Rによるデータ可視化と地図表現

2022年8月28日に実施された統計思考院 オンラインワークショップ「探索的ビッグデータ解析と再現可能研究」https://sites.google.com/view/ws-ebda-rr-2022/ の発表資料です。
スライドの中のRコードは https://github.com/uribo/220828ism_ws にあげています。

Transcript

1. 2022-08-28@
   統
   計
   思
   考
   院
   探
   索
   的
   解
   析
   再
   現
   可
   能
   研
   究
   ⽠
   ⽣
   真
   也
   徳
   島
   ⼤
   学
   デ
   ザ
   イ
   ン
   型
   AI
   教
   育
   研
   究
   セ
   ン
   タ
   

   View Slide

2. ⾃⼰紹介
   (Uryu Shinya)
   u_ribo
   AI
   2021 10
   uribo
   2
   

   View Slide

3. Showcase
   IUUQTHJUIVCDPNVSJCP%BZ.BQ$IBMMFOHF
   IUUQTHJUIVCDPNVSJCPVSCBOTQBDFNBQ
   IUUQTHJUIVCDPNVSJCPWPMDBOP
   IUUQTHJUIVCDPNVSJCPTIPXZPVSTUSJQFT
   IUUQTHJTUHJUIVCDPNVSJCPCCDEFCFCBBDC
   IUUQTHJTUHJUIVCDPNVSJCPGFCFGDFEEBCEFEG
   IUUQTHJTUHJUIVCDPNVSJCPEEBDBGBFDEGCE
   

   View Slide

4. IUUQTXXXLTQVCDPKQCPPLEFUBJMIUNM
   分析 可視化⼊⾨
   
   
   (2021).
   Kieran Healy
   
   
   Data Visualization: A practical introduction
   
   
   (
   2
   019
   ). Princeton University Press.
   4
   

   View Slide

5. 品書
   1. 総
   3.
   ggplot
   2
   2.
   3
   5
   ৮Εͳ͍಺༰
   

   View Slide

6. 対象 ⼈
   R ggplot
   2
   6
   R
   

   View Slide

7. 前置
   7
   {ggplot
   2
   }
   GitHub
   {magrittr} %>% R
   4
   .
   1
   .
   0
   |> び
   R
   4
   .
   2
   .
   1
   Speaker Deck
   01 # σʔλૢ࡞ͷͨΊͷύοέʔδ
   
   
   02 library(dplyr, warn.conflicts = FALSE)
   
   
   03 penguins_xy
   < -
   
   04 palmerpenguins
   : :
   penguins
   | >
   
   05 select(flipper_length_mm, bill_length_mm, species)
   | >
   
   06 f
   i
   lter(!is.na(flipper_length_mm))
   

   View Slide

8. 前置
   8
   

   01 packages
   < -
   c(
   
   
   02 "tidyverse", "sf", "zipangu", "tabularmaps", "geofacet",
   
   
   03 "palmerpenguins", "datasauRus", "gt", "gapminder", "statebins",
   
   
   04 "ggtext", "ggrepel", "gghighlight", "patchwork",
   
   
   05 "rnaturalearth", "ggokabeito")
   
   
   06 install.packages(setdiff(packages, rownames(installed.packages())))
   
   
   07
   
   
   08 ropensci_pkgs
   < -
   c("rnaturalearthhires")
   
   
   09 install.packages(setdiff(ropensci_pkgs, rownames(installed.packages())),
   
   
   10 repos = "https:
   / /
   ropensci.r
   -
   universe.dev")
   
   
   11
   
   
   12 wilkelab_pkgs
   < -
   c("gridtext")
   
   
   13 install.packages(setdiff(wilkelab_pkgs, rownames(installed.packages())),
   
   
   14 repos = "https:
   / /
   wilkelab.r
   -
   universe.dev")
   
   
   15
   
   
   16 uris_pkgs
   < -
   c("ssdse")
   
   
   17 install.packages(setdiff(uris_pkgs, rownames(installed.packages())),
   
   
   18 repos = "https:
   / /
   uribo.r
   -
   universe.dev")
   

