --> 00:29.760 Cool, so the first thing is once you've got 00:29.760 --> 00:31.170 that downloaded, go to ChatGPT. 00:31.170 --> 00:33.030 I want you to click on this plus sign, 00:33.030 --> 00:36.375 click from Upload from computer, go to students.csv. 00:36.375 --> 00:40.650 You can also drag the CSV or any file directly 00:40.650 --> 00:42.360 into the chat box as well. 00:42.360 --> 00:43.740 That's also possible. 00:43.740 --> 00:44.910 Then we're gonna say, 00:44.910 --> 00:48.510 can you find out how many students 00:48.510 --> 00:50.380 there are in this dataset? 00:56.187 --> 00:57.330 And what it will do 00:57.330 --> 01:00.450 is ChatGPT will analyze and write code 01:00.450 --> 01:02.310 and you'll see if you click on this Analyzing, 01:02.310 --> 01:03.840 you can see that it's actually going 01:03.840 --> 01:06.570 and generating some code in Python 01:06.570 --> 01:08.310 and it's returning a result. 01:08.310 --> 01:11.310 And then it shows you there are 50 students in the dataset. 01:11.310 --> 01:14.970 So ChatGPT has a way of being able to run code for you 01:14.970 --> 01:17.160 in Python, execute that code, 01:17.160 --> 01:19.290 and then return the answer of that to you 01:19.290 --> 01:20.490 in natural language. 01:20.490 --> 01:24.720 We could say how many males versus females 01:24.720 --> 01:26.823 are there in the CSV? 01:29.220 --> 01:31.290 And again, it's running some Python code. 01:31.290 --> 01:33.150 You don't always have to show the details of this. 01:33.150 --> 01:35.790 You can hide it, but I quite like to see exactly 01:35.790 --> 01:36.870 what's been happening. 01:36.870 --> 01:38.970 So you'll see here it actually ran into an error. 01:38.970 --> 01:41.220 See there's this KeyError Gender 01:41.220 --> 01:43.554 and then it's decided to print out the columns 01:43.554 --> 01:46.170 so that it can specifically figure out 01:46.170 --> 01:47.820 which columns it needs to query. 01:47.820 --> 01:50.250 It then creates a new specific piece of code. 01:50.250 --> 01:51.083 And you can see here, 01:51.083 --> 01:54.090 it is doing a .value_counts on the gender, 01:54.090 --> 01:55.320 which allows it to find out 01:55.320 --> 01:58.410 that there's these 25 females and 25 males. 01:58.410 --> 02:01.323 If we scroll up and have a look at the dataset, 02:02.640 --> 02:06.060 we could also say which major 02:06.060 --> 02:09.423 has the highest attendance rate? 02:12.210 --> 02:13.500 So now it's deciding to group it 02:13.500 --> 02:17.160 by the major and it's using the attendance rate 02:17.160 --> 02:18.120 and getting the mean of that. 02:18.120 --> 02:20.850 And then it finds the one that has the highest maximum. 02:20.850 --> 02:23.100 So math had the highest attendance rate, 02:23.100 --> 02:23.940 and so you can see 02:23.940 --> 02:26.520 that we don't necessarily need to put this data 02:26.520 --> 02:28.110 directly into Microsoft Excel. 02:28.110 --> 02:30.600 We can just put a CSV directly 02:30.600 --> 02:32.730 inside of ChatGPT and start asking it 02:32.730 --> 02:34.080 natural language questions. 02:34.080 --> 02:35.400 We can even plot bar charts. 02:35.400 --> 02:36.750 So I can say, can you plot 02:36.750 --> 02:41.750 a bar chart of attendance rates, broken down by major? 02:46.320 --> 02:48.750 And now it's decided to use a plotting library. 02:48.750 --> 02:49.980 So this is just a piece of code 02:49.980 --> 02:52.953 that allows it to generate graphs dynamically. 02:54.420 --> 02:55.830 So we'll just wait for that to load. 02:55.830 --> 02:57.690 And here you go, here's your bar chart, 02:57.690 --> 03:00.210 and you can easily download this or change this. 03:00.210 --> 03:01.950 You can click on this Download chart 03:01.950 --> 03:04.080 and then you can download that directly to your desktop. 03:04.080 --> 03:05.580 You can also click on, right click, 03:05.580 --> 03:08.100 and then Save Image As if you're on Mac. 03:08.100 --> 03:10.890 So it's possible for you to not only ask questions 03:10.890 --> 03:11.820 in the natural language, 03:11.820 --> 03:13.800 but get automatic data visualizations 03:13.800 --> 03:16.200 made out of ChatGPT. 03:16.200 --> 03:17.033 As an exercise, 03:17.033 --> 03:19.200 I want you spend two minutes asking ChatGPT 03:19.200 --> 03:21.540 some follow-up questions on this dataset, 03:21.540 --> 03:22.860 and also playing around with 03:22.860 --> 03:25.980 what kind of data visualizations you can get it to create. 03:25.980 --> 03:27.730 Cool, I'll see you in the next one.