Letoto lena la lingoloa le etselitsoe boithuto ba mosebetsi oa kaho toropong e kholo ea Silicon Valley - San Francisco. San Francisco ke "Moscow" ea theknoloji ea lefats'e la rona, e sebelisa mohlala oa eona (ka thuso ea boitsebiso bo bulehileng) ho shebella tsoelo-pele ea indasteri ea kaho metseng e meholo le metse e meholo.
Kaho ea li-graph le lipalo e ile ea etsoa ka (sethaleng sa Kaggle.com).
Lintlha tse mabapi le litumello tsa ho aha tse fetang milione (litlaleho ho li-dataset tse peli) tse tsoang Lefapheng la Kaho la San Francisco - li lumella sekaseka eseng feela mosebetsi oa kaho toropong, empa hape nahana ka hloko mekhoa ea morao-rao le nalane ea nts'etsopele ea indasteri ea kaho lilemong tse 40 tse fetileng, pakeng tsa 1980 le 2019.
Lintlha tse bulehileng li etsa hore ho khonehe ho hlahloba lintlha tse kholo tse susumelitseng le tse tla susumetsa nts'etsopele ea indasteri ea kaho motseng, ho li arola ka "ka ntle" (mathata a moruo le mathata) le "ka hare" (tšusumetso ea matsatsi a phomolo le linako tsa selemo le selemo).
Tse ka hare
Bula data le tlhahlobo ea li-parameter tsa mantlha.
Sena ha se phetolelo ea sengoloa. Ke ngola ho LinkedIn mme e le hore ke se ke ka etsa litšoantšo ka lipuo tse 'maloa, litšoantšo tsohle li ka Senyesemane.
Kopana le karolo ea bobeli:
Lintlha tsa Tumello ea Kaho ea Toropo ea San Francisco - Ho tsoa ho Open Data Portal - . Sebaka sa marang-rang se na le li-dataset tse 'maloa ka sehlooho sa kaho. Li-dataset tse peli tse joalo li boloka le ho ntlafatsa datha mabapi le litumello tse fanoeng bakeng sa kaho kapa tokiso ea lintho teropong:
- (litlaleho tse likete tse 850)
- (Lirekoto tse likete tse 280, data e jarollotsoe le ho ntlafatsoa beke le beke)
Li-dataset tsena li na le tlhahisoleseling mabapi le litumello tsa kaho tse fanoeng, tse nang le litšobotsi tse fapaneng tsa ntho eo tumello e fanoeng bakeng sa eona. Kakaretso ea lipehelo (litumello) tse amohetsoeng nakong ea 1980-2019 - litumello tse 1.
Lintlha tse ka sehloohong tsa dataset ena tse sebelisitsoeng ho hlahloba:
- tumello_letsatsi_la_ho bopa - letsatsi la tlhahiso ea kopo (ha e le hantle, letsatsi leo mosebetsi oa kaho o qalang ka lona)
- tlhaloso - tlhaloso ea kopo (mantsoe a mabeli kapa a mararo a hlalosang morero oa kaho (mosebetsi) oo tumello e entsoeng bakeng sa oona)
- khakanyo_chelete - chelete e hakanyetsoang (e hakantsoeng) ea mosebetsi oa kaho
- revised_cost - litšenyehelo tse ntlafalitsoeng (litšenyehelo tsa mosebetsi ka mor'a ho hlahlobisisoa, ho eketseha kapa ho fokotseha ha meqolo ea pele ea kopo)
- e leng_sebedisa - mofuta oa matlo (ntlo e le 'ngoe, ea malapa a mabeli, lifolete, liofisi, tlhahiso, joalo-joalo)
- zipcode, sebaka — khoutu ea poso le likhokahano tsa lintho
Mosebetsi oa kaho oa selemo le selemo San Francisco
Kerafo e ka tlase e bontša li-parameter khakanyo_chelete и revised_cost e hlahisoang e le kabo ea kakaretso ea litšenyehelo tsa mosebetsi ka khoeli.
data_cost_m = data_cost.groupby(pd.Grouper(freq='M')).sum()Ho fokotsa "li-outliers" tsa khoeli le khoeli, lintlha tsa khoeli le khoeli li hlophisoa ka selemo. Kerafo ea chelete e tsetelitsoeng ka selemo e fumane mokhoa o utloahalang le o ka hlahlojoang haholoanyane.
data_cost_y = data_cost.groupby(pd.Grouper(freq='Y')).sum()
Ho ipapisitsoe le motsamao oa selemo le selemo oa kakaretso ea litšenyehelo (litumello tsohle tsa selemo) ho ea litsing tsa toropo lintlha tsa moruo tse susumelitseng ho tloha 1980 ho fihlela 2019 li bonahala ka ho hlaka ka palo le litšenyehelo tsa merero ea kaho, kapa ho seng joalo matsete a San Francisco Real Estate.
