Random

Kopi Sianida : Are The Police Right OR Is It Just A New Form of The Prosecutor Fallacy?

Pernah dengar tentang Prosecutor Fallacy?

Prosecutor Fallacy mudahnya adalah kesalahan penggunaan konsep peluang ketika ingin menunjukkan peluang seorang tersangka bersalah.

Misal :
Terjadi pembunuhan dan pelaku meninggalkan bukti berupa darah, sedemikian sehingga darah ini ternyata hanya ada 1 di setiap 1000 orang. Lalu ada si X yang darahnya matching dengan darah yang ada di TKP.
Maka oleh prosecutor dibilang wah ini darah cuma ada 1:1000, artinya si X ini peluang tidak bersalahnya cuma 1:1000 alias 0.1% alias peluang salahnya 99.9%.

Waw this prosecutor brought probability to the court!!!

BUT… this logic is sooo far from flawless. Kenapa? Karena penggunaan teorema Bayes di sini definisi peluangnya terbalik.

Peluang 0.1% ini adalah peluang darah matching jika diketahui tidak bersalah {P(A|B)}, sedangkan yang kita ingin tau adalah berapa sih peluang tidak bersalah jika diketahui darah matching {P(B|A)}.

Lalu apa hubunganny ama kasus kopi sianida? Prosecutor fallacy mengajarkan kita jangan sampai kesesatan berfikir ini terbawa ke pengadilan dan akhirnya membawa imbas jelek bagi siapa pun yang terlibat, dalam kasus ini polisi. Punya firasat bahwa si X adalah pelaku ya sah-sah saja, tapi jangan sampai firasat ini nutupin bukti atau keganjilan-keganjilan lain yang ada di TKP.

Kasus Prosecutor fallacy yang terkenal ada namanya The Sally Clark case, di mana Sally dituduh membunuh 2 bayi nya yg baru lahir, hanya karena seorang pakar bilang kalo peluang bayi dalam satu keluarga meninggal karena SIDS adalah 1:73juta, waw artinya peluang kalo Sally yang bunuh anaknya besar sekali. TAPI sayangnya si pakar menghitung peluang ini salah, tapi di tahun 1998 ga ada yang menangkap kesesatan cara mendapatkan 1:73juta ini, akhirnya Sally ditangkap tahun 1999, dan baru dibebaskan tahun 2003 ketika profesor dibidang matematika menghitung ulang peluang ini, Sally meninggal tahun 2007 karena problem psikis, dan jadi alkoholic.

See, this is dangerous!

Sama seperti kasus kopi sianida, si J datang duluan ke TKP, si J yang beliin kopi, si J bayar kopi terburu-buru, jadi pasti si J nih yang bersalah! O…o… This kind of logic is dangerous, tapi yah ga tau juga sih, karena mungkin saja kasus ini sebenernya se-simple itu.

Trus ada yang nyamain kasus ini dengan kasus di Conan, emang sih di Conan pake sianida juga dan emang sih di situ yg bunuhnya yang beliin kopi.

Trus gue inget ada tv series barat jadul taun 2006-an judulnya Boston Legal episode The Verdict, di mana di kasus ini polisi udah yakin banget kalo Scott membunuh seorang wanita hanya karna ada psikiatris yg ngasi rekaman terapi antara dia dan scott, di mana di rekaman itu scott bilang kalo dia mimpi ngebunuh si wanita. Kasus langsung masuk pengadilan. Di sini pengacara scott keren abis, dia bisa “membujuk” semua juri yg hadir bahwa scott ini ga bersalah, karena ada namanya “Reasonable doubt”, ini “bujukan” dari jaksa penuntut dan pengacara scott di akhir cerita :

Jaksa penuntut : Did anybody fall for that? First, he said the husband did it. Then, it was all those other men she was having an affair with. Then, I can’t be sure, but I think he actually seemed to accuse the father, at least briefly. Oh, and let’s not forget Lincoln Meyer, the neighbor. He wanted you to think that he was the killer. And then finally, in the most desperate and transparent attempt of all, he goes after Scott’s mother. And she played along perfectly, didn’t she? “I want an attorney. I assert my Fifth Amendment right.” Gee, that ought to give them all reasonable doubt. The evidence, all of it, points to one person being in the house—the defendant, a sick twisted, psychologically infirm young man, loaded with anger, self-loathing. He confessed to his therapist he had compulsions of killing her. Gee, I wonder who could have done it. If any of you were fooled by the defense lawyer’s ridiculous Perry Mason theatrics, then shame on you.

