Small talk with Daughter on US Election

just had a small talk with Tanya on US election, she was super angry and there was a big demonstration against Trump in her school too

T:
I don’t want him to win
I don’t want him to do well
Or else another racist gets electedMe:

Me:
neither did I
IF he does very badly, he will be impeached;
or at least he will not be reelected in 4 years.
But now that he is, we can keep an open mind.
There is an element of sentiment he is representing: so-called silent majority, that is why most polls were wrong.

By the way, many have praised my social media analysis just before the election, mine was way better than all the popular polls such as CNN.  This is not by accident, this is power of big data and high tech in the information age:

Final Update of Social Media Sentiment Statistics Before Election

with deep NLP and social media, we can pick up sentiments way more reliable and statistical than the traditional polls, which usually only call 500 to 1000 for opinions to hope they represent 200 million voters.  My mining and analysis are based on millions and millions of data points.  So in future we have to utilize and bring the automatic NLP into things like this as one important indicator of insights and public opinions and sentiments

So in future, we have to utilize and bring NLP into things like this as one important indicator of insights and public opinions and sentiments.

T:
daddy
you’re amazing
Your technology is amazing

Me:
I got lots of compliments for that, but yours mean the most to me.

What happened in the election as I had been tracking using our NLP sentiment tool was:

1. Clinton was clearly leading in the period after the recording scandal of Trump and before the FBI started reopening Clinton’s email case: Big data mining shows clear social rating decline of Trump last month.

2. Clinton has always been leading in Spanish speaking communities and media, but that did not seem to be sufficient to help revert the case:  Trump sucks in social media big data in Spanish.

3. The event of FBI re-opening the email investigation gave Clinton the most damage: Trump’s scandal was cooling down and the attention was all drawn to Clinton’s email case so that the sentiment has a sharp drop for Clinton (【社煤挖掘:大数据告诉我们,希拉里选情告急】)

4. When FBI finally reissued a statement that there was no evidence to charge Clinton only 2 days before the election, time was too short to remedy the damage FBI did in their first event of reopening the case: my big data tracking found that there was some help but not as significant (【大数据跟踪美大选每日更新,希拉里成功反击,拉川普下水】).

5. Then just before the election, I did a final update of the big data sentiment tracking for the last 24 hours versus last 3 months, and found that Trump had a clear leading status in public opinion and sentiments, so I decided to let the world know it although at the point most everyone believed that Clinton was almost sure to win.

T:
Oh my god dad your machine is the smartest tracker on the market
Dad your system is genius
This is exactly what media needs
You should start your own company
This is amazing
I think this would be the planets smartest machine

Me:
I do not disagree, :=)It was a tight competition and with good skills, things could turn different in result.  In terms of popularity votes, they are too to be statistically different, so anything at the right timing could have changed the result.

It was in fact a tight competition and with good skills, things could turn different in result.  In terms of popularity votes, they are too to be statistically different, so anything at the right timing could have changed the result.

On retrospect, FBI did a terrible thing to mess up with the election:
they reopened a case which they did not know the results
just 10 days before the election which made a huge difference.
On the other hand, the recording scandal was released too early
so that although it hurt Trump severely at the time, yet it allowed FBI to revert the attention to Clinton

In future, there should be a strict law disallowing a government agency
which is neutral politically by nature to mess up with an election within a time frame, so Trump’s winning the case to my mind has 80%+ credit from the FBI events.
What a shame

 

[Related]

【社煤挖掘:川普的葛底斯堡演讲使支持率飙升了吗?】

【社煤挖掘:为什么要选ta而不是ta做总统?】

Big data mining shows clear social rating decline of Trump last month

Clinton, 5 years ago. How time flies …

【社媒挖掘:川大叔喜大妈谁长出了总统样?】

【川普和希拉里的幽默竞赛】

【大数据舆情挖掘:希拉里川普最近一个月的形象消长】

欧阳峰:论保守派该投票克林顿

【立委科普:自动民调】

【关于舆情挖掘】

《朝华午拾》总目录

发布者

liweinlp

立委博士,自然语言处理(NLP)资深架构师,Research Director, Beyond AI.前 Principle Scientist, jd-valley, 主攻深度解析和知识图谱及其应用。Netbase前首席科学家,期间指挥研发了18种语言的理解和应用系统。特别是汉语和英语,具有世界一流的解析(parsing)精度,并且做到鲁棒、线速,scale up to 大数据,语义落地到数据挖掘和问答产品。Cymfony前研发副总,曾荣获第一届问答系统第一名(TREC-8 QA Track),并赢得联邦政府17个小企业创新研究的信息抽取项目(PI for 17 SBIRs)。立委NLP工作的应用方向包括大数据舆情挖掘、客户情报、信息抽取、知识图谱、问答系统、智能助理、语义搜索等等。

发表评论