   View Slide

9. View Slide

10. ࣮ࡏ ৘ใ
    :
    : #609527
    :
    : 29.3cm
    : 1
    実在 情報、 変換 過程 情報量 損失 ⽣
    写像・ ・可視化
    σʔλ
     
    mapping
    #020202
    10
    

    View Slide

11. ࣮ࡏ ৘ใ
    情報 含 上 写像
    写像・ ・可視化
    ՄࢹԽ
    σʔλ
    mapping
    x
    y
    x
    y
    etc.
    11
    

    View Slide

12. 実在 情報、可視化 写像 失 情報量 補
    写像・ ・可視化
    12
    良 悪 ⼀ 観点: 読 取 情報 性質
    

    View Slide

13. 可視化 重要性
    実在 情報 効率的 伝 ⽅法 必要
    σʔλՄࢹԽҎ֎ͷख๏ͷྫ
    13
    

    View Slide

14. 可視化 重要性
    例: 可視化 重要性 説明 例
    総
    x,y 総
    14
    

    View Slide

15. 可視化 重要性
    Alberto Cairo,
    2 0
    1 6
    :
    15
    

    View Slide

16. 法則 基 情報 理解
    16
    ۙ઀ ྨࣅ ғ͍ࠐΈ
    ด࠯ ࿈ଓੑ ઀ଓ
    視覚的情報 基 、要素間 関係性 推論
    

    View Slide

17. 統計 歴史
    17
    ग़య൧ౢوࢠ σʔλࢹ֮Խͷਓྨ࢙άϥϑͷൃ໌͔Β࣌ؒͱۭؒͷՄࢹԽ·Ͱ੨౔ࣾ
    

    View Slide

18. 近代的 発明者 ・
    18
    8JMMJBN1MBZGBJS 1VCMJDEPNBJO WJB8JLJNFEJB$PNNPOT
    IUUQTVQMPBEXJLJNFEJBPSHXJLJQFEJBDPNNPOT1MBZGBJS@5JNF4FSJFTQOH
    18 総
    

    1786
    The Commercial and Political Atlas,
    1
    7 86
    Statistical Breviary,
    1
    80
    1
    

    View Slide

19. 1
    .
    
    
    2
    . 総 肢
    
    
    3
    . 18
    

    View Slide

20. View Slide

21. {ggplot2}
    The Grammar of Graphics (Wilkinson,
    20
    05
    )
    
    
    R
    The Grammar of Graphics
    21
    

    View Slide

22. 可視化、ggplot2 推 理由
    22
    

    View Slide

23. base
    23
    縛
    IUUQTHJUIVCDPN1PJTPO"MJFOCBTFHSBQIJDT
    

    View Slide

24. ggplot2 構成 3 柱
    1.
    2.
    3.
    24
    

    View Slide

25. ggplot2 利⽤可能
    01 # install.packages("ggplot2")
    
    
    02 # ggplot2ύοέʔδͷಡΈࠐΈ
    
    
    03 # ൃදࢿྉͰ͸2022೥8݄࣌఺ͰͷCRAN࠷৽όʔδϣϯͰ͋Δ3.3.6Λར༻͠·͢
    
    
    04 library(ggplot2)
    
    
    05
    
    
    06 # tidyverseύοέʔδʹ಺แ͞Ε͍ͯΔͨΊɺͪ͜ΒΛಡΈࠐΜͰ΋OKͰ͢
    
    
    07 # install.packages("tidyverse")
    
    
    08 # library(tidyverse)
    25
    

    View Slide

26. Tidyverse
    26
    ಡΈࠐΈ ੔ܗ Ճ޻
    ՄࢹԽ
    Ϟσϧ
    ఻ୡ
    Garrett and Hadley (2016)
    

    

    View Slide

27. ggplot2 第⼀歩
    fl
    ipper_length_mm bill_length_mm species
    181 39.1 Adelie
    186 39.5 Adelie
    195 40.3 Adelie
    … … …
    198 50.2 Chinstrap
    27
    01 # σʔλૢ࡞ͷͨΊͷύοέʔδ
    
    
    02 library(dplyr, warn.conflicts = FALSE)
    
    
    03 penguins_xy
    < -
    
    04 palmerpenguins
    : :
    penguins
    | >
    
    05 select(flipper_length_mm, bill_length_mm, species)
    | >
    
    06 f
    i
    lter(!is.na(flipper_length_mm))
    

    View Slide

28. ggplot2 第⼀歩
    01 ggplot(data = penguins_xy)
    テ
    data.frame
    data penguins_xy
    28
    