Palo ea tumello ea ho haha (palo ea mesebetsi ea kaho kapa palo ea matsete) lilemong tse 40 tse fetileng e amana haufi-ufi le mesebetsi ea moruo Silicon Valley.
Tlhōrō ea pele ea mosebetsi oa kaho e ne e amahanngoa le hype ea elektronike bohareng ba lilemo tsa bo-80 phuleng. Ho putlama ha moruo oa elektroniki le libanka ka 1985 ho ile ha etsa hore 'maraka oa libaka o theohe moo o sa kang oa hlaphoheloa ka lilemo tse ka bang leshome.
Ka mor'a moo, ka makhetlo a mabeli (ka 1993-2000 le 2009-2016) pele ho oa ha bubble ea Dotcom le theknoloji e ntseng e eketseha ea lilemong tsa morao tjena. Indasteri ea kaho ea San Francisco e bile le kholo e makatsang ea liperesente tse likete tse 'maloa..
Ka ho tlosa litlhōrō tse mahareng le lithupa le ho siea boleng bo tlase le bo phahameng ka ho fetisisa bakeng sa potoloho e 'ngoe le e' ngoe ea moruo, ho hlakile hore na liphetoho tse kholo tsa mebaraka li aparetse indasteri joang lilemong tse 40 tse fetileng.
Keketseho e kholo ka ho fetisisa ea matsete a kaho e etsahetse nakong ea dot-com boom, ha pakeng tsa 1993 le 2001 $10 bilione e ne e tseteloa ho nchafatsa le kaho, kapa hoo e ka bang $1 bilione ka selemo. Haeba re bala ka lisekoere metres (litšenyehelo tsa 1 m² ka 1995 ke $3000), sena se ka ba 350 m000 ka selemo bakeng sa lilemo tse 2, ho qala ka 10.
Kholo ea matsete a selemo le selemo nakong ena e fihlile ho 1215%.
Lik'hamphani tse hirileng thepa ea kaho nakong ena li ne li tšoana le lik'hamphani tse neng li rekisa likharafu nakong ea ho potlaka ha khauta (sebakeng sona seo bohareng ba lekholo la bo19 la lilemo). Feela ho e-na le likharafu, lilemong tsa bo-2000 ho ne ho se ho ntse ho e-na le li-cranes le lipompo tsa konkreite bakeng sa lik'hamphani tse sa tsoa thehoa tsa kaho tse neng li batla ho etsa chelete ka boomo ea kaho.
Kamora mathata a mangata ao indasteri ea kaho e bileng le 'ona ho theosa le lilemo, lilemong tse peli tse tlang tsa kamora koluoa, matsete (palo ya dikopo tsa diphemiti) bakeng sa kaho e theohile ka bonyane 50% nako le nako.
Likoluoa tse kholo indastering ea kaho ea San Francisco li etsahetse lilemong tsa bo-90. Moo, ka periodicity ea lilemo tse 5, indasteri e ka 'na ea oa (-85% nakong ea 1983-1986), eaba e tsoha hape (+895% nakong ea 1988-1992), e setseng ka selemo sa 1981, 1986, 1988. , 1993 - boemong bo tšoanang.
Kamora 1993, litheolelo tsohle tse ileng tsa latela indastering ea kaho ha lia ka tsa feta 50%. Empa mathata a moruo a atamelang (ka lebaka la COVID-19) e ka baka mathata a rekotiloeng indastering ea kaho nakong ea 2017-2021, ho fokotseha ha eona ho seng ho ntse ho le teng nakong ea 2017-2019 e leng kakaretso ea ho feta 60%.
Keketseho ea baahi ba San Francisco dynamics nakong ea 1980-1993 hape e bontshitse kgolo e batlang e feta. Matla a moruo le matla a macha a Silicon Valley e ne e le motheo o tiileng oo phetiso ea Moruo o Mocha, Renaissance ea Amerika, le dot-coms e hahiloeng holim'a eona. E ne e le sehlohlolong sa moruo o mocha. Empa ho fapana le ho phahama ha matsete a thekiso ea matlo le meaho, kamora tlhoro ea dot-com, palo ea batho e ile ea hla ea tlala.
