1
00:00:00,381 --> 00:00:02,640
Now, let's talk about Amazon Forecast.

2
00:00:02,640 --> 00:00:03,860
And this is a very easy one

3
00:00:03,860 --> 00:00:06,080
because it allows you to do forecasts.

4
00:00:06,080 --> 00:00:07,630
So this is a fully managed services

5
00:00:07,630 --> 00:00:08,750
that will use machine learning

6
00:00:08,750 --> 00:00:10,790
to deliver a highly accurate forecast.

7
00:00:10,790 --> 00:00:13,280
For example, you want to predict the future sales

8
00:00:13,280 --> 00:00:14,510
of a raincoat.

9
00:00:14,510 --> 00:00:16,660
The idea is that it's going to be 50% more accurate,

10
00:00:16,660 --> 00:00:18,450
than looking at the data itself.

11
00:00:18,450 --> 00:00:21,280
And you reduce forecasting time from months to hours

12
00:00:21,280 --> 00:00:22,990
by using a managed service.

13
00:00:22,990 --> 00:00:24,330
So the use cases can be multiple,

14
00:00:24,330 --> 00:00:25,380
whenever you need a forecast.

15
00:00:25,380 --> 00:00:28,340
For example, product demand planning, financial planning,

16
00:00:28,340 --> 00:00:30,110
resource planning, and so on.

17
00:00:30,110 --> 00:00:31,090
So how does it work?

18
00:00:31,090 --> 00:00:33,160
Well, you take your historical time-series data,

19
00:00:33,160 --> 00:00:35,700
for example, and you also add your product features,

20
00:00:35,700 --> 00:00:38,206
prices, discounts, website traffic, store locations,

21
00:00:38,206 --> 00:00:40,010
basically, any kind of data you can

22
00:00:40,010 --> 00:00:41,380
to then enhance your model.

23
00:00:41,380 --> 00:00:43,790
Then you upload this into Amazon S3.

24
00:00:43,790 --> 00:00:46,560
You then start the Amazon Forecast service

25
00:00:46,560 --> 00:00:48,620
which will create a forecasting model.

26
00:00:48,620 --> 00:00:50,570
And you can use that forecasting model, for example,

27
00:00:50,570 --> 00:00:52,010
to say that your future sales

28
00:00:52,010 --> 00:00:56,060
of raincoats are going to be $500,000 next year.

29
00:00:56,060 --> 00:00:57,360
And that's it, super simple.

30
00:00:57,360 --> 00:00:58,920
Whenever you see forecasts at the exam,

31
00:00:58,920 --> 00:01:00,503
think Amazon Forecast.