Pengacara Scott: Shame on you? How ‘bout shame on the police for investigating nobody else but Scott Little? As soon as they saw a video of Scott talking about killing Marcia Hooper, it was “Case Closed” for them. Could they not have looked at the husband? He had motive; he was there; he has no real alibi; he made no attempt to save her. Could they not have looked at Lincoln Meyer, the card-carrying nut-job from next door? He was obsessed with Marcia Hooper. She had to get a restraining order to keep him away. His flowers were found in the room where she was murdered. And yes, Barbara Little. Let’s face it: Love obsessions rarely occur in a vacuum. Could she not have loved him just as much? Her whole life is her son. The idea of him falling into the clutches of another woman—? Why was that house so clean of prints? The killer will typically wipe clean just what he or she has touched. But if she didn’t know what Scott had previously touched, she would have cleaned everything. How can any of us be sure here? As for physical evidence, the only thing the prosecution can establish is that Scott Little made love to Marcia Hooper the night of the murder. Well, he told them that. He came forward, as innocent people tend to do, and volunteered the information. The only reason they arrested him was because of a video and an opinion proffered by an eccentric therapist who runs around telling people he’s kissed movie stars he’s never even met. This can’t turn on his credibility. Talk about reasonable doubt. Reasonable doubt! Come on. It could have been him, could have been him, could have been her. This whole case screams, “Reasonable doubt,” which is typically the case when the police don’t investigate.

Isn’t that line beautiful? Soo smooth.

And let’s not forget this quote from Patrick Jane The Mentalist – No wife should glow at her husband’s funeral“, in this case : “No husband(family/friend) should glow at his wife’s funeral”.

So am i saying the husband did it? No.
Jessica? No.

I don’t know, i’m not an expert in this.

See ya!

Random

Goodbye 2015

Last post in 2015.

Banyak target 2015 yang belom tercapai, seperti diet misalnya kikikikikikikik, mudah-mudahan di 2016 bisa tercapai #ahem.
Target buat nabung yang belom tercapai sama sekali, masi ngangsur duit motor ke abang gue sih, gaji jadi banyak terpotong, belom lagi belanja online mulu nambah koleksi sepatu, dvd, baju, dan buku, belom lagi jajan es krim, sate,martabak wkwkwkwkwk *diet woy!!!*

Beberapa minggu belakangan ini gue lagi intensif chat ama si saricang, obrolan mulai dari yg bener sampe yg menyimpang hahahahahahaha. Mulai dari pake Whatsapp sampe pindah lapak ke Line demi sticker-sticker yang dramatis dan mendukung.

Anyway, do you know Kim Seulgi?
Itu loh yang main di Oh my ghost yg jadi hantunya.
Itu loh yang main di surplus princess yg jadi blogger gila makan.
Itu loh yang main di flower boy next door yang jadi webtoon editor.
Itu loh yang kalo main drama jadi pemeran sampingan mulu tapi selalu ninggalin kesan yang dalam cieee suit suiiiit.

Pas meta ngasi tau kalo dia main drama lain dan jadi tokoh utama, gue langsung Line tusi ngasi tau dia, soalnya kita pernah diskusi (?) soal si Kim Seulgi ini ga bakal pernah jadi pemeran utama, soalnya suara dia ga mendukung untuk jadi tokoh utama yg kerjaannya (biasanya) nangis. Eh ternyata kita salah! Dia jadi pemeran utama di Splash-splash Love yang langsung gue download ngorbanin paket flash di hp, demi Kim Seulgi kikikikikikikik. Sayangnya cuma 2 episod, dan terlalu romantis bikin gue pengen garuk-garuk dinding wkwkwkwkwkwk.

Watch out for Kim Seulgi cuteness!

Wishing you people a great day.
I’ll see you next year. BHAY!

Random

TCash Telkomsel

Sisa pulsa anda 7000, sisa kuota flash 30 Mb. Crap!

Oh iya, kan ada aplikasi MyTelkomsel, bisa beli pulsa online pake CC (Credit Card). Untung MasterCard BNI ada fitur VCN (Virtual Card Number) *bisa minta VCN buat beli pulsa online*.

….. (ceritanya lagi sms banking minta vcn, habis deh pulsa 2ribu)

Buka aplikasi MyTelkomsel, ga bisa login. Bang to the Ke. Bangke.
Login pake token, balik lagi ke home hp, minta token ke *323#, token cuma berlaku 3 menit. Balik lagi ke MyTelkomsel, bisa login.

Mau ngisi nomer CC ternyata ga bisa copas, matek, balik lagi ke menu sms, nyatet nomer CC di notepad komputer. Ketik ketik ketik, send.

Transaksi Gagal!