    View Slide

29. 01 ggplot(data = penguins_xy)
    02 aes(x = flipper_length_mm,
    
    
    03 y = bill_length_mm)
    ggplot2 第⼀歩
    デ
    x y
    

    +
    +
    

    29
    data penguins_xy
    

    x
    fl
    ipper_length_mm
    

    y bill_length_mm
    

    View Slide

30. 01 ggplot(data = penguins_xy) +
    
    
    

    View Slide

31. 審美的 (aesthetic) 要素
    x
    fl
    ipper_length_mm y bill_length_mm
    

    species
    31
    01 ggplot(data = penguins_xy) +
    
    
    

    View Slide

32. 審美的 (aesthetic) 要素
    aes(άϥϑͷதͷ໾ׂ = ม਺)
    x, y
    color,
    fi
    ll
    size
    shape, linetype
    alpha
    group
    32
    linewidth* linewidth v
    3
    .
    4
    .
    0
    

    View Slide

33. 審美的 (aesthetic) 要素
    33
    審美的要素 機能 値 対応
    

    View Slide

34. 審美的要素 指定⽅法
    34
    01 # 1. aes()͸ϨΠϠͱͯ͠ggplotΦϒδΣΫτʹ௥Ճͯ͠΋ྑ͍
    
    
    02 ggplot(data = penguins_xy) +
    
    
    03 aes(x = flipper_length_mm,
    
    
    04 y = bill_length_mm)
    
    
    05
    
    
    06 # 2. ggplot(mapping = aes(...))ͱͯ͠༩͑ͯ΋ྑ͍
    
    
    07 ggplot(data = penguins_xy,
    
    
    08 mapping = aes(x = flipper_length_mm,
    
    
    09 y = bill_length_mm))
    
    
    10
    
    
    11 # 3. ggplot(data = ,mapping = aes(x = ,y = ))͸҉໧తʹҾ਺Λলུͯ͠هड़Ͱ͖Δ
    
    
    12 ggplot(penguins_xy,
    
    
    13 aes(flipper_length_mm, bill_length_mm))
    

    View Slide

35. 審美的要素 指定⽅法
    35
    共通 内
    04 color = species,
    
    
    05 group = species
    aes(color = species)
    
    
    aes(group = species)
    (species)
    

    01 ggplot(data = penguins_xy) +
    
    
    

    View Slide

36. ⼈⼝ 可視化
    01 dplyr
    : :
    glimpse(df_ssdse_b)
    
    
    02 #> Rows: 564
    
    
    03 #> Columns: 7
    
    
    04 #> $ prefecture "๺ւಓ", "๺ւಓ", "๺ւಓ", "๺ւಓ", …
    
    
    05 #> $ year 2019, 2018, 2017, 2016, 2015, 2014, 2013, …
    
    
    06 #> $ population 5250000, 5286000, 5320000, 5352000, 5381733, …
    
    
    07 #> $ birth_male 15988, 16681, 17503, 17888, 18838, 19010, …
    
    
    08 #> $ birth_female 15032, 15961, 16537, 17237, 17857, 18048, …
    
    
    09 #> $ spending 294682, 281054, 286698, 287325, 272124, …
    
    
    10 #> $ food_expenses 72912, 69044, 69640, 69445, 65912, 65450, …
    
    
    11
    
    
    12 df_ssdse_b2019
    < -
    
    13 df_ssdse_b
    | >
    
    14 f
    i
    lter(year
    = =
    2019)
    符
    符
    36
    統計 提供 教育⽤標準 (SSDSE-B) 加⼯ IUUQTXXXOTUBDHPKQVTFMJUFSBDZTTETF
    

    View Slide

37. ⼈⼝ 可視化
    37
    01 df_ssdse_b2019
    | >
    
    02 ggplot() +
    
    
    03 aes(prefecture, population) +
    
    
    04 geom_bar(stat = "identity")
    x
    

    View Slide

38. ⼈⼝ 可視化
    38
    01 df_ssdse_b2019
    
    
    06 coord_flip()
    

    View Slide

39. ⼈⼝ 可視化
    39
    02 # ͓͓ΑͦͷҢ౓ͷॱ൪ʹ഑ஔ͢Δ
    
    
    03 mutate(prefecture =
    
    
    04 forcats
    : :
    fct_rev(
    
    
    05 forcats
    : :
    fct_inorder(
    
    
    06 prefecture)))
    | >
    
    01 df_ssdse_b2019
    
    
    

