WEBVTT 00:01.810 --> 00:02.310 Hey, everyone. 00:02.470 --> 00:03.030 Welcome back. 00:03.480 --> 00:06.460 So we declared the tweets from the Twitter server. 00:06.900 --> 00:14.940 Now, before it read every single tweet, let us use this tax blob library and see how to use to get 00:14.940 --> 00:16.960 the sentiment out of a tax. 00:17.520 --> 00:24.420 And for the illustration, as I hope, capital, we variable like Highmore bag cricket player. 00:24.650 --> 00:27.240 I am a good cricket player and I am cricket player. 00:27.630 --> 00:29.270 So that would be one positive sign. 00:29.280 --> 00:32.130 The main one is neutral and one is a negative. 00:32.580 --> 00:39.950 And let's see how it goes with all of this text block library so we can use like a text block. 00:41.950 --> 00:43.570 Taken Boss E! 00:45.380 --> 00:47.670 We will be applying this sentiment. 00:49.620 --> 00:51.590 So if you just execute this much part. 00:53.250 --> 00:56.130 Ops is not defined being defined as ABC. 00:57.930 --> 00:58.320 All right. 00:58.800 --> 01:03.630 So it has given us the information, clarity and subjectivity. 01:04.020 --> 01:07.320 So here the polarity will be zero point sixty nine. 01:07.370 --> 01:11.070 But that is a negative because of this bad word. 01:12.350 --> 01:19.070 If you want to get, let's say, just a polarity, that will be a negative number. 01:19.860 --> 01:20.360 Same way. 01:20.540 --> 01:22.610 Let's do it for me. 01:22.610 --> 01:24.950 Variable, that will be a positive rate even. 01:26.350 --> 01:27.700 So it's said you don't point seven. 01:27.900 --> 01:29.310 That means a positive sentiment. 01:30.180 --> 01:33.880 And the last one, see, that will be a neutral. 01:35.440 --> 01:37.110 So it's a zero point zero. 01:37.800 --> 01:44.980 All right, so now we know how to get the sentiment out of the texts and we know already how to fetch 01:44.980 --> 01:45.910 this data. 01:46.270 --> 01:54.730 So let's just do it every single text or a tweet and supply to this text block to find a sentiment available 01:54.730 --> 01:56.100 in an individual tweet. 01:57.040 --> 01:59.790 So let me just make FuelCell here. 02:02.640 --> 02:10.860 Now, from this whole Twitter sentiment, we are going to get a four parameter like out of Kozin three. 02:11.070 --> 02:12.590 How many of them are positive? 02:12.660 --> 02:13.980 How many of them are negative? 02:14.460 --> 02:15.830 How many of them are neutral? 02:16.380 --> 02:17.180 And what is it? 02:17.410 --> 02:18.510 Average polarity. 02:18.780 --> 02:23.640 So polarities about zero below zero on an average. 02:23.670 --> 02:24.360 So we'll see. 02:24.780 --> 02:29.130 So I defining default radio positive, negative, neutral polarity. 02:29.220 --> 02:32.370 As of now, everything is zero for now. 02:32.520 --> 02:36.680 Let me be clear to every single Twitter tweets. 02:37.170 --> 02:40.710 So that will be tweets, radio and so forth. 02:41.550 --> 02:44.380 Each tweet in. 02:49.820 --> 02:50.180 It's. 02:51.390 --> 02:56.100 Ledger's first name putting this tweet tax. 02:58.060 --> 03:01.570 So it will just this display, the tax associated with individual. 03:03.590 --> 03:04.100 All right. 03:04.230 --> 03:10.350 So you can see all those thousand tweet it has displayed and still it's going on. 03:11.380 --> 03:13.390 So it will take good amount of pain, I guess. 03:14.020 --> 03:15.070 All right, let's finish. 03:16.730 --> 03:16.920 Right. 03:17.620 --> 03:19.660 So these are the housing tweets. 03:21.570 --> 03:24.180 Now we are going to supply this to packs. 03:25.140 --> 03:28.920 Inside this tax block function to get the polarity. 03:29.610 --> 03:34.500 So let me copy of let us use your tax block. 03:35.400 --> 03:39.540 Let me assign it to somebody who let's say analysis. 03:41.370 --> 03:45.030 And next, we'll be finding from this analysis of polarity. 03:46.580 --> 03:48.740 So let's have polarity. 03:49.660 --> 03:50.420 Who will be? 03:51.380 --> 03:53.390 So as of now, the polarity zero. 03:53.420 --> 03:55.700 So we'll just keep upending the polarity. 03:56.550 --> 03:57.980 So that will be a whole. 03:58.450 --> 03:59.720 Is equal to polarity. 03:59.960 --> 04:03.390 Plus, we'll get the polarity for current. 04:03.850 --> 04:04.060 We. 04:05.200 --> 04:12.090 We'll be analysis, not sentiment, not clarity. 04:13.050 --> 04:13.440 All right. 04:14.020 --> 04:17.680 So we grabbed the polarity inside the polarity medium. 04:18.430 --> 04:25.300 Now, based on this polarity, we want to find how many of them said negative, positive and a neutral. 04:26.260 --> 04:28.150 So that will be so simple. 04:28.330 --> 04:34.600 Like, if the polarity is equal to zero, we'll just increment a counter for this neutral. 04:35.560 --> 04:43.630 If the polarity is less than ideal, incremental counter negative, has it given for the positive. 04:44.630 --> 04:44.990 All right. 04:45.160 --> 04:49.240 So if something embedded within the indication, we can just get. 04:50.090 --> 04:50.440 All right. 04:50.500 --> 04:51.280 So let me. 04:51.630 --> 04:52.010 Ranny. 