2.4.1.1 Taksi a New York City

A bincike amfani da babban data daga taxi mita yi nazarin yanke shawara na taksi direbobi a New York. Wadannan bayanai da aka kyau dace domin wannan bincike.

Daya misali daga cikin sauki ikon kirgawa da hakkin abu ya zo daga Henry Farber ta (2015) nazarin hali na New York City taxi direbobi. Ko da yake wannan kungiya zai ba sauti inherently m yana da wani dabarun bincike site for gwada biyu gasar theories a aiki tattalin arziki. Ga dalilai na Farber ta gudanar da bincike, akwai biyu da muhimmanci fasali game da aiki yanayi na taksi direbobi: 1) su hourly ijara fluctuates daga rana-to-day, tushen a part a dalilai kamar weather da kuma 2) yawan hours suna aiki iya canza kowace rana bisa ga direba ta yanke shawara. Wadannan siffofin kai ga mai ban sha'awa tambaya game da dangantaka tsakanin hourly Hakkin kuma hours aiki. Neoclassical model a tattalin arziki hango ko hasashen cewa taxi direbobi za su yi aiki more on kwanaki, inda suka yi mafi girma hourly Hakkin. A madadin, model daga halin tattalin arziki hango ko hasashen daidai akasin. Idan direbobi kafa wani albashi manufa-ce $ 100 a kowace rana-da aikin har sai wannan manufa da aka hadu, to direbobi zai ƙarasa aiki m hours a kwanaki da suke mãsu kamfata more. Alal misali, idan ka kasance a manufa earner, za ka iya kawo karshen sama aiki 4 hours a kan mai kyau rana ($ 25 a kowace awa) da kuma 5 hours a kan wani mummunan rana ($ 20 a kowace awa). Saboda haka, kada direbobi aiki fiye da sa'o'i a kwanaki da hakan hourly Hakkin (as annabta da neoclassical model) ko fiye da sa'o'i a kwana tare da ƙananan hourly Hakkin (as annabta da halin tattalin arziki da model)?

Don amsa wannan tambaya Farber samu bayanai a kan kowane taxi tafiya dauka New York City cabs daga 2009 - 2013, data da cewa yanzu a fili available . Wannan data-wanda aka tattara da lantarki mita cewa birnin bukatar taksin yi amfani da-hada da dama guda na bayani ga kowane tafiya: fara lokaci, fara location, karshen lokaci, karshen location, kudin tafiya, kuma tip (idan tip aka biya da katin bashi). A cikin duka, Farber ta data dauke bayani a kan kimanin 900 miliyan tafiye-tafiye dauka a lokacin kamar miliyan 40 canjawa (a miƙa shi ne wajen wata rana aikin daya direba). A gaskiya ma, akwai sosai data, wanda Farber kawai amfani da bazuwar samfurin da shi domin bincike. Amfani da wannan taksi mita data, Farber gano cewa mafi direbobi aiki more on days lokacin Hakkin ne mafi girma, m da neoclassical ka'idar. Baya ga wannan babban binciken, Farber ya iya yin amfani da girman da data for mafi fahimtar heterogeneity da kuzarin kawo cikas. Farber gano cewa a kan lokaci sababbin direbobi hankali koyi yin aiki fiye da sa'o'i a high ijara days (misali, su koyi yin kamar neoclassical model predicts). Kuma, new direbobi suke yin kamar manufa waɗanda suke tsiwurwurin ne mafi kusantar su sallama kasancewa taksi direba. Duka wadannan more dabara binciken, wanda taimaka bayyana lura hali na yanzu direbobi, su kawai zai yiwu saboda girman da dataset. Suka dã ya kasance ba zai yiwu ba a gane a baya karatu cewa amfani da takarda tafiya zanen gado daga karamin yawan taxi direbobi kan wani kankanen lokaci (misali, Camerer et al. (1997) ).

Farber ta binciken da aka kusa da wani m-harka for wani binciken yin amfani da babban data. Na farko, da data kasance ba wadanda ba wakilin saboda birnin da ake bukata direbobi suyi amfani da digital mita. Kuma, da data kasance ba bai cika saboda bayanai da aka tattara daga cikin birnin da aka m kusa da bayanai da Farber zai tattara idan yana da zabi (daya bambanci shi ne cewa Farber zai yi so da bayanai a kan total Hakkin-Fares da tips- amma birnin data kawai hada tips biya da katin bashi). The kewayawa don Farber ta gudanar da bincike da aka hada da mai kyau tambaya da kyau data. The data kaɗai ne bai isa ba.