Pele ho tlhoro ea dot-com ka 2001, kholo ea selemo le selemo ea baahi ho tloha ka 1950 e bile hoo e ka bang 1% ka selemo. Joale, ka mor'a ho putlama ha bubble, tšubuhlellano ea baahi ba bacha e ile ea fokotseha 'me ho tloha ka 2001 e bile karolo ea 0.2 lekholong feela ka selemo.
Ka selemo sa 2019 (ka lekhetlo la pele ho tloha 1950), matla a kholo a bonts'itse phallo ea baahi (-0.21% kapa batho ba 7000) ho tsoa toropong ea San Francisco.
Tebello le 'nete ha u lokisetsa likhakanyo tsa litšenyehelo
Ho li-datasets tse sebelisitsoeng, lintlha tse mabapi le litšenyehelo tsa tumello ea morero oa kaho li arotsoe ka:
- theko e lekantsoeng ea mantlha (khakanyo_chelete)
- litšenyehelo tsa mosebetsi ka mor'a ho hlahlojoa (revised_cost)
Nakong ea linako tsa boom, morero o ka sehloohong oa ho tsosolosa ke ho eketsa litšenyehelo tsa pele, ha motseteli (moreki oa kaho) a bontša takatso ea lijo ka mor'a ho qala ho haha.
Nakong ea tlokotsi, ba leka ho se fete litšenyehelo tse hakantsoeng, 'me likhakanyo tsa pele ha li na liphetoho (ntle le tšisinyeho ea lefatše ea 1989).
Ho ea ka kerafo e hahiloeng holim'a phapang lipakeng tsa litšenyehelo tse tsosolositsoeng le tse hakantsoeng (revised_cost - measure_cost), ho ka hlokomeloa hore:
Palo ea ho eketseha ha litšenyehelo ha ho ntšuoa boleng ba mosebetsi oa kaho ka kotloloho ho ipapisitse le ho phahama ha moruo
data_spread = data_cost.assign(spread = (data_cost.revised_cost-data_cost.estimated_cost))
Nakong ea kholo ea moruo e potlakileng, bareki ba mosebetsi (batseteli) ba sebelisa chelete ea bona ka seatla se bulehileng, ba eketsa likopo tsa bona kamora ho qala mosebetsi.
Moreki (motseteletsi), o ikutloa a itšepa licheleteng, o kopa rakonteraka oa kaho kapa setsebi sa meralo ea meralo ho atolosa tumello ea ho haha e seng e fanoe. Sena e ka 'na ea e-ba qeto ea ho eketsa bolelele ba pele ba letamo kapa ho eketsa sebaka sa ntlo (kamora ho qala mosebetsi le ho fana ka tumello ea ho haha).
Sehlohlolong sa mehla ea dot-com, litšenyehelo tse joalo tsa “tlatsetso” li ile tsa fihla ho “eketsehileng” 1 bilione ka selemo.
Haeba u sheba tafole ena e se e ntse e le phetohong ea liperesente, joale keketseho ea tlhōrō ea khakanyo (100% kapa makhetlo a 2 ho feta tefiso e lekantsoeng ea pele) e etsahetse ka selemo pele ho tšisinyeho ea lefatše e ileng ea etsahala ka 1989 haufi le motse. Ke nahana hore ka mor’a tšisinyeho ea lefatše, merero ea kaho e qalileng ka 1988, ka mor’a tšisinyeho ea lefatše ka 1989, e ne e hloka nako le chelete e eketsehileng bakeng sa ho e phethahatsa.
Ka lehlakoreng le leng, tokiso e tlase ea litšenyehelo tse hakantsoeng (tse etsahetseng hanngoe feela nakong ea 1980 ho isa 2019) lilemo tse 'maloa pele ho tšisinyeho ea lefatše e bakoa ke taba ea hore merero e meng e qalileng ka 1986-1987 e ile ea emisoa kapa matsete a merero ena a fokotsehile. fatshe. Ka kemiso ka karolelano bakeng sa morero o mong le o mong o qalileng ka 1987 - phokotso ea litšenyehelo tse hakantsoeng e ne e le -20% ea morero oa pele..
data_spred_percent = data_cost_y.assign(spred = ((data_cost_y.revised_cost-data_cost_y.estimated_cost)/data_cost_y.estimated_cost*100))
Keketseho ea theko e hakanyetsoang ea pele ea ho feta 40% e bontšitsoeng kapa mohlomong e bakiloe ke ho ata ha limmaraka tsa lichelete le tsa kaho.
