joshdavham commited on
Commit
c21347b
·
1 Parent(s): 360a122

reorganize grammar and word origin tables

Browse files
Files changed (1) hide show
  1. app.py +86 -75
app.py CHANGED
@@ -9,6 +9,17 @@ st.set_page_config(
9
  page_icon='favicon.svg'
10
  )
11
 
 
 
 
 
 
 
 
 
 
 
 
12
  @st.cache_data
13
  def load_dataframes():
14
 
@@ -18,9 +29,81 @@ def load_dataframes():
18
 
19
  return video_df, word_coverage_df, num_video_df
20
 
21
- video_df, word_coverage_df, num_video_df = load_dataframes()
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
22
 
23
 
 
 
 
 
24
  st.markdown("Note: this analysis is meant to viewed on a computer and not a phone (sorry!)")
25
 
26
  st.markdown("[Code can be found [here](https://github.com/joshdavham/cij-analysis)]")
@@ -1315,49 +1398,8 @@ st.altair_chart(sconj_hist, use_container_width=True)
1315
 
1316
  st.markdown("We also notice differences in the use of other types of words.")
1317
 
1318
- data = {
1319
- 'Complete Beginner': [0.02638719922016275 ,0.0192492959834, 0.00476028625918155, 0.2503071253071253],
1320
- 'Beginner': [0.0473047304730473, 0.0266429840142095, 0.005813953488372, 0.2454068241469816],
1321
- 'Intermediate': [0.06625719079578135, 0.03514773095199635, 0.0087719298245614, 0.23239271705403663],
1322
- 'Advanced': [0.0766787658802177, 0.0373056994818652, 0.0108588351431391, 0.2237101220953131]
1323
- }
1324
- df = pd.DataFrame(data)
1325
-
1326
- row_labels = ['Median Perc. Subordinating Conjunctions', 'Median Perc. Adverbs', 'Median Perc. Determiners', 'Median Perc. Nouns']
1327
- df.index = row_labels
1328
-
1329
- styled_df = df.style.set_table_styles(
1330
- {
1331
- 'Complete Beginner': [
1332
- {'selector': 'th.col_heading.level0', 'props': [('background-color', 'rgba(165, 190, 228, 0.45)')]},
1333
- {'selector': 'td:hover', 'props': [('background-color', '#e0f7fa')]}
1334
- ],
1335
- 'Beginner': [
1336
- {'selector': 'th.col_heading.level0', 'props': [('background-color', 'rgba(154, 214, 216, 0.45)')]},
1337
- {'selector': 'td:hover', 'props': [('background-color', '#e0f7fa')]}
1338
- ],
1339
- 'Intermediate': [
1340
- {'selector': 'th.col_heading.level0', 'props': [('background-color', 'rgba(199, 174, 205, 0.45)')]},
1341
- {'selector': 'td:hover', 'props': [('background-color', '#e0f7fa')]}
1342
- ],
1343
- 'Advanced': [
1344
- {'selector': 'th.col_heading.level0', 'props': [('background-color', 'rgba(221, 158, 158, 0.45)')]},
1345
- {'selector': 'td:hover', 'props': [('background-color', '#e0f7fa')]}
1346
- ]
1347
- }).set_properties(**{'background-color': 'white'}).format("{:.2%}")
1348
-
1349
- st.markdown(
1350
- """
1351
- <style>
1352
- .dataframe-divv {
1353
- background-color: white;
1354
- }
1355
- </style>
1356
- """, unsafe_allow_html=True
1357
- )
1358
-
1359
  st.markdown(
1360
- '<div class="dataframe-divv">' + styled_df.to_html() + "</div>"
1361
  , unsafe_allow_html=True)
1362
 
1363
  ###
@@ -1527,39 +1569,8 @@ st.markdown("In Japanese, Kango are somewhat analogous to French words in Englis
1527
 
1528
  st.markdown("We also notice orderings when counting the percentage of Wago and Gairaigo as well.")
1529
 