04:54.660 --> 04:56.850 All right, so you can see immediately what that is. 04:57.560 --> 05:03.240 And let me just display how many positive tweets that the pope's. 05:05.600 --> 05:07.230 So says that detail. 05:07.520 --> 05:10.910 So what if something goes wrong, if you try to. 05:11.620 --> 05:15.460 Lastly, let's see, just do it for. 05:17.870 --> 05:18.990 And let me see. 05:18.990 --> 05:20.920 More loosening in legislating. 05:22.560 --> 05:24.400 So earlier, we got those thousand into. 05:24.710 --> 05:26.940 And now you can see we got almost GDP. 05:27.080 --> 05:29.400 So what just goes wrong here? 05:30.650 --> 05:33.020 So you can see we're dealing with here. 05:33.230 --> 05:41.510 Of course, the object that meets once Lucozade reach to the end of our Triplette object. 05:42.050 --> 05:43.190 Now, that is No. 05:43.430 --> 05:49.420 Two, it exists because all those thousand three one time we have already created over. 05:50.060 --> 05:55.050 So what we have to do, we have to again fetch those result inside the coaster. 05:55.700 --> 05:57.550 Let me run this one again. 05:58.590 --> 05:59.150 In No. 06:00.790 --> 06:02.050 Let me run this quick. 06:04.100 --> 06:11.490 Oh, so now you can see it's taking a little bit more time because it is trying to get every single 06:11.500 --> 06:20.340 tweet, all those thousand tweet, finding the polarity and find a counterpart is neutral, negative 06:20.340 --> 06:20.940 and a positive. 06:21.730 --> 06:24.750 Let us display how many positive reviews. 06:25.940 --> 06:28.760 So total 247 positive reviews. 06:33.170 --> 06:34.370 How many negatives did. 06:35.880 --> 06:37.300 176 and negative. 06:38.040 --> 06:43.970 And how many nutrients, 577 nuchal. 06:44.670 --> 06:45.600 What is the polarity? 06:45.670 --> 06:48.350 So it's towards the positive order. 06:48.810 --> 06:49.350 Negative. 06:51.630 --> 06:52.070 All right. 06:52.150 --> 06:54.380 You can see twenty two point zero eight. 06:54.720 --> 06:58.650 But we have to normalize it because we'll just added all those politely. 06:58.680 --> 07:00.030 We've done Hosein scores. 07:00.030 --> 07:00.930 Or we can just simply. 07:02.310 --> 07:02.860 Normalizing. 07:04.140 --> 07:07.260 And let us define one function before Ben. 07:08.250 --> 07:09.630 Finding the percentage. 07:12.000 --> 07:12.420 All right. 07:12.460 --> 07:15.190 So you can see I only find one percentage function. 07:15.380 --> 07:17.890 We're just going to find a percentage between your. 07:19.530 --> 07:21.330 Numerator Hannah denominator. 07:22.200 --> 07:25.020 So here we are just passing the positive. 07:25.090 --> 07:27.270 Total number of use out of whole. 07:28.330 --> 07:30.580 Total number of tweets, which we are read, right? 07:30.700 --> 07:36.180 Same for positive or negative in a neutral and simply for political, Sylvia, just marginally right. 07:36.270 --> 07:39.400 They got a number of tweets, hopes. 07:40.420 --> 07:45.300 We'll go find this dysfunction first and then call this function for all four. 07:45.550 --> 07:47.280 We do it now. 07:47.310 --> 07:48.620 Ledgers format that is in. 07:50.930 --> 07:51.320 All right. 07:51.920 --> 08:00.950 So now if you just display this positive, you can see positives is twenty four point seventy. 08:01.970 --> 08:03.820 That means it is divided by thousand. 08:04.140 --> 08:04.430 Similar. 08:04.590 --> 08:06.830 You will get the result for the remaining three or so. 08:07.910 --> 08:09.860 Now, let's try to display. 08:11.010 --> 08:13.260 Our overall summary of the result. 08:14.200 --> 08:22.450 So here what we have done for reaction of people, for some particular search done by analyzing total 08:22.450 --> 08:28.750 number of priests, I just use the plain statement and displayed rather qualities, did or not. 08:28.780 --> 08:32.260 So if it is zero, we are just going to display his neutral. 08:32.980 --> 08:39.020 Otherwise, it will be a negative and positive respectively, based on polarities, negative or positive 08:39.570 --> 08:40.540 ledges and any. 08:42.660 --> 08:45.230 All right, so the reaction of people on monkey. 08:45.870 --> 08:47.340 By analyzing hosen tweet. 08:47.790 --> 08:49.440 It's a positive reviews. 08:50.220 --> 08:50.580 All right. 08:51.060 --> 08:55.320 Now, let's try to display this result in some sort of paycheck's. 08:56.730 --> 09:01.770 So to display all those results in a paycheck form, we are only just this MacCulloch live legatee and 09:01.770 --> 09:06.390 I hope created this small lutein self matplotlib call that. 09:07.640 --> 09:16.010 We are going to display a positive, negative and a neutral value in some sort of paychecks for me. 09:16.490 --> 09:23.780 And it will give us an indication that out of full paycheck check this yellow part indicates that it's 09:23.780 --> 09:30.380 kind of neutral, really indicate that this much amount of negativeness are left out of those thousand. 09:31.370 --> 09:35.030 And this green part indicates some positive reviews. 09:35.730 --> 09:36.130 All right. 09:36.170 --> 09:41.360 So that is the whole story of this Twitter sentiment analysis.