What? Why? Ternyata CC yang bisa dipake cuma CC beneran, ga bisa yg virtual. Crap!

Oh, wait….. ada pilihan TCash.

Langsung googling cara aktivasi dan ngisi TCash. *800*88#

Aktivasi kelar, buka *800# ngisi saldo TCash via sms banking. Done.

Buka MyTelkomsel lagi. Ngetikin satu-satu nomer akun TCash.

Login Gagal! Ngetik lagi.

Login Gagal! Ngetik lagi.

Token salah! Ngetik lagi.

Token salah! Ngetik lagi.

Dst.. sampe setengah jam kemudian…

Telkomsel kampreeeeeeeeet.

Kuota tinggal 1200 Kb.

Tunggu. Di *800# tadi kayakny ada pilihan isi pulsa.

Balik lagi ke *800#, ketik ketik ketik, masukin pin, done.

Pulsa udah nambah, langsung beli paket Flash.

Uninstall aplikasi MyTelkomsel plus ngasi bintang 1 di Playstore.

Itulah sekilas yang terjadi saat makan siang di kantor tadi.

*****

WOW. TCash!!! Why did i never try this before?! Seriously.

Jadi Tcash bisa diisi via sms banking di *800#. Bisa isi pulsa sekalian. Ngisiin pulsa nyokap udh ga perlu repot ke kios.
Jadi gue sekarang udah ga perlu repot keluar rumah malem-malem cuma buat beli paket internet. Tinggal isi TCash hihihihihihihihi.

Family · Random

Random Monday

Time to move on!!!
Ternyata nyari yang jual full size SUM37 itu susaaaaah banget. Ada jual di lazada tapi hargany 1jutaan, duh bangkrut gue, jadi yak gue balik lagi ke Laneige, ini lebih gampang nyari yang jual full size-nya. Pas banget tadi pagi toner sum37 gue abis, eh sorenya essential skin refiner Laneige nyampe, ini semacam toner lah ya, buat ngangkat sel kulit mati juga. Toner SUM37 memuaskan, sangat memuaskan malah, tapi ya itu, susah nyarinya.

Gue sekalian beli konjac sponge baru, green tea lagi, it smells so good.
Gue juga beli series barat lagi wkwkwkwk, baru dikirim tadi siang, mungkin 3 hari lagi baru nyampe. Lupa deh pesan apa, yg gue inget cuma The mentalist.

Btw beberapa minggu yang lalu gue nemu foto ini di 9gag, dan gue pake buat foto profil di Whatsapp
1439347552
Ini komentar kakak gue, keluarga yang sangat solid wkwkwkwk
Screenshot_2015-08-15-20-55-21
Chat1
Ini chat sama Mimi
Screenshot_2015-09-28-21-29-39
Ini chat sama Hani
chat2

Btw background whatsapp gue pake gambar karya c-cassandra I love her work, visit her tumblr people!!

Udah jam 22.00
I’ll see you when i see you, bye~

Data viz · Tableau

Tableau Web Data Connector

First of all, please pardon my English. English is not my first language.
Second of all, i post this as a reminder for myself, if you find it helpful then good for you 😀

Tableau brings data visualization to a whole new level. You can extract data from any website using Web Data Connector (Tableau 9.1).
Alex Ross in his blog Tableau Junkie posted some examples of his web data connector (facebook, twitter, open weather map) check out his Github to download the code HERE.

After reading Alex Ross blog post, i start wondering, what is the very first step to make a connector? i’m not a programmer, i never use javascript, where do i write this whole code? How do i run the code?
Turns out that the answer is available on Tableau’s online help HERE.

Easy huh? But NOT FOR ME. I don’t even understand how to do the first part, it took me three days to work on it. So here it is.
## Obviously you’re gonna need Tableau Web Data Connector SDK (WDC SDK), download HERE and then I extract the file in C:\WDC SDK (for the rest of this post i assumed you also use the same location)
## To work with WDC SDK you need web server running in your computer (source: Tableau Online Help)
Or you can just install Python (https://www.python.org/downloads/) *i use Pyhton 2* and follow the instruction how to use Python for WDC HERE

BUT if you’re as doomed as i am in programming world, then follow my steps below:

1. Download and Install python (example location C:\Pyhton27) >> finish
2. Go to Computer properties >> Advanced system settings >> Environment Variables.. >> Edit Path (Path is in System variables box) >> type ;C:\Python27\ at the end of variable value (do not delete variable value that already exist, just type ;C:\Python27\ at the end of it) >> ok >> ok >> ok
3. Open your command prompt (type cmd in Start search for windows user)
4. Next step, Tableau online help said “Change folders to the location where you extracted the Web Data Connector SDK.”
>> In your command prompt type cd c:\WDC SDK (change c:\WDC SDK to the location where you extracted WDC SDK file) >> enter
5. Still in command prompt (which now show c:\WDC SDK>)
for Python 2.x type python -m SimpleHTTPServer 8888
for Python 3.x type python -m http.server 8888
(if you are already using port 8888 for another purpose, specify a different port – i quoted this from tableau onlinehelp)
6. Done. Congratulation!!! now you can move on to the next instruction a.k.a Part 2

The -supposed to be- easiest part is done.
Good luck in creating the code.