Ke lebaka lefe la ho fokotseha ha ho ata (phapang) lipakeng tsa litšenyehelo tse hakantsoeng le tse ntlafalitsoeng kamora 2007?
Mohlomong bo-ramatsete ba ile ba qala ho sheba lipalo ka hloko (karolelano ea chelete e fetang lilemo tse 20 e eketsehile ho tloha ho $ 100 ho isa ho $ 2 milione) kapa mohlomong lefapha la kaho, ho thibela le ho thibela li-bubble tse hlahang 'marakeng oa thekiso ea matlo, li hlahisitse melao le lithibelo tse ncha ho fokotsa ho qhekella ho ka khonehang. le likotsi tse ka ’nang tsa hlaha nakong ea lilemo tsa tlokotsi.
Mosebetsi oa kaho ho latela nako ea selemo
Ka ho hlophisa lintlha ka libeke tsa almanaka tsa selemo (libeke tse 54), u ka bona ts'ebetso ea kaho toropong ea San Francisco ho latela nako ea selemo le nako ea selemo.
Ka Keresemese, mekhatlo eohle ea kaho e leka ho fumana tumello ea merero e mecha "e kholo" ka nako. (ka nako e tšoanang! palo! ea litumello likhoeling tsona tsena e boemong bo tšoanang ho pholletsa le selemo). Bahoebi, ba rerileng ho fumana thepa ea bona nakong ea selemo se tlang, ba kena likonteraka likhoeling tsa mariha, ba itšetlehile ka litheolelo tse kholo (kaha likonteraka tsa lehlabula, boholo ba tsona, li fela qetellong ea selemo le lik'hamphani tsa kaho li thahasella. ho amohela likopo tse ncha).
Pele ho Keresemese, palo e kholo ka ho fetisisa ea likopo e romelloa (eketseha ho tloha ho karolelano ea limilione tse likete tse 1-1,5 ka khoeli ho ea ho limilione tse likete tse 5 ka December feela). Ka nako e ts'oanang, palo eohle ea likopo ka khoeli e lula e le boemong bo tšoanang (sheba karolo e ka tlase: lipalo-palo tsa palo eohle ea likopo ka khoeli le letsatsi)
Ka mor'a matsatsi a phomolo a mariha, indasteri ea kaho e rera ka mafolofolo le ho kenya ts'ebetsong litaelo tsa "Keresemese" (ho batla ho se na keketseho ea palo ea tumello) e le ho lokolla lisebelisoa bohareng ba selemo (pele ho matsatsi a phomolo a Letsatsi la Boipuso) pele ho letsatsi le lecha. leqhubu la likonteraka tsa lehlabula le qala hang ka mor'a matsatsi a phomolo a June.
data_month_year = data_month_year.assign(week_year = data_month_year.permit_creation_date.dt.week)
data_month_year = data_month_year.groupby(['week_year'])['estimated_cost'].sum()
E tšoanang peresente ya data (orange line) e boetse e bontša hore indasteri e sebetsa "boreleli" ho pholletsa le selemo, empa pele le ka mor'a matsatsi a phomolo, mosebetsi oa litumello o eketseha ho 150% ka nako e pakeng tsa beke 20-24 (pele ho Letsatsi la Boipuso), le e fokotseha hang ka mor'a matsatsi a phomolo ho fihlela ho -70%.
Pele ho Halloween le Keresemese, mesebetsi ea indasteri ea kaho ea San Francisco ka libeke tse 43-44 e eketseha ka 150% (ho tloha tlase ho ea holimo) ebe e fokotseha ho ea ho zero nakong ea matsatsi a phomolo.
Kahoo, indasteri e nakong ea potoloho ea likhoeli tse tšeletseng, e arohanngoa ke matsatsi a phomolo "Letsatsi la Boipuso la US" (beke ea 20) le "Keresemese" (beke ea 52).
