Spaces:
Sleeping
Sleeping
switching to plotly graphs
Browse files- .gitignore +2 -0
- app.py +106 -83
.gitignore
ADDED
@@ -0,0 +1,2 @@
|
|
|
|
|
|
|
1 |
+
# ignore pycache
|
2 |
+
__pycache__/
|
app.py
CHANGED
@@ -4,6 +4,8 @@ from nltk.util import ngrams
|
|
4 |
from collections import Counter
|
5 |
import pandas as pd
|
6 |
import plotly.express as px
|
|
|
|
|
7 |
import matplotlib.pyplot as plt
|
8 |
|
9 |
# Load the dataset and convert it to a Pandas dataframe
|
@@ -28,7 +30,7 @@ df["ari"] = df["no-contractions"].apply(
|
|
28 |
+ (0.5 * (len(x.split()) / len(x.split("."))))
|
29 |
- 21.43
|
30 |
)
|
31 |
-
|
32 |
written = df[df["categories"] == "Written"]
|
33 |
spoken = df[df["categories"] == "Spoken"]
|
34 |
|
@@ -39,115 +41,136 @@ with gr.Blocks() as demo:
|
|
39 |
# A Dashboard to Analyze the State of the Union Addresses
|
40 |
"""
|
41 |
)
|
42 |
-
|
43 |
df,
|
44 |
x="date",
|
45 |
y="word_count",
|
46 |
-
title="Total Number of Words in
|
47 |
-
|
|
|
|
|
|
|
|
|
48 |
)
|
|
|
49 |
# group by president and category and calculate the average word count sort by date
|
50 |
avg_word_count = (
|
51 |
-
df.groupby(["
|
52 |
)
|
53 |
-
|
54 |
-
gr.BarPlot(
|
55 |
avg_word_count,
|
56 |
x="potus",
|
57 |
y="word_count",
|
58 |
-
title="Average Number of Words in
|
59 |
color="categories",
|
60 |
-
|
61 |
-
height=400,
|
62 |
-
min_width=160,
|
63 |
-
fill_height=True,
|
64 |
-
container=True,
|
65 |
-
scale=2,
|
66 |
)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
67 |
with gr.Row():
|
68 |
ari = df[["potus", "date", "ari", "categories"]]
|
69 |
-
|
70 |
ari,
|
71 |
x="date",
|
72 |
y="ari",
|
73 |
-
title="Automated Readability Index",
|
|
|
|
|
|
|
|
|
|
|
74 |
)
|
|
|
75 |
# get all unique president names
|
76 |
presidents = df["potus"].unique()
|
77 |
# convert presidents to a list
|
78 |
presidents = presidents.tolist()
|
79 |
# create a dropdown to select a president
|
80 |
-
president = gr.Dropdown(label="Select a President", choices=
|
81 |
grams = gr.Slider(minimum=1, maximum=4, step=1, label="N-grams", interactive=True)
|
82 |
-
with gr.Row():
|
83 |
-
# if president is not of type string
|
84 |
-
@gr.render(inputs=president)
|
85 |
-
def show_text(potus):
|
86 |
-
if potus != "All" and potus is not None:
|
87 |
-
ari = df[df["potus"] == potus][
|
88 |
-
["date", "categories", "word_count", "ari"]
|
89 |
-
]
|
90 |
-
gr.DataFrame(ari, height=200)
|
91 |
|
92 |
-
|
93 |
-
|
94 |
-
#
|
95 |
-
|
96 |
-
#
|
97 |
-
|
98 |
-
|
99 |
-
|
100 |
-
|
101 |
-
|
102 |
-
|
103 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
104 |
|
105 |
-
|
106 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
107 |
|
108 |
-
|
109 |
-
|
110 |
-
|
111 |
-
|
112 |
-
|
113 |
-
print(n_grams)
|
114 |
-
# create a Counter object from the trigrams
|
115 |
-
potus_df = df[df["potus"] == potus]
|
116 |
-
# decode the tokens-nostop column from a byte array to a list of string
|
117 |
-
trigrams = (
|
118 |
-
potus_df["tokens-nostop"]
|
119 |
-
.apply(lambda x: list(ngrams(x, n_grams)))
|
120 |
-
.apply(Counter)
|
121 |
-
.sum()
|
122 |
-
)
|
123 |
-
# get the most common trigrams
|
124 |
-
common_trigrams = trigrams.most_common(20)
|
125 |
-
# unzip the list of tuples and plot the trigrams and counts as a bar chart
|
126 |
-
trigrams, counts = zip(*common_trigrams)
|
127 |
-
# join the trigrams into a single string
|
128 |
-
trigrams = [" ".join(trigram) for trigram in trigrams]
|
129 |
-
# create a dataframe from the trigrams and counts
|