    View Slide

40. ggplot2 提供 関数群
    ggplot(data = )
    aes()
    geom_*()
    40
    ggplot
    

    View Slide

41. ggplot2 提供 関数群
    stat_*()
    scale_*()
    coord_*()
    facet_*()
    theme(),theme_*()
    41
    ggplot(data = )
    aes()
    geom_*()
    ggplot
    

    View Slide

42. : geom_*()
    種類、描画 対象 定義
    42
    geom_density()
    geom_sf()
    geom_smooth()
    geom_point()
    geom_line()
    geom_bar()
    geom_histogram()
    geom_boxplot()
    

    View Slide

43. 種類、描画 対象 定義
    43
    geom_bar(stat = "identity")
    df_ssdse_b2019
    
    
    

    View Slide

44. 統計処理
    44
    stat = "identity" ?
    geom_bar
    
    
    #> function (mapping = NULL, data = NULL, stat = "count", position = "stack",
    
    
    #>
    . . .
    , width = NULL, na.rm = FALSE, orientation = NA, show.legend = NA,
    
    
    #> inherit.aes = TRUE)
    geom_bar(stat = "count")
    01 df_ssdse_b2019
    | >
    
    02 f
    i
    lter_shikoku()
    | >
    
    03 ggplot() +
    
    
    04 aes(prefecture, population) +
    
    
    05 geom_bar(stat = "identity")
    

    View Slide

45. 統計処理 : stat_*()
    上 変形・集計 ⾏
    45
    stat = "count" ... aes(x = ) y
    01 tibble
    : :
    tibble(
    
    
    02 group = c(rep("a", 4), rep("b", 2), "c")
    
    
    03 )
    | >
    
    04 ggplot() +
    
    
    05 aes(x = group) +
    
    
    06 geom_bar(stat = "count")
    01 df_ssdse_b2019
    | >
    
    02 f
    i
    lter_shikoku()
    | >
    
    03 ggplot() +
    
    
    04 aes(prefecture, population) +
    
    
    05 geom_bar(stat = "identity")
    stat = "identity" ... aes(y = )
    stat_count() group
    

    View Slide

46. 統計処理 : stat_*()
    geom_*(stat = ) or stat_*() 利⽤
    46
    01 p
    < -
    
    02 df_ssdse_b2019
    | >
    
    03 ggplot() +
    
    
    04 aes(x = spending)
    p + geom_bar(stat = "bin", bins = 10) p + geom_histogram(bins = 10) p + stat_bin(bins = 10)
    

    View Slide

47. x
    47
    変数 属性 審美的要素
    

    View Slide

48. : scale__*()
    48
    01 df_ssdse_b
    | >
    
    02 f
    i
    lter_shikoku()
    | >
    
    03 ggplot() +
    
    
    04 aes(year, population,
    
    
    05 group = prefecture,
    
    
    06 color = prefecture) +
    
    
    07 geom_line()
    

    View Slide

49. : scale__*()
    49
    01 df_ssdse_b
    
    
    08 scale_x_continuous(
    
    
    09 breaks = seq.int(2008, 2019, by = 2)) +
    
    
    10 scale_y_log10() +
    
    
    11 scale_color_viridis_d()
    x y
    

    View Slide

50. 座標系
    50
    値 配置 決
    2
    ggplot
    2
    x y
    x y
    

    ) x 3
    

    y 10 (0,10, 20,...)
    

    View Slide

51. 座標系: coord_*()
    51
    p
    < -
    
    ggplot(data = penguins_xy) +
    
    
    aes(x = flipper_length_mm,
    
    
    y = bill_length_mm) +
    
    
    geom_point()
    p +
    
    
    coord_flip()
    x y x y
    p +
    
    
    coord_f
    i
    xed(ratio = 1)
    

    View Slide

52. 52
    任意 変数 基 分割、 形式 表⽰
    'SBODJT"8BMLFS 1VCMJDEPNBJO WJB8JLJNFEJB$PNNPOT
    IUUQTVQMPBEXJLJNFEJBPSHXJLJQFEJBDPNNPOTG1FSTPOT@8JUI@(BJOGVM@0DDVQBUJPOT@BOE@"UUFOEJOH@4DIPPM@JO@KQH
    9 (1874)
    

