Hadley Ecosystem 2016

Tokyo.R#53 での発表資料

Transcript

1. ϋυϦʔ͞Μɺ
   ਐ௙Ͳ͏Ͱ͔͢ ӝੜਅ໵
   
   
   
   
   
   
   
   !V@SJCP ! " 
   5PLZP3!4IJCVZB )BEMFZ
   &DPTZTUFN
   
2. ຊ೔ͷओ໾
3. Hadley Wickham はどりー うぃっかむ
4. # ୭ʁ
5. Hadley Wickham 34UVEJP
   νʔϑαΠΤϯςΟετ $ ΦʔΫϥϯυେֶʢχϡʔδʔϥϯυʣͷ
   ඇৗۈߨࢣ $ 3քͷϨδΣϯυ $ Ӌௗʢ೔ຊͰͷѪশʣ $ !
6. ʮ3Ͱ̏ԯԁʯ ʹ࠷΋͍ۙஉ
7. )BEMFZ
   &DPTZTUFN
   d Hadley Ecosystem: Reshape, Plyr, GGplot @Etienne https://speakerdeck.com/lowdecarie/hadley-ecosystem-reshape-plyr-ggplot
8. Hadley Ecosystem
9. σʔλαΠΤϯε σʔλऔಘ ΤϯδχΞϦϯά $ $ $
10. Fry to the Hadley World
11. 
    άϥϑͷछྨʹԠͨ͡࡞ਤؔ਺
    plot(x, y, ...) barplot(height, ...) boxplot(x, ...) hist(x, ...) pie(x, labels = names(x), ...) # ؔ਺ͷҾ਺͕
    ଗ͓ͬͯΒͣɺ
    ֮͑ʹ͍͘ \HHQMPU^ύοέʔδ
12. 
    άϥϑඳըͷͨΊͷจ๏ʹجͮ͘ϨΠϠʔͷߟ͑Λಋೖ
    library(ggplot2) ggplot(data, aes(x, y)) + geom_*() + xlab() + ylab() \HHQMPU^ύοέʔδ
13. 
    άϥϑඳըͷͨΊͷจ๏ʹجͮ͘ϨΠϠʔͷߟ͑Λಋೖ
    library(ggplot2) ggplot(data, aes(x, y)) + geom_*() + xlab() + ylab() άϥϑʹ༻͍ΔཁૉΛࢦఆ
    BFT
    ΤεςςΟοΫ৹ඒతཁૉΛఆٛ͢Δؔ਺ Y Z [   ʜ   ʜ   ʜ \HHQMPU^ύοέʔδ
14. 
    άϥϑඳըͷͨΊͷจ๏ʹجͮ͘ϨΠϠʔͷߟ͑Λಋೖ
    library(ggplot2) ggplot(data, aes(x, y)) + geom_*() + xlab() + ylab() 
    ඳը͢ΔάϥϑͷछྨΛࢦఆ FY
    HFPN@QPJOU ʜ
    ࢄ෍ਤɺ
    
    
    
    
    