1530
- data = {
1531
- 'Complete Beginner': [0.06999874574159035, 0.8578043261266064, 0.03301790801790795],
1532
- 'Beginner': [0.0955284552845528, 0.8399311531841652, 0.0279441117764471],
1533
- 'Intermediate': [0.1165702954621605, 0.8259877335615461, 0.0241447813837379],
1534
- 'Advanced': [0.1303328645100797, 0.8225274725274725, 0.0157535445475231],
1535
- }
1536
- df = pd.DataFrame(data)
1537
-
1538
- row_labels = ['Median Perc. Kango (漢語)', 'Median Perc. Wago (和語)', 'Median Perc. Garaigo (外来語)']
1539
- df.index = row_labels
1540
-
1541
- styled_df = df.style.set_table_styles(
1542
- {
1543
- 'Complete Beginner': [
1544
- {'selector': 'th.col_heading.level0', 'props': [('background-color', 'rgba(165, 190, 228, 0.45)')]},
1545
- {'selector': 'td:hover', 'props': [('background-color', '#e0f7fa')]}
1546
- ],
1547
- 'Beginner': [
1548
- {'selector': 'th.col_heading.level0', 'props': [('background-color', 'rgba(154, 214, 216, 0.45)')]},
1549
- {'selector': 'td:hover', 'props': [('background-color', '#e0f7fa')]}
1550
- ],
1551
- 'Intermediate': [
1552
- {'selector': 'th.col_heading.level0', 'props': [('background-color', 'rgba(199, 174, 205, 0.45)')]},
1553
- {'selector': 'td:hover', 'props': [('background-color', '#e0f7fa')]}
1554
- ],
1555
- 'Advanced': [
1556
- {'selector': 'th.col_heading.level0', 'props': [('background-color', 'rgba(221, 158, 158, 0.45)')]},
1557
- {'selector': 'td:hover', 'props': [('background-color', '#e0f7fa')]}
1558
- ],
1559
- }).set_properties(**{'background-color': 'white'}).format("{:.2%}")
1560
-
1561
  st.markdown(
1562
- '<div class="dataframe-divv">' + styled_df.to_html() + "</div>"
1563
  , unsafe_allow_html=True)
1564
 
1565
  ###
 
9
  page_icon='favicon.svg'
10
  )
11
 
12
+ # colors white the index columns of rendered dataframes
13
+ st.markdown(
14
+ """
15
+ <style>
16
+ .dataframe-div {
17
+ background-color: white;
18
+ }
19
+ </style>
20
+ """, unsafe_allow_html=True
21
+ )
22
+
23
  @st.cache_data
24
  def load_dataframes():
25
 
 
29
 
30
  return video_df, word_coverage_df, num_video_df
31
 
32
+ def get_grammar_table():
33
+
34
+ data = {
35
+ 'Complete Beginner': [0.02638719922016275 ,0.0192492959834, 0.00476028625918155, 0.2503071253071253],
36
+ 'Beginner': [0.0473047304730473, 0.0266429840142095, 0.005813953488372, 0.2454068241469816],
37
+ 'Intermediate': [0.06625719079578135, 0.03514773095199635, 0.0087719298245614, 0.23239271705403663],
38
+ 'Advanced': [0.0766787658802177, 0.0373056994818652, 0.0108588351431391, 0.2237101220953131]
39
+ }
40
+ df = pd.DataFrame(data)
41
+
42
+ row_labels = ['Median Perc. Subordinating Conjunctions', 'Median Perc. Adverbs', 'Median Perc. Determiners', 'Median Perc. Nouns']
43
+ df.index = row_labels
44
+
45
+ styled_df = df.style.set_table_styles(
46
+ {
47
+ 'Complete Beginner': [
48
+ {'selector': 'th.col_heading.level0', 'props': [('background-color', 'rgba(165, 190, 228, 0.45)')]},
49
+ {'selector': 'td:hover', 'props': [('background-color', '#e0f7fa')]}
50
+ ],
51
+ 'Beginner': [
52
+ {'selector': 'th.col_heading.level0', 'props': [('background-color', 'rgba(154, 214, 216, 0.45)')]},
53
+ {'selector': 'td:hover', 'props': [('background-color', '#e0f7fa')]}
54
+ ],
55
+ 'Intermediate': [
56
+ {'selector': 'th.col_heading.level0', 'props': [('background-color', 'rgba(199, 174, 205, 0.45)')]},
57
+ {'selector': 'td:hover', 'props': [('background-color', '#e0f7fa')]}
58
+ ],
59
+ 'Advanced': [
60
+ {'selector': 'th.col_heading.level0', 'props': [('background-color', 'rgba(221, 158, 158, 0.45)')]},
61
+ {'selector': 'td:hover', 'props': [('background-color', '#e0f7fa')]}
62
+ ]
63
+ }).set_properties(**{'background-color': 'white'}).format("{:.2%}")
64
+
65
+ return styled_df
66
+
67
+ def get_word_origin_table():
68
+
69
+ data = {
70
+ 'Complete Beginner': [0.06999874574159035, 0.8578043261266064, 0.03301790801790795],
71
+ 'Beginner': [0.0955284552845528, 0.8399311531841652, 0.0279441117764471],
72
+ 'Intermediate': [0.1165702954621605, 0.8259877335615461, 0.0241447813837379],
73
+ 'Advanced': [0.1303328645100797, 0.8225274725274725, 0.0157535445475231],
74
+ }
75
+ df = pd.DataFrame(data)
76
+
77
+ row_labels = ['Median Perc. Kango (漢語)', 'Median Perc. Wago (和語)', 'Median Perc. Garaigo (外来語)']
78
+ df.index = row_labels
79
+
80
+ styled_df = df.style.set_table_styles(
81
+ {
82
+ 'Complete Beginner': [
83
+ {'selector': 'th.col_heading.level0', 'props': [('background-color', 'rgba(165, 190, 228, 0.45)')]},
84
+ {'selector': 'td:hover', 'props': [('background-color', '#e0f7fa')]}
85
+ ],
86
+ 'Beginner': [
87
+ {'selector': 'th.col_heading.level0', 'props': [('background-color', 'rgba(154, 214, 216, 0.45)')]},
88
+ {'selector': 'td:hover', 'props': [('background-color', '#e0f7fa')]}
89
+ ],
90
+ 'Intermediate': [
91
+ {'selector': 'th.col_heading.level0', 'props': [('background-color', 'rgba(199, 174, 205, 0.45)')]},
92
+ {'selector': 'td:hover', 'props': [('background-color', '#e0f7fa')]}
93
+ ],
94
+ 'Advanced': [
95
+ {'selector': 'th.col_heading.level0', 'props': [('background-color', 'rgba(221, 158, 158, 0.45)')]},
96
+ {'selector': 'td:hover', 'props': [('background-color', '#e0f7fa')]}
97
+ ],
98
+ }).set_properties(**{'background-color': 'white'}).format("{:.2%}")
99
+
100
+ return styled_df
101
 
102
 
103
+ video_df, word_coverage_df, num_video_df = load_dataframes()
104
+ grammar_table = get_grammar_table()
105
+ word_origin_table = get_word_origin_table()
106
+
107
  st.markdown("Note: this analysis is meant to viewed on a computer and not a phone (sorry!)")
108
 
109
  st.markdown("[Code can be found [here](https://github.com/joshdavham/cij-analysis)]")
 
1398
 
1399
  st.markdown("We also notice differences in the use of other types of words.")
1400
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1401
  st.markdown(
1402
+ '<div class="dataframe-div">' + grammar_table.to_html() + "</div>"
1403
  , unsafe_allow_html=True)
1404
 
1405
  ###
 
1569
 
1570
  st.markdown("We also notice orderings when counting the percentage of Wago and Gairaigo as well.")
1571
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1572
  st.markdown(
1573
+ '<div class="dataframe-div">' + word_origin_table.to_html() + "</div>"
1574
  , unsafe_allow_html=True)
1575
 
1576
  ###