    View Slide

53. : facet_*()
    53
    01 p
    < -
    
    02 df_ssdse_b
    | >
    
    03 f
    i
    lter_shikoku()
    | >
    
    04 select(year, prefecture, starts_with("birth"))
    | >
    
    05 tidyr
    : :
    pivot_longer(cols = starts_with("birth"),
    
    
    06 names_to = "gender",
    
    
    07 values_to = "value",
    
    
    08 names_pref
    i
    x = "birth_")
    | >
    
    09 ggplot() +
    
    
    10 aes(year, value) +
    
    
    11 geom_line(aes(group = prefecture, color = prefecture)) +
    
    
    12 scale_x_continuous(breaks = seq.int(2008, 2019, by = 2)) +
    
    
    13 scale_color_manual(values = custom_pals[1
    :
    4])
    
    
    14
    
    
    15 p$data
    

    View Slide

54. : facet_*()
    54
    p_facet
    < -
    p +
    
    
    facet_wrap(~ gender,
    
    
    scales = "free_y",
    
    
    ncol = 2)
    p +
    
    
    facet_grid(rows = vars(gender),
    
    
    scales = "free_y")
    

    View Slide

55. : theme(), theme_*()
    55
    ggplot2 豊富 存在
    

    View Slide

56. : theme(), theme_*()
    56
    01 p_facet +
    
    
    02 theme_classic() +
    
    
    03 theme(
    
    
    04 strip.background = element_rect(f
    i
    ll = "gray"),
    
    
    05 strip.text = element_text(color = "white",
    
    
    06 face = "bold"),
    
    
    07 legend.position = "top",
    
    
    08 axis.title = element_text(color = "#28a87d"),
    
    
    09 axis.text = element_text(color = "#28a87d"),
    
    
    10 axis.ticks = element_line(color = "#28a87d"))
    細 部位 theme()、element_*() 制御
    element_blank()
    element_rect()
    element_line()
    element_text()
    

    View Slide

57. {ggtext} ⽂字 装飾
    57
    01 library(palmerpenguins)
    
    
    02 p_penguins_scatter
    < -
    
    03 penguins
    | >
    
    04 ggplot() +
    
    
    05 aes(flipper_length_mm, bill_length_mm) +
    
    
    06 geom_point(aes(color = species),
    
    
    07 alpha = 0.5,
    
    
    08 show.legend = FALSE) +
    
    
    09 geom_smooth(method = "lm",
    
    
    10 aes(group = species),
    
    
    11 se = FALSE,
    
    
    12 color = "#0E0E0E") +
    
    
    13 scale_color_manual(values = custom_pals) +
    
    
    14 theme_light() +
    
    
    15 labs(title = "ϖϯΪϯͷମͷେ͖͞ͷؔ܎",
    
    
    16 x = "ཌྷͷ௕͞(mm)",
    
    
    17 y = "ͪ͘͹͠ͷ௕͞(mm)")
    

    View Slide

58. {ggtext} ⽂字 装飾
    58
    01 library(ggtext)
    
    
    03 labs(title = "
    * *
    ϖϯΪϯͷମͷେ͖͞ͷؔ܎
    * *
    ",
    
    
    04 x = "*F
    l
    ipper length
    *
    (mm)",
    
    
    05 y = "*Bill length
    *
    (mm)") +
    
    
    06 theme(
    
    
    07 plot.title = ggtext
    : :
    element_markdown(
    
    
    08 color = custom_pals[4],
    
    
    09 f
    i
    ll = "gray80"
    
    
    10 ),
    
    
    11 axis.title.x = element_markdown(),
    
    
    12 axis.title.y = element_markdown())
    02 p_penguins_scatter +
    Markdown
    

    View Slide

59. {gghighlight} 使 強調
    59
    02 p_penguins_scatter +
    
    
    03 gghighlight(species
    = =
    "Adelie")
    

    View Slide

60. 対 注釈 {ggrepel} 追加
    60
    01 library(ggrepel)
    
    
    02 p_gapminder
    < -
    
    03 gapminder
    : :
    gapminder
    | >
    
    04 f
    i
    lter(year
    = =
    2007,
    
    
    05 continent
    = =
    "Americas")
    | >
    
    06 ggplot() +
    
    
    07 aes(gdpPercap, lifeExp) +
    
    
    08 geom_point()
    

    View Slide

61. 対 注釈 {ggrepel} 追加
    61
    01 p_gapminder +
    
    
    02 ggrepel
    : :
    geom_text_repel(
    
    
    03ɹ aes(label = country))
    

    View Slide

62. {ggrepel} {gghighlight} 合 技
    62
    01 p_gapminder +
    : :
    gghighlight(
    
    
    03 lifeExp < 70) +
    
    
    
    

    View Slide

63. 1
    . {ggplot
    2
    } R
    
    
    2
    .
    
    
    3
    . {ggplot
    2
    }
    

    View Slide

64. View Slide

65. 地図表現 過程
    65
    空間 他 定量的 紐付 、地図上 表⽰
    ஍ཧۭؒ৘ใ σʔλ ओ୊ਤ
    +
    etc.
    

    View Slide

66. 階級区分図 誕⽣
    66
    $IBSMFT%VQJO
    1VCMJDEPNBJO WJB8JLJNFEJB$PNNPOT
    IUUQTDPNNPOTXJLJNFEJBPSHXJLJ'JMF$BSUF@
    fi
    HVSBUJWF@EF@MJOTUSVDUJPO@QPQVMBJSF@EF@MB@'SBODFKQH
    1826
    1
    
 

    18 19
    

    

    View Slide

67. 01 ne_jpn
    
    
    02 #> Simple feature collection with 47 features and 3 f
    i
    elds
    
    
    03 #> Geometry type: MULTIPOLYGON
    
    
    04 #> Dimension: XY
    
    
    05 #> Bounding box: xmin: 122.9382 ymin: 24.2121 xmax: 153.9856 ymax: 45.52041
    
    
    06 #> CRS
    : +
    proj=longlat
    +
    datum=WGS84
    +
    no_defs
    +
    ellps=WGS84
    +
    towgs84=0,0,0
    
    
    07 #> First 10 features:
    
    
    08 #> iso_3166_2 prefecture region geometry
    
    
    09 #> 1 01 ๺ւಓ Hokkaido MULTIPOLYGON (((143.8965 44
    . . .
    
    10 #> 2 02 ੨৿ݝ Tohoku MULTIPOLYGON (((139.9438 40
    . . .
    
    11 #> 3 03 ؠखݝ Tohoku MULTIPOLYGON (((141.681 40....
    
    
    12 #> 4 04 ٶ৓ݝ Tohoku MULTIPOLYGON (((141.6403 38
    . . .
    
    13 #> 5 05 ळాݝ Tohoku MULTIPOLYGON (((139.8809 39
    . . .
    地図 作
    67
    地理空間情報 拡張可能 {sf} 便利
    sf
    sfg(geometry)
    sfc (column)
    sf
    3
    Natural Earth {rnaturalearth}
    

    View Slide

68. 地図 作
    68
    01 ne_jpn
    | >
    
    02 ggplot() +
    
    
    03 aes() +
    
    
    04 geom_sf()
    

    View Slide

69. 地図 作
    69
    01 ne_jpn
    | >
    
    02 ggplot() +
    
    
    03 aes(f
    i
    ll = prefecture) +
    
    
    04 geom_sf(show.legend = FALSE)
    

    View Slide

70. 地理空間情報 facet
    70
    

    View Slide

71. 地理空間情報 facet
    71
    01 sf_jpn_population
    | >
    
    02 f
    i
    lter(prefecture %in% c("ಙౡݝ", "߳઒ݝ", "Ѫඤݝ", "ߴ஌ݝ"))
    | >
    
    03 ggplot() +
    
    
    04 aes(f
    i
    ll = population) +
    
    
    05 geom_sf(color = "transparent") +
    
    
    06 guides(f
    i
    ll = guide_colorbar(title = "૯ਓޱ")) +
    
    
    07 scico
    : :
    scale_f
    i
    ll_scico(palette = "imola",
    
    
    08 labels = zipangu
    : :
    label_kansuji(),
    
    
    09 breaks = c(700000, 1000000, 1400000)) +
    
    
    10 labs(title = "࢛ࠃ4ݝͷਓޱ",
    
    
    11 subtitle = "2008೥͔Β2019೥ͷਪҠ",
    
    
    12 caption = "Source: ౷ܭηϯλʔ ڭҭ༻ඪ४σʔληοτ\nSSDSE-ݝผਪҠʢSSDSE-Bʣ") +
    
    
    13 theme_void() +
    
    
    14 theme(legend.position = "top",
    
    
    15 legend.key.width = unit(3.0, "line")) +
    
    
    16 facet_wrap(~ year, nrow = 2)
    

    View Slide

72. 地図表現 ⾏ 際 遭遇 課題
    72
    2.
    1.
    3.
    4.
    