    HFPN@CBS ʜ
    ๮άϥϑ \HHQMPU^ύοέʔδ
15. 
    HFPN@Ͱ࢝·Δؔ਺Λදࣔ͢Δ
    ls("package:ggplot2") %>% grep("^geom_", ., value = TRUE) \HHQMPU^ύοέʔδ
16. 
    άϥϑඳըͷͨΊͷจ๏ʹجͮ͘ϨΠϠʔͷߟ͑Λಋೖ
    library(ggplot2) ggplot(data, aes(x, y)) + geom_*() + xlab() + ylab() 
    ඞཁͳཁૉΛϨΠϠʔͱͯ͠௥Ճ \HHQMPU^ύοέʔδ # ౷Ұ͞Εͨه๏ͱؔ਺ʹΑΓɺ
    άϥϑ࡞੒͕؆ܿʹ
17. 
    3ඪ४ͷจࣈྻૢ࡞ؔ਺ͱ\TUSJOHJ^ύοέʔδ
    paste("ほく", "のうむ") # [1] "ほく のうむ" stringi::stri_c("ほく", "のうむ") # [1] "ほくのうむ" \TUSJOHS^ύοέʔδ # \TUSJOHS^͸*$6Λαϙʔτ͢Δ\TUSJOHJ^ͷ ϥούʔύοέʔδ
18. 
    TUS@ ͱ͍͏ؔ਺໊
    library(stringr) ls("package:stringr") %>% grep("^str_", ., value = TRUE) stringr::str_c("ほく", "のうむ") # [1] "ほくのうむ" \TUSJOHS^ύοέʔδ
19. library(stringi) stri_locale_list() %>% length() # [1] 683 stri_datetime_symbols() %>% names() # [1] "Month" "Weekday" "Quarter" "AmPm" "Era" stri_datetime_symbols(locale="ja_JP_TRADITIONAL")$Era %>% tail() # [1] "元治" "慶応" "明治" "大正" "昭和" "平成" \TUSJOHS^ύοέʔδ # ϩέʔϧɺϢχίʔυਖ਼نԽͳͲ͸
    \TUSJOHJ^ύοέʔδͷؔ਺ͰରԠʂ
20. stri_trans_general("㈱ﾎｸｿｴﾑ", id = "nfkd") stri_trans_general("ת京", "Any-ch_FONIPA") stri_trans_general("Nippon", "es-ja") stri_trans_general("ת京", "Simplified-Traditional") \TUSJOHS^ύοέʔδ :PVS
    5VSO
21. stri_trans_general("㈱ﾎｸｿｴﾑ", id = "nfkd") # [1] "(株)ホクソエム" stri_trans_general("ת京", "Any-ch_FONIPA") # [1] "dōng jīng" stri_trans_general("Nippon", "es-ja") # [1] "ニッポン" stri_trans_general("ת京", "Simplified-Traditional") # [1] "東京" \TUSJOHS^ύοέʔδ
22. 
    3Ͱ೔෇ɾ࣌ؒΦϒδΣΫτΛѻ͏ؔ਺
    c("2016-04-30") %>% as.Date() # [1] "2016-04-30" Sys.Date() # [1] "2016-04-30" \MVCSJEBUF^ύοέʔδ # ϢʔβʔϑϨϯυϦʔͰͳ͍ؔ਺໊ͱ
    ࣌ܥྻσʔλΛѻ͏ύοέʔδͷ܈༤ׂڌʜ
23. library(lubridate) today() # [1] "2016-04-30" ymd("2016年4月30日", tz = "Asia/Tokyo") %>% month() # [1] 4 \MVCSJEBUF^ύοέʔδ λΠϜκʔϯ΍ϩέʔϧͷ
    มߋ͕ؔ਺಺Ͱ࣮ߦՄೳ
24. 
    ೔෇ɾ࣌ؒͷՃ޻ͱߏจղੳ
    today() + weeks(1) # [1] "2016-05-06" x <- dmy("30/04/2016") day(x) <- 26 x # [1] "2016-04-26" \MVCSJEBUF^ύοέʔδ # ௚ײతʹཧղ
    ͠΍͍ؔ͢਺໊ɻ
    ೔෇ɾ࣌ؒσʔλͷ
    ѻ͍ʹศརͳ
    ิॿؔ਺Λఏڙ
25. 
    ศརͳิॿؔ਺܈
    (time1 <- ymd_hms("2016-04-01 09:00:00", tz = "Asia/Tokyo")) # [1] "2016-04-01 09:00:00 JST" with_tz(time1, "America/New_York") # [1] "2016-03-31 20:00:00 EDT" \MVCSJEBUF^ύοέʔδ ࢦఆՄೳͳλΠϜκʔϯ͸0MTPO/BNFT
    Ͱ֬ೝ
26. (chl <- interval(ymd("1989年11月27日", tz = "Asia/Tokyo"), mdy("April 30, 2016", tz = "Asia/Tokyo"))) # [1] 1989-11-27 JST--2016-04-30 JST time1 %within% chl # [1] TRUE \MVCSJEBUF^ύοέʔδ
27. \SFBES^ύοέʔδ 
    طଘͷදܗࣜϑΝΠϧಡΈࠐΈؔ਺ͷվྑ
    library(readr) read_csv(file, col_types = NULL, locale = default_locale()) # SFBEDTW DPM$MBTTFT Λجૅͱͨ͠ DPMMFDUPS ؔ਺ͷ੔උͱϩέʔϧରԠ
28. \SFBES^ύοέʔδ ܕ ؔ਺ লུ࣌ͷදه ࿦ཧܕ col_logical() l (logical) ੔਺ܕ col_integer() i (integer) ࣮਺ܕ col_double() d (double) จࣈྻܕ col_character c (character) ೔෇ܕʢ:NEදهʣ col_date(format) D (date) ೔࣌ܕʢ<*40>ʣ col_datetime(format, tz) T (time) ࣌ؒܕ col_time(format) ਺஋ܕ col_number() n ཁҼܕ col_factor(levels, ordered) ಡΈࠐΈର৅ͱ͠ͳ͍ col_skip() _, - collector()
29. read_csv("x,y,z\n1,2,a\n3,4,b", col_types = cols(x = "i", y = "c", z = "_")) \SFBES^ύοέʔδ Source: local data frame [2 x 2] x y <int> <chr> 1 1 2 2 3 4 ม਺ͷܕ͸ೖྗͷઌ಄ߦ͔Β
    ࣗಈతʹਪଌ͞ΕΔ #
30. library(httr) GET(…) POST(…) library(rvest) xml2::read_html(x = …) %>% html_nodes(xpath = …) ΢Σϒσʔλऔಘ # IUUQϝιουʹରԠͨؔ͠਺܈ɻ
    ೝূɺIUUQϔομʔ౳ͷαϙʔτ IUNM
    YNM
    ͲΜͱདྷ͍
31. 3ύοέʔδ։ൃิॿ library(devtools) install_*(…) use_*(…) check(run_dont_test = FALSE, …) build(…) install(…) $3"/ϦϙδτϦ֎͔ΒͷύοέʔδΠϯετʔϧ ܧଓతΠϯςάϨʔγϣϯɺHJUϦϙδτϦʜ
    Πϯϑϥ੔උ # \SPYZHFO^
    \UFTUUIBU^ύοέʔδ
    ʹΑΔࣗಈυΩϡϝϯτੜ੒ɺ
    ςετ࣮ߦ
32. まさに "神業" ggmap::ggimage(hadley)
33. 2014~ {dplyr}
34. \EQMZS^ύοέʔδ σʔλϑϨʔϜૢ࡞ʹಛԽʢલॲཧͷաఔɾ࡞ۀΛ੔ཧʣ $ Ճ޻
    طଘͷม਺Λݩʹ৽ͨͳม਺Λ௥Ճͨ͠Γɺ
    