Kakaretso ea matsete a matlo a San Francisco
Ho ipapisitsoe le lintlha tsa tumello ea ho aha teropong:
Kakaretso ea matsete mererong ea kaho San Francisco ho tloha 1980 ho isa 2019 ke $91,5 bilione.
sf_worth = data_location_lang_long.cost.sum()
Kakaretso ea boleng ba 'maraka ba meaho eohle ea bolulo ho la San Francisco e lekantsoeng ka lekhetho la thepa (e leng boleng bo lekantsoeng ba litša le thepa eohle ea San Francisco) .
Ke libaka life tsa San Francisco tse tsetelelitseng ho tsona lilemong tse 40 tse fetileng?
Re sebelisa laebrari ea Folium, ha re boneng moo $91,5 bilione ena e ileng ea tseteloa ke sebaka. Ho etsa sena, ha re se re hlophisitse data ka khoutu ea zip, re tla emela litekanyetso tse hlahisoang re sebelisa li-circles (ts'ebetso ea Circle ho laebraring ea Folium).
import folium
from folium import Circle
from folium import Marker
from folium.features import DivIcon
# map folium display
lat = data_location_lang_long.lat.mean()
long = data_location_lang_long.long.mean()
map1 = folium.Map(location = [lat, long], zoom_start = 12)
for i in range(0,len(data_location_lang_long)):
Circle(
location = [data_location_lang_long.iloc[i]['lat'], data_location_lang_long.iloc[i]['long']],
radius= [data_location_lang_long.iloc[i]['cost']/20000000],
fill = True, fill_color='#cc0000',color='#cc0000').add_to(map1)
Marker(
[data_location_mean.iloc[i]['lat'], data_location_mean.iloc[i]['long']],
icon=DivIcon(
icon_size=(6000,3336),
icon_anchor=(0,0),
html='<div style="font-size: 14pt; text-shadow: 0 0 10px #fff, 0 0 10px #fff;; color: #000";"">%s</div>'
%("$ "+ str((data_location_lang_long.iloc[i]['cost']/1000000000).round()) + ' mlrd.'))).add_to(map1)
map1
Ho hlakile ho tsoa literekeng hore Bongata ba pie ka ho utloahalang bo ile ba ea DownTown. Kaha re nolofalitse lihlopha tsa lintho tsohle ho ea bohareng ba toropo le nako ea ho fihla bohareng ba toropo (ehlile, matlo a theko e boima a ntse a hahoa lebōpong la leoatle), litumello tsohle li arotsoe ka lihlopha tse 4: 'Downtown' , '<0.5H Downtown', '< 1H Downtown', 'Kantle ho SF'.
from geopy.distance import vincenty
def distance_calc (row):
start = (row['lat'], row['long'])
stop = (37.7945742, -122.3999445)
return vincenty(start, stop).meters/1000
df_pr['distance'] = df_pr.apply (lambda row: distance_calc (row),axis=1)
def downtown_proximity(dist):
'''
< 2 -> Near Downtown, >= 2, <4 -> <0.5H Downtown
>= 4, <6 -> <1H Downtown, >= 8 -> Outside SF
'''
if dist < 2:
return 'Downtown'
elif dist < 4:
return '<0.5H Downtown'
elif dist < 6:
return '<1H Downtown'
elif dist >= 6:
return 'Outside SF'
df_pr['downtown_proximity'] = df_pr.distance.apply(downtown_proximity)Ho tse limilione tse likete tse 91,5 tse tsetelitsoeng toropong, hoo e ka bang limilione tse likete tse 70 (75% ea matsete ohle) a tsetelitsoeng ho lokisoa le kahong e bohareng ba toropo. (sebaka se setala) le ho ea sebakeng sa toropo ka har'a sebaka sa 2 km. ho tloha bohareng (sebaka se seputsoa).
Karolelano e hakanyetsoang ea litšenyehelo tsa kopo ea kaho ho latela setereke sa toropo
Lintlha tsohle, joalo ka kakaretso ea chelete ea matsete, li arotsoe ka zip code. Tabeng ena feela ka kakaretso (.mean()) e hakantsoeng ea litšenyehelo tsa ts'ebeliso ka khoutu ea zip.
data_location_mean = data_location.groupby(['zipcode'])['lat','long','estimated_cost'].mean()
Libakeng tse tloaelehileng tsa toropo (ho feta 2 km ho tloha bohareng ba toropo) - ka karolelano litšenyehelo tsa kopo ea kaho ke $ 50 tse likete.
Karolelano e hakantsoeng ea litšenyehelo sebakeng sa bohareng ba toropo e batla e phahame ka makhetlo a mararo ($ 150 sekete ho $ 400 sekete) ho feta libakeng tse ling ($ 30-50 sekete).