    View Slide

73. 地球(3次元) 平⾯(2次元) 表現 ?
    73
    球体 平⾯ 表現 過程 歪 ⽣
    → 地図 嘘
    

    View Slide

74. 座標参照系（Coordinate Reference System: CRS）
    74
    空間 座標 地球上 位置 特定 仕組
    ஍ཧ࠲ඪܥ
    ౤Ө࠲ඪܥ
    

    

    2 XY
    
    
    
    
    UTM
    
    
    

    View Slide

75. 地図投影法 変換
    75
    01 # Natural Earth͔ΒશٿϙϦΰϯΛऔಘ
    
    
    02 ne_world
    < -
    
    03 rnaturalearth
    : :
    ne_countries(scale = 10,
    
    
    04 returnclass = "sf")
    
    
    05
    
    
    06 p
    < -
    
    07 ne_world
    | >
    
    08 ggplot() +
    
    
    09 geom_sf()
    sf
    

    View Slide

76. 地図投影法 変換
    76
    1
    . coord_sf()
    2. CRS
    01 # ϞϧϫΠσਤ๏ʹΑΔੈք஍ਤͷඳը
    
    
    02 p +
    
    
    03 coord_sf(crs = "
    +
    proj=moll")
    01 # st_transform()ʹΑΔ࠲ඪࢀরܥͷมߋ
    
    
    02 ne_world_moll
    < -
    
    03 sf
    : :
    st_transform(ne_world, crs = "
    +
    proj=moll")
    
    
    04
    
    
    05 ggplot(data = ne_world_moll) +
    
    
    06 geom_sf()
    

    View Slide

77. 地理空間 属性 紐付
    77
    prefecture geometry
    Hokkaido POLYGON(…)
    Aomori-ken POLYGON(…)
    … …
    Okinawa-ken POLYGON(…)
    prefecture population
    ๺ւಓ 5.281
    ੨৿ݝ 1.249
    … …
    ԭೄݝ 1.457
    distinct geometry
    Hokkaido POLYGON(…)
    Honsyu POLYGON(…)
    … …
    Kyusyu POLYGON(…)
    prefecture geometry
    ๺ւಓ POLYGON(…)
    ੨৿ݝ POLYGON(…)
    … …
    ԭೄݝ POLYGON(…)
    ×
    ×
    ✓ +
    NG
    NG
    

    =
    prefecture population geometry
    ๺ւಓ 5.281 POLYGON(…)
    ੨৿ݝ 1.249 POLYGON(…)
    … … …
    ԭೄݝ 1.457 POLYGON(…)
    

    View Slide

78. 必要 応 名寄 集計単位 変更
    78
    01 ne_jpn_region
    < -
    
    02 ne_jpn
    | >
    
    03 group_by(region)
    | >
    
    04 summarise()
    01 ne_jpn_region
    | >
    
    02 left_join(df_region_pops_2019,
    
    
    03 by = "region")
    | >
    
    04 ggplot() +
    
    
    05 aes(f
    i
    ll = population) +
    
    
    06 geom_sf(color = "transparent")
    geometry
    

    View Slide

79. 地図 使 表現
    79
    IUUQTXXXTUPQDPWJEKQ
    空間的 位置関係 ⼤ 保持 形式 表現
    ΧϥϜ஍ਤ
    ※⾒ ⼈ 対 、空間的 配置 対 理解 暗黙的 求
    {statebins}
    

    View Slide

80. 01 library(tabularmaps)
    