130 |
-
trigrams_df = pd.DataFrame({"trigrams": trigrams, "counts": counts})
|
131 |
-
# plot the trigrams and counts as a bar chart from matplotlib
|
132 |
-
"""
|
133 |
-
fig, ax = plt.subplots(figsize=(12, 4))
|
134 |
-
ax.barh(trigrams_df["trigrams"], trigrams_df["counts"])
|
135 |
-
ax.set_title("Top 20 Trigrams")
|
136 |
-
ax.set_ylabel("Count")
|
137 |
-
ax.set_xlabel("Trigrams")
|
138 |
-
plt.xticks(rotation=45)
|
139 |
-
# make it tight layout
|
140 |
-
plt.tight_layout()
|
141 |
-
"""
|
142 |
-
fig = px.scatter(
|
143 |
-
trigrams_df,
|
144 |
-
x="counts",
|
145 |
-
y="trigrams",
|
146 |
-
title="Top 20 Trigrams",
|
147 |
-
orientation="h",
|
148 |
-
)
|
149 |
-
print(fig)
|
150 |
-
gr.Plot(value=fig, container=True, visible=True)
|
151 |
|
152 |
|
153 |
demo.launch(share=True)
|
|
|
4 |
from collections import Counter
|
5 |
import pandas as pd
|
6 |
import plotly.express as px
|
7 |
+
import plotly.graph_objects as go
|
8 |
+
from plotly.subplots import make_subplots
|
9 |
import matplotlib.pyplot as plt
|
10 |
|
11 |
# Load the dataset and convert it to a Pandas dataframe
|
|
|
30 |
+ (0.5 * (len(x.split()) / len(x.split("."))))
|
31 |
- 21.43
|
32 |
)
|
33 |
+
df = df.sort_values(by="date")
|
34 |
written = df[df["categories"] == "Written"]
|
35 |
spoken = df[df["categories"] == "Spoken"]
|
36 |
|
|
|
41 |
# A Dashboard to Analyze the State of the Union Addresses
|
42 |
"""
|
43 |
)
|
44 |
+
fig1 = px.line(
|
45 |
df,
|
46 |
x="date",
|
47 |
y="word_count",
|
48 |
+
title="Total Number of Words in Addresses",
|
49 |
+
line_shape="spline",
|
50 |
+
)
|
51 |
+
fig1.update_layout(
|
52 |
+
xaxis=dict(title="Date of Address"),
|
53 |
+
yaxis=dict(title="Word Count"),
|
54 |
)
|
55 |
+
gr.Plot(fig1)
|
56 |
# group by president and category and calculate the average word count sort by date
|
57 |
avg_word_count = (
|
58 |
+
df.groupby(["potus", "categories"])["word_count"].mean().reset_index()
|
59 |
)
|
60 |
+
fig2 = px.bar(
|
|
|
61 |
avg_word_count,
|
62 |
x="potus",
|
63 |
y="word_count",
|
64 |
+
title="Average Number of Words in Addresses by President",
|
65 |
color="categories",
|
66 |
+
barmode="group",
|
|
|
|
|
|
|
|
|
|
|
67 |
)
|
68 |
+
fig2.update_layout(
|
69 |
+
xaxis=dict(
|
70 |
+
title="President",
|
71 |
+
tickangle=-45, # Rotate labels 45 degrees counterclockwise
|
72 |
+
),
|
73 |
+
yaxis=dict(
|
74 |
+
title="Average Word Count",
|
75 |
+
tickangle=0, # Default label angle (horizontal)
|
76 |
+
),
|
77 |
+
legend=dict(orientation="h", yanchor="bottom", y=1.02, xanchor="right", x=1),
|
78 |
+
)
|
79 |
+
gr.Plot(fig2)
|
80 |
with gr.Row():
|
81 |
ari = df[["potus", "date", "ari", "categories"]]
|
82 |
+
fig3 = px.line(
|
83 |
ari,
|
84 |
x="date",
|
85 |
y="ari",
|
86 |
+
title="Automated Readability Index in each Address",
|
87 |
+
line_shape="spline",
|
88 |
+
)
|
89 |
+
fig3.update_layout(
|
90 |
+
xaxis=dict(title="Date of Address"),
|
91 |
+
yaxis=dict(title="ARI Score"),
|
92 |
)
|
93 |
+
gr.Plot(fig3)
|
94 |
# get all unique president names
|
95 |
presidents = df["potus"].unique()
|
96 |
# convert presidents to a list
|
97 |
presidents = presidents.tolist()
|
98 |
# create a dropdown to select a president
|
99 |
+
president = gr.Dropdown(label="Select a President", choices=presidents)
|
100 |
grams = gr.Slider(minimum=1, maximum=4, step=1, label="N-grams", interactive=True)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
101 |
|
102 |
+
def plotly_bar(n_grams, potