1 00:00:08,822 --> 00:00:12,510 Matplotlib has support for several kinds of bar charts. 2 00:00:12,510 --> 00:00:16,520 The most general case, we plot a bar chart by sending in a parameter 3 00:00:16,520 --> 00:00:20,740 of the x components, and a parameter of the height of the bar. 4 00:00:20,740 --> 00:00:22,930 Let's do this with our linear data. 5 00:00:22,930 --> 00:00:26,840 Here I'm going to just create a list of values as long as our linear data list. 6 00:00:28,790 --> 00:00:32,620 Then, we'll send this along with the linear data's bar heights, 7 00:00:32,620 --> 00:00:34,180 to the bar function. 8 00:00:34,180 --> 00:00:37,109 We pass in the width of the bar as 0.3. 9 00:00:45,428 --> 00:00:49,387 To add a second bar, we simply call the bar plot again with new data, 10 00:00:49,387 --> 00:00:53,209 keeping in mind that we need to adjust the x component to make up for 11 00:00:53,209 --> 00:00:54,763 the first bar we plotted. 12 00:01:00,957 --> 00:01:03,271 Of course, we'll throw this bar in as red. 13 00:01:10,890 --> 00:01:14,230 So the plot looks okay, but not great. 14 00:01:14,230 --> 00:01:17,520 All of the x labels are to the far left of the items being described, and 15 00:01:17,520 --> 00:01:19,000 not centered. 16 00:01:19,000 --> 00:01:23,100 Now, we can center them with a given bar, using the align parameter, but 17 00:01:23,100 --> 00:01:25,880 it's a bit frustrating when you have to make multiple bars, and 18 00:01:25,880 --> 00:01:27,260 you have to calculate it yourself. 19 00:01:27,260 --> 00:01:30,950 And it's a pain to go through and add a new data series manually. 20 00:01:30,950 --> 00:01:34,790 And this gets worse, we actually have to manually iterate through and 21 00:01:34,790 --> 00:01:36,080 add in dates if we want them. 22 00:01:36,080 --> 00:01:39,640 Because otherwise we couldn't set the bar left location appropriately. 23 00:01:41,020 --> 00:01:41,750 I'll be honest, 24 00:01:41,750 --> 00:01:45,900 bar charts are one of my least favorite things to create in matplotlib. 25 00:01:45,900 --> 00:01:48,550 I feel like the foundation is there, 26 00:01:48,550 --> 00:01:51,300 but some of the ways I would want to use the API, 27 00:01:51,300 --> 00:01:55,940 such as plotting several series of data in groups across time, are sadly missing. 28 00:01:57,150 --> 00:02:00,470 There are some other nice features of bar charts, though. 29 00:02:00,470 --> 00:02:04,050 For instance, you can add error bars to each bar as well, 30 00:02:04,050 --> 00:02:05,590 using the y-error parameter. 31 00:02:06,680 --> 00:02:11,250 For example, each of our pieces of data in the linear data might actually be a mean 32 00:02:11,250 --> 00:02:14,950 value, computed from many different observations. 33 00:02:14,950 --> 00:02:17,720 So we could just create a list of error values. 34 00:02:17,720 --> 00:02:22,050 I'll import from a random function which can generate some random integers, and 35 00:02:22,050 --> 00:02:25,890 then I'll use a list comprehension to generate this new list. 36 00:02:25,890 --> 00:02:29,505 Then we could just call plt.bar again, and see our new data appear. 37 00:02:38,428 --> 00:02:41,740 We can also do stacked bar charts as well. 38 00:02:41,740 --> 00:02:45,814 For instance, if we wanted to show cumulative values while also keeping 39 00:02:45,814 --> 00:02:50,155 the series independent, we could do this by setting the bottom parameter and 40 00:02:50,155 --> 00:02:53,251 our second plot to be equal to first set of data to plot. 41 00:03:18,071 --> 00:03:22,955 Finally, we can pivot this bar graph into a horizontal bar graph by calling 42 00:03:22,955 --> 00:03:24,825 the barh function instead. 43 00:03:25,900 --> 00:03:30,373 But note that we have to change the bottom to a left and the width to a height. 44 00:03:47,491 --> 00:03:51,355 So those are the basics of charting and plotting in Matplotlib. 45 00:03:51,355 --> 00:03:54,315 I think it's important to realize that there really isn't any magic. 46 00:03:54,315 --> 00:03:57,955 The scripting layer is really a set of convenience functions 47 00:03:57,955 --> 00:03:59,305 on top of the object layer. 48 00:04:00,600 --> 00:04:03,890 Some people share a preference for one or the other, but at the core, 49 00:04:03,890 --> 00:04:06,370 they manipulate the same data elements. 50 00:04:06,370 --> 00:04:09,990 And I don't think you should think of them as being opposed to one another or 51 00:04:09,990 --> 00:04:11,060 duplicative. 52 00:04:11,060 --> 00:04:14,769 They work together, and being able to move back and forth is really important. 53 00:04:16,230 --> 00:04:19,000 We've talked about some fundamental charts in this module. 54 00:04:19,000 --> 00:04:22,760 Scatter plots, line graphs, and bar charts. 55 00:04:22,760 --> 00:04:24,980 There are more, which we'll dig into in the next module. 56 00:04:24,980 --> 00:04:26,180 But these are a good basis. 57 00:04:27,240 --> 00:04:30,050 The assignment for this module is a tough one. 58 00:04:30,050 --> 00:04:34,050 It's meant to show you real world data visualization problems. 59 00:04:34,050 --> 00:04:36,930 You're going to have to lever your knowledge of the Pandas Library from 60 00:04:36,930 --> 00:04:38,300 the first course and 61 00:04:38,300 --> 00:04:42,120 couple this with your new knowledge of Matplotlib to come to an elegant solution. 62 00:04:43,140 --> 00:04:46,660 But the final result should be a portfolio ready piece. 63 00:04:46,660 --> 00:04:49,230 So I'll see you in the forums for discussions and for help.