    
    
    
    
    
    
    
    
    ม਺ͷ஋ʹमਖ਼ΛՃ͑Δ
    બ୒
    ม਺໊ͷมߋ΍ෆཁͳม਺ྻͷ࡟আ͢Δ
    நग़
    ಛఆͷج४ʹର͠ɺ৚݅ʹҰக͢Δ߲໨Λநग़͢Δ
    ूܭ
    ͦΕͧΕͷ߲໨ͷूܭ஋ΛٻΊΔ 3%#͔Βͷσʔλऔಘɾ42-ΫΤϦʔ࣮ߦ؀ڥ΋੔උ $
35. library(dplyr) data("Butterflies", package = "gpk") glimpse(Butterflies) \EQMZS^ύοέʔδ 
    Πϯυࠃ಺ͷ͔̑ॴͰه࿥͞Εͨ௏ྨͷ෼෍σʔλ Observations: 44 Variables: 9 $ Serial_Number (int) 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, ... $ Area (fctr) Indian Subcontinent, Western Himalaya, Western Himalaya, Western Himalaya, Western Himalaya, Western Himalaya, Western Himalaya, Western Himal... $ Locality (fctr) Indian Subcontinent, Western Himalaya, Kangra Hills, Shimla Hills, Dehradun Valley, Mussoorie Hills, Mussoorie Town, Kumaon Hills, Central Him... $ Total_Species_count (int) 1439, 417, 228, 299, 148, 323, 146, 371, 623, 962, 853, 690, 262, 423, 321, 510, 596, 276, 167, 124, 151, 109, 77, 110, 78, 78, 59, 145, 164, 1... $ Skippers (int) 307, 63, 25, 41, 22, 54, 14, 52, 125, 211, 189, 159, 27, 67, 119, 98, 132, 59, 32, 18, 19, 15, 11, 12, 14, 11, 4, 22, 19, 23, 54, 60, 20, 37, 4... $ Swallow_tails (int) 94, 31, 23, 21, 11, 23, 10, 26, 43, 69, 62, 55, 29, 38, 18, 42, 49, 13, 10, 9, 10, 7, 4, 4, 6, 5, 5, 10, 10, 11, 16, 18, 9, 15, 15, 23, 12, 5, ... $ Whites_Yellows (int) 99, 42, 37, 34, 19, 32, 13, 37, 49, 57, 52, 51, 32, 30, 1, 36, 40, 26, 19, 12, 23, 19, 20, 29, 21, 19, 18, 25, 50, 19, 21, 31, 25, 26, 30, 35, ... $ Blues (int) 458, 129, 56, 88, 42, 88, 44, 109, 185, 284, 258, 162, 48, 110, 126, 128, 166, 77, 57, 46, 38, 33, 22, 31, 20, 22, 17, 46, 41, 35, 66, 80, 50, ... $ Brush_Footed (int) 482, 152, 87, 115, 54, 126, 65, 147, 221, 342, 292, 263, 126, 178, 57, 207, 209, 101, 49, 39, 61, 35, 20, 34, 17, 21, 15, 42, 44, 42, 66, 89, 3...
36. Butterflies %>% filter(Total_Species_count >= 200 | Area %in% c("Indian Subcontinent", "North East India+ North Myanmar", "Other parts of India"), grepl("Hills", Locality), between(Serial_Number, 10, 20)) \EQMZS^ύοέʔδ
37. Butterflies %>% filter(Total_Species_count >= 200 | Area %in% c("Indian Subcontinent", "North East India+ North Myanmar", "Other parts of India"), grepl("Hills", Locality), between(Serial_Number, 10, 20)) \EQMZS^ύοέʔδ ࿦ཧ࿨ ਅِ஋Λฦؔ͢਺ ൣғΛநग़ʜ
    \EQMZS^ͷิॿؔ਺ ཁૉͷू߹
38. 
    ʮϓϦΩϡΞʯγϦʔζͷ"1*
    !TVF
    req.girls <- GET("https://rubicure.herokuapp.com/ girls.json") %>% content() req.girls %>% class() # [1] "list" req.girls %>% .[[1]] %>% names() # [1] "girl_name" "human_name" "precure_name" "cast_name" "created_date" "color" "birthday" "transform_message" # [9] "extra_names" "attack_messages" "transform_calls" \EQMZS^
    
    \QVSSS^
    
    \IUUS^
39. req.girls %>% map_df(., ~ .[c("color")]) %>% group_by(color) %>% tally(sort = TRUE) Source: local data frame [8 x 2] color n <chr> <int> 1 pink 10 2 blue 9 … 7 green 2 8 black 1 JF
    TVNNBSJTF O
    
    O
    
    
    
    
    
    
    
    
    