Mo godimo ga ditshenyegelo tsa setsha, go na le mabaka a le mararo a a laolang palo-kakaretso ya ditshenyegelo tsa go aga matlo: badiri, didirisiwa, le dituelo tsa puso. Likarolo tsena tse tharo li holimo California ho feta naheng eohle. Likhoutu tsa mohaho oa California li nkoa e le tse ling tsa mekhoa e mengata ka ho fetisisa le e thata ka ho fetisisa naheng (ka lebaka la tšisinyeho ea lefatše le melao ea tikoloho), hangata e hlokang thepa e theko e boima le basebetsi.
Ka mohlala, ’muso o hloka hore lihahi li sebelise thepa ea mohaho ea boleng bo holimo (lifensetere, litšila, tsamaiso ea ho futhumatsa le ho pholisa) ho finyella litekanyetso tse phahameng tsa tšebeliso ea matla.
Ho latela lipalo-palo tse akaretsang tsa litšenyehelo tse tloaelehileng tsa kopo ea tumello, libaka tse peli lia ikhetha:
- Sehlekehleke sa Letlotlo - sehlekehleke sa maiketsetso se San Francisco Bay. Ka karolelano litšenyehelo tsa tumello ea ho haha ke $ 6,5 milione.
- Setsi sa Bay — (baahi ba 2926) Ka karolelano litšenyehelo tse hakantsoeng tsa tumello ea ho haha ke liranta tse limilione tse 1,5.
Ha e le hantle, karolelano e phahameng ea tšebeliso libakeng tsena tse peli e amana ka palo e fokolang ea likopo tsa libaka tsena tsa poso (145 le 3064 ka ho latellana, kaho sehlekehlekeng sena e fokotsehile haholo), athe bakeng sa lintlha tse ling tsa poso - XNUMX.mme nako ya 1980-2019 e amohetse dikopo tse ka bang 1300 ka selemo (kakaretso ka kakaretso 30 -50 likete likopo bakeng sa nako eohle).
Ho ea ka "palo ea likopo" parameter, kabo e lekanang hantle ea palo ea likopo ka khoutu ea poso ho pholletsa le toropo ea bonahala.
Lipalopalo tsa palo eohle ea likopo ka khoeli le letsatsi
Lipalopalo tse akaretsang tsa palo eohle ea likopo ka khoeli le letsatsi la beke lipakeng tsa 1980 le 2019 li bonts'a hore Likhoeli tse khutsitseng ka ho fetisisa lefapheng la kaho ke likhoeli tsa selemo le tsa mariha. Ka nako e ts'oanang, chelete ea matsete e boletsoeng lits'ebetsong e fapana haholo 'me e fapana khoeli le khoeli ka linako tse ling. (sheba hape "Mosebetsi oa kaho ho latela nako ea selemo"). Har'a matsatsi a beke, ka Mantaha mojaro oa lefapha o batla o le ka tlase ho 20% ho feta matsatsing a mang a beke.
months = [ 'January', 'February', 'March', 'April', 'May','June', 'July', 'August', 'September', 'October', 'November', 'December' ]
data_month_count = data_month.groupby(['permit_creation_date']).count().reindex(months)
Le hoja Phuptjane le Phupu li batla li tšoana ho latela palo ea likopo, ho latela kakaretso ea litšenyehelo tse hakantsoeng phapang e fihla ho 100% (libilione tse 4,3 ka Mots'eanong le Phupu le 8,2 bilione ka Phuptjane).
data_month_sum = data_month.groupby(['permit_creation_date']).sum().reindex(months)
Bokamoso ba indasteri ea kaho ea San Francisco, ho bolela esale pele tšebetso ka lipaterone.
Qetellong, a re bapiseng chate ea mosebetsi oa kaho San Francisco le chate ea theko ea Bitcoin (2015-2018) le chate ea theko ea khauta (1940 - 1980).
Mohlala (ho tloha ho mokhoa oa Senyesemane - mohlala, mohlala) - tlhahlobo ea tekheniki e tsitsitseng e pheta-phetoang motsoako oa theko, molumo kapa lintlha tsa matšoao li bitsoa. Tlhahlobo ea mohlala e thehiloe ho e 'ngoe ea li-axioms tsa tlhahlobo ea tekheniki: "histori e ipheta" - ho lumeloa hore ho kopana khafetsa ha data ho lebisa sephethong se ts'oanang.
Mokhoa o ka sehloohong o ka bonoang chate ea tšebetso ea selemo le selemo ke Ena ke "Hlooho le Mahetla" mokhoa oa ho fetola mokhoa. E bitsoa joalo hobane chate e shebahala joaloka hlooho ea motho (tlhōrō) le mahetla mahlakoreng (litlhōrō tse nyenyane). Ha theko e tlōla moeli o kopanyang likoti, mohlala o nkoa o feletse 'me ho ka etsahala hore motsamao o theohele tlaase.
Metsamao ea ts'ebetso ea indasteri ea kaho ea San Francisco e batla e lumellana ka ho feletseng le ho phahama ha theko ea khauta le bitcoin. Ts'ebetso ea nalane ea lichate tsena tse tharo tsa litheko le mesebetsi e bonts'a ho ts'oana ho hlokomelehang.
Ho tseba ho bolela esale pele boitšoaro ba 'maraka oa kaho nakong e tlang, hoa hlokahala ho bala coefficient ea ho hokahanya ka e 'ngoe le e' ngoe ea mekhoa ena e 'meli.
Liphetoho tse peli tse sa reroang li bitsoa correlated haeba nako ea tsona ea correlation (kapa coefficient coefficient) e fapane le zero; 'me li bitsoa lipalo tse sa amaneng haeba nako ea tsona ea khokahano e le zero.
Haeba boleng bo hlahisoang bo le haufi le 0 ho feta ho 1, joale ha ho na thuso ho bua ka mokhoa o hlakileng. Bona ke bothata bo rarahaneng ba lipalo, bo ka nkuoang ke bo-mphato ba baholo ba ka bang ba thahasella sehlooho sena.
Haeba! ha e lumellane le mahlale! sheba sehlooho sa tsoelo-pele e eketsehileng ea indasteri ea kaho San Francisco: haeba mohlala o tsoela pele ho lumellana le theko ea Bitcoin, joale ho latela khetho ena e hlokang tšepo - ho tsoa maqakabetsing a indasteri ea kaho San Francisco ho ke ke ha e-ba bonolo nakong ea hang-hang ka mor'a tlokotsi.
Ka khetho ea "tšepo" ho feta ntshetsopele, khōlo e phetoang khafetsa indastering ea kaho e ka khoneha haeba mosebetsi mona o latela boemo ba "theko ea khauta". Tabeng ena, lilemong tsa 20-30 (mohlomong ho 10), lekala la kaho le tla tobana le ts'ebetso e ncha ea mosebetsi le nts'etsopele.
Ke tla shebisisa likarolo tse ling tsa kaho (ho lokisa marulelo, likhitla, ho hahoa ha litepisi, likamore tsa ho hlapela, haeba u na le litlhahiso mabapi le liindasteri kapa lintlha tse ling - ka kopo ngola litlhaloso) 'me u bapise theko ea lichelete bakeng sa mefuta ea mosebetsi ka bomong. litefiso tse tsitsitseng tsa likalimo tsa mortgage le phaello ea libonto tsa mmuso tsa US (Fixed Mortgage Rates & US Treasury Yield).
Kopana le karolo ea bobeli:
Sehokelo ho Jupyter Notebook:
Ka kopo, bakeng sa ba nang le Kaggle, fa Notebook tlatsetso (Kea leboha!).
(Litlhaloso le litlhaloso tsa khoutu li tla kenyelletsoa hamorao ho Notebook)
Khokahano ea mofuta oa Senyesemane:
Haeba u natefeloa ke litaba tsa ka, ka kopo nahana ka ho nthekela kofi.
ke leboha tšehetso ea hau!
Ke basebelisi ba ngolisitsoeng feela ba ka kenyang letsoho phuputsong. ka kopo.
Bokamoso ba indasteri ea kaho ea San Francisco ke bofe?
-
66,7%Lefapha la kaho le ka 'na la latela tsela ea Bitcoin2
-
0,0%Lekala la kaho le ka latela tsela ea litheko tsa khauta0
-
0,0%Lekala le lebelletse hype nakong ea lilemo tse 10 tse tlang
-
33,3%Ntlafatso ea lekala ha e tsamaee ho latela mekhoa1
Basebelisi ba 3 ba ile ba khetha. Basebelisi ba 6 ba ile ba hana.
Source: www.habr.com