    
    02 jpn77
    | >
    
    03 select(jis_code,
    
    
    04 prefecture = prefecture_kanji,
    
    
    05 x,
    
    
    06 y)
    | >
    
    07 left_join(df_ssdse_b2019,
    
    
    08 by = "prefecture")
    | >
    
    09 tabularmap(x,
    
    
    10 y,
    
    
    11 group = jis_code,
    
    
    12 label = prefecture,
    
    
    13 f
    i
    ll = population,
    
    
    14 size = 3) +
    
    
    15 scale_f
    i
    ll_viridis_c() +
    
    
    16 theme_tabularmap()
    {tabularmaps}
    80
    

    View Slide

81. {geofacet}
    81
    01 library(geofacet)
    
    
    02 jp_prefs_grid1
    < -
    
    03 jp_prefs_grid1
    | >
    
    04 left_join(
    
    
    05 zipangu
    : :
    jpnprefs
    | >
    
    06 select(jis_code, prefecture = prefecture_kanji),
    
    
    07 by = c("code_pref_jis" = "jis_code"))
    01 p
    < -
    
    02 df_ssdse_b
    | >
    
    03 ggplot() +
    
    
    04 aes(year, population) +
    
    
    05 geom_line() +
    
    
    06 theme_gray(base_size = 6)
    
    
    07
    
    
    08 p +
    
    
    09 facet_geo(~ prefecture,
    
    
    10 grid = "jp_prefs_grid1",
    
    
    11 scales = "free_y")
    

    View Slide

82. 1
    . R {sf}
    
    
    2
    . {sf} ggplot
    2
    ::geom_sf()
    
    
    3
    .
    
    
    4
    . 総
    

    View Slide

83. View Slide

84. 利⽤
    84
    IUUQTXXXSTUVEJPDPNSFTPVSDFTDIFBUTIFFUT ggplot
    2
    dplyr
    

    View Slide

85. 多 図・ ⾒
    TidyTuesday
    85
    #
    30
    DayMapChallenge
    > A weekly social data project in R.
    IUUQTHJUIVCDPNSGPSEBUBTDJFODFUJEZUVFTEBZ
    from Data to Vis
    Twitter #tidytuesday
    R Python D
    3
    .js
    THE R GRAPH GALLERY IUUQTSHSBQIHBMMFSZDPN
    IUUQTXXXEBUBUPWJ[DPN
    awesome-ggplot2 IUUQTHJUIVCDPNFSJLHBIOFSBXFTPNFHHQMPU
    ggplot
    2
    ggplot2 extensions IUUQTFYUTHHQMPUUJEZWFSTFPSH
    ggplot
    2
    #rtistry
    

    View Slide

86. 最後 〜今回 発表 伝 〜
    1
    .
    
    
    2
    .
    
    
    3.
    

    View Slide

87. 参考⽂献・URL
    87
    Kieran Healy, 2019. Data Visualization: A practical introduction. Princeton University Press (⽠⽣真也、江⼝哲史、三村喬⽣
    訳, 2021. 分析 可視化⼊⾨. 講談社 )
    Claus Wilke, 2019. Fundamentals of Data Visualization: A Primer on Making Informative and Compelling Figures. OʼReilly
    (⼩林儀匡、瀬⼾⼭雅⼈訳, 2022. 基礎: 明確 、魅⼒的 、説得⼒ ⾒
    ⽅・伝 ⽅. ・ )
    Michael Friendly and Howard Wainer, 2021. A History of Data Visualization and Graphic Communication. Harvard
    University Press (飯嶋貴⼦訳, 2021. 視覚化 ⼈類史: 発明 時間 空間 可視化 . ⻘⼟社)
    Jonathan Schwabish, 2021. Better Data Visualizations: A Guide for Scholars, Researchers, and Wonks Fundamentals of
    Data Visualization. Columbia University Press
    Hadley Wickham, Danielle Navarro, and Thomas Lin Pedersen, (2022). ggplot2: elegant graphics for data analysis. Springer
    Cédric Scherer 2022. Graphic Design with ggplot2: How to Create Engaging and Complex Visualizations in R
    IUUQTDMBVTXJMLFDPNEBUBWJ[
    IUUQTGSJFOEMZHJUIVCJP)JTU%BUB7JT
    IUUQTHHQMPUCPPLPSH
    IUUQTTPDWJ[DP
    IUUQTSTUVEJPDPOGHJUIVCJPHHQMPUHSBQIJDEFTJHO
    

    View Slide

88. ENJOY!
    

    View Slide