    
    BSSBOHF O \EQMZS^
    
    \QVSSS^
    
    \IUUS^
40. ໰୊ GSPN
    ŴŞťŚŷŵŕţſŞŽŲƄ LJOUPSF
41. ೴τϨԋश 
    TFMFDUIFMQFST var_1 var_2 third_variable 4th_variable start_with("var_") end_with("variable") contains("_") matches("[[:punct:]]") num_range("var_", 1:2) one_of("third_variable") everything() :PVS
    5VSO
    Ͳͷม਺͕બ୒͞ΕΔ͔ʁ
42. 
    TFMFDUIFMQFST var_1 var_2 third_variable 4th_variable start_with("var_") ◦ ◦ ✖ ✖ end_with("variable") ✖ ✖ ◦ ◦ contains("_") ◦ ◦ ◦ ◦ matches("[[:punct:]]") ◦ ◦ ◦ ◦ num_range("var_", 1:2) ◦ ◦ ✖ ✖ one_of("third_variable") ✖ ✖ ◦ ✖ everything() ◦ ◦ ◦ ◦ ೴τϨԋश
43. point <- c(-2:2, NA) if_else(point < 0, "あし", "よし", "なし") soum <- c("そうむ", "ほうむ", "ポエム", "のうむ") na_if(soum, "ポエム") nth(soum, 2) 
    ϕΫτϧΛର৅ʹͨ͠ૢ࡞ :PVS
    5VSO ೴τϨԋश
44. point <- c(-2:2, NA) if_else(point < 0, "あし", "よし", "なし") # [1] "あし" "あし" "よし" "よし" "よし" "なし" soum <- c("そうむ", "ほうむ", "ポエム", "のうむ") na_if(soum, "ポエム") # [1] "そうむ" "ほうむ" NA "のうむ" nth(soum, 2) # [1] "ほうむ" 
    ϕΫτϧΛର৅ʹͨ͠ૢ࡞ ೴τϨԋश
45. library(magrittr); library(stringr) df.q1 <- data_frame( point = c("10/13/13/16", "9/12/8/15")) %$% point %>% str_split_fixed(., "/", 4) %>% as.data.frame() %>% set_colnames(str_c("var", 1:4)) ೴τϨԋश
46. library(tibble) df.q2 <- frame_data( ~fruit, ~sales, "apple/banana/orange", "1/2/4", "orange/melon/applef/pear", "1/3/5/1") %$% data_frame( fruit = fruit %>% str_split("/"), sales = sales %>% str_split("/")) ೴τϨԋश
47. df.q1 var1 var2 var3 var4 1 10 13 13 16 2 9 12 8 15 df.q2 Source: local data frame [2 x 2] fruit sales <list> <list> 1 <chr [3]> <chr [3]> 2 <chr [4]> <chr [4]> library(tidyr) df.q3 <- df.q2 %>% unnest() ೴τϨԋश
48. df.q3 Source: local data frame [7 x 2] fruit sales <chr> <chr> 1 apple 1 2 banana 2 3 orange 4 4 orange 1 5 melon 3 6 apple 5 7 pear 1 ೴τϨԋश
49. with 2015~ 2014~ adv anced & DPLyr
50. σʔλ͕࣋ͭ֊૚ੑ " # $ ʜ ʜ B ʜ ʜ B ʜ ʜ C ʜ ʜ D ʜ ʜ D ʜ ʜ D ^ ^ ^ EBUB B EBUB EBUB D EBUB EBUB C EBUB $ SFTVMU SFTVMU B ʜ ʜ C ʜ ʜ D ʜ ʜ
51. σʔλ͕࣋ͭ֊૚ੑ " # $ ʜ ʜ B ʜ ʜ B ʜ ʜ C ʜ ʜ D ʜ ʜ D ʜ ʜ D ^ ^ ^ EBUB B EBUB EBUB D EBUB EBUB C EBUB $ SFTVMU SFTVMU B ʜ ʜ C ʜ ʜ D ʜ ʜ UJEZSOFTU UJEZSVOOFTU QVSSSNBQ EQMZSEP
52. library(tidyr) by_area <- group_by(Butterflies, Area) %>% nest() \EQMZS^
    
    \UJEZS^ Source: local data frame [8 x 2] Area data <fctr> <list> 1 Indian Subcontinent <tbl_df [1,8]> 2 Western Himalaya <tbl_df [7,8]> 3 Central Himalaya <tbl_df [1,8]> 4 North East India+ North Myanmar <tbl_df [9,8]> 5 Other parts of India <tbl_df [21,8]> 6 Sri Lanka <tbl_df [1,8]> 7 North Myanmar <tbl_df [3,8]> 8 South Myanmar <tbl_df [1,8]> # HSPVQ@CZ Ͱࢦఆ
    ͨ͠ά ϧʔϓ͝ͱʹ UCM@EGΫϥε
    ΦϒδΣΫτͱͯ͠ ֨ೲ
53. library(purrr) mod_res <- map(by_area$data, ~ glm(Total_Species_count ~ Skippers + Blues, data = ., family = poisson)) \EQMZS^
    
    \QVSSS^ # άϧʔϓ͝ͱʹؔ਺Λద༻ʢNBQQJOHʣ
54. mod_res[[4]] mod_res %>% class() # [1] "list" \EQMZS^
    
    \QVSSS^ # Call: glm(formula = Total_Species_count ~ Skippers + Blues, family = poisson, data = .) Coefficients: (Intercept) Skippers Blues 5.380924 0.003370 0.002889 Degrees of Freedom: 8 Total (i.e. Null); 6 Residual Null Deviance: 932.4 Residual Deviance: 100.3 AIC: 178.6 NBQ ͷฦΓ஋͸
    ϦετΦϒδΣΫτ
    NBQ@EG Ͱ͸
    σʔλϑϨʔϜ
55. library(broom) (do_mod_res <- group_by(Butterflies, Area) %>% do(mod = glm(Total_Species_count ~ Skippers + Blues, data = ., family = poisson) %>% glance())) do_mod_res %>% class() # [1] "rowwise_df" "tbl_df" "tbl" "data.frame" \EQMZS^
    
    \CSPPN^
56. \EQMZS^
    
    \CSPPN^ do_mod_res %>% unnest() Source: local data frame [8 x 8] Area null.deviance df.null logLik AIC BIC deviance df.residual <fctr> <dbl> <int> <dbl> <dbl> <dbl> <dbl> <int> 1 Central Himalaya 0.0000 0 -4.136346 10.272691 8.272691 0.000000e+00 0 2 Indian Subcontinent 0.0000 0 -4.554848 11.109697 9.109697 -1.723066e-13 0 3 North East India+ North Myanmar 932.3645 8 -86.309184 178.618368 179.210041 1.002874e+02 6 4 North Myanmar 55.5590 2 -10.891251 27.782502 25.078339 2.664535e-14 0 5 Other parts of India 630.0408 20 -94.738568 195.477137 198.610704 4.746547e+01 18 6 South Myanmar 0.0000 0 -4.253793 10.507587 8.507587 5.240253e-14 0 7 Sri Lanka 0.0000 0 -3.663752 9.327503 7.327503 -4.884981e-15 0 8 Western Himalaya 255.5851 6 -33.529065 73.058130 72.895860 1.535052e+01 4 #֊૚ߏ଄Λ࣋ͭσʔλϑϨʔϜΛ࠶݁߹
57. ਐԽ͠ଓ͚Δ
    \EQMZS^ύοέʔδ
58. Ұํʜͱ͋Δࣝऀͷಗ໊ݟղ ˞ϓϥΠόγʔอޢͷͨΊɺ
    Ұ෦΅͔͠ΛೖΕ͍ͯ·͢ \EQMZS^ͱ͸
    Կͩͬͨͷ͔ʜ a͜ͷޙͷ-5ʹظ଴ʂ
59. Hadley Ecosystem 2016~
60. )BEMFZ
    &DPTZTUFNͷ৽੕
61. )BEMFZ
    &DPTZTUFNͷ৽੕ \EQMZS^ͷؔ਺ΛϚϧνίΞॲཧ
    QBSUJUJPO Ͱάϧʔϓ෼ׂɺDPMMFDU Ͱ݁ՌΛճऩ
    Ϋϥελʔ਺͸TFU@EFGBVMU@DMVTUFS ؔ਺Ͱௐ੔Մೳ
62. )BEMFZ
    &DPTZTUFNͷ৽੕ \EQMZS^ͷྲّྀΛܧঝͨ͠σʔλϑϨʔϜੜ੒
    EQMZSBEE@SPXOBNFT
    ͔Β
    
    UJCCMFSPXOBNFT@UP@DPMVNO
    ΁
    σʔλϑϨʔϜΛྻ͝ͱʹهड़͢ΔGSBNF@EBUB ͱ͔ศར͔΋
    
63. )BEMFZ
    &DPTZTUFNͷ৽੕ 3
    1ZUIPO
    ؒͰޓ׵ੑͷ͋Δ௒ܰྔσʔλܗࣜ
    $MPVEFSBͷ8FT
    .D,JOOFZͱͷڞಉ։ൃ
    "QBDIF
    "SSPX
    1SPKFDUʜ
    ରԠܗࣜ΋ͬͱ૿͑Δ͔΋
    3ඪ४ͷอଘػߏʹΑΔσʔλͷอଘ͕ܰྔʹͳΔ͜ͱ΋
64. ホクソエム神話 第四の書 "R4DS"
65. ѹ౗తےྗΛ
    ಘΑ͏ʂ
66. enjoy!!!