Agent Nia is signing off.

Data viz · Tableau

Vizlicious

Data visualization is delicious. Just like 3 pieces of my Doughnut (this doughnut is killing me, it’s 9.24 pm and i want it!!!).

This post is planned to be published a week ago, but due to unexpected laziness of writing down the narration, so…

Well here we go. This post is about my sudden obsession about Data visualization and Tableau. Here is some blog and post from data viz guru and Tableau Zen Master that i found insightful and helpful. Add it to your bookmarks or whatever, it’s up to you.

Data visualization.

Communicating ideas and information. Visually. -Ryan Robitaille- See his post HERE and his blog ryrobes.com

Based on my need, data viz is not about pretty graph or colorful chart, but it’s about presenting information in un-misleading simplicity (is my grammar okay?).

Anyway, there are pros and contras about using one of the most loving chart (since excel still has it so i guess people still loves it, like really loves it) which known as Pie Chart. Who hates Pie right? It’s an amazing dessert, i love chocolate pie. I called this pro and contra as PIE WARRR.

Let the war begin.

Timo Elliot post THIS in his blog, he said that there is a research about how our brain shows no difference in pie chart and bar chart perception (read the comment section too).
And then data viz guru Stephen Few also post an article in his blog in reply to Timo Elliot HERE.
Oh my god this is sooo FUN hihihihihi.
There are so much hate on Pie chart. Below is some example of it:

1. http://datarevelations.com/with-great-power-comes-great-responsibility-or-think-before-you-use-a-donut-chart.html

2. http://darkhorseanalytics.com/blog/salvaging-the-pie/

3. http://100yrsofbrinton.tumblr.com/post/98378563825/the-circular-chart-misused-100yrs-ago

4. https://eagereyes.org/criticism/march-chart-madness

5. http://www.perceptualedge.com/blog/?p=799

But, you might want to read this too >> http://speakingppt.com/2013/03/18/why-tufte-is-flat-out-wrong-about-pie-charts/

So, is it okay to use Pie chart or not? It depends. Depends on your situation, but if you can use other charts to tell your data better, then why bother?

***

Enough about the Pie. I have a list of blog you might want to read.

1. http://darkhorseanalytics.com/blog/
2. http://vizwiz.blogspot.com
3. http://vizcandy.blogspot.com/ (i love Kelly Martin :D)
4. http://juiceanalytics.com/writing/
5. http://datarevelations.com/blog-page
6. http://datavizdoneright.com/
7. http://perceptualedge.com/blog/
8. http://www.wannabeawesomeme.com
9. Mike Bostock awesome visualization >>> http://bost.ocks.org/mike/ If you are brave enough, learn how to make it here >>> http://d3js.org/
10. Must read blog post >> http://vizcandy.blogspot.com/2015/02/the-elusive-executive-dashboard.html

Tableau is Love, Tableau is Life
1. http://wannabedatarockstar.blogspot.com/2014/01/is-tableau-elitist.html
2. http://www.wannabeawesomeme.com/viz-a-viz/identity-crisis
3. list of dataviz blog >>> http://vizwiz.blogspot.com/p/data-viz-blogs.html
4. list of Tableau Zen Master blog >>> http://vizwiz.blogspot.com/p/cool-data-viz-blogs.html

Tableau knowledge :
1. http://evolytics.com/blog/tableau-201-how-to-make-bullet-graphs/
But i make bullet chart using this one (change line chart to gantt bar) >>> http://vizwiz.blogspot.com/2011/08/tableau-tip-7-easy-steps-to-create.html
2. http://theinformationlab.co.uk/2013/08/12/cascading-quick-filters-in-tableau/
3. http://kb.tableau.com/articles/knowledgebase/creating-custom-color-palettes
4. http://kb.tableau.com/articles/knowledgebase/barsintooltips
5. Data blending >> https://tc14.tableau.com/schedule/content/1045
6. http://thevizioneer.blogspot.com/2013/10/tableau-and-auto-save-finally-solution.html (No more losing your hours of work)

There. Done. Finally post this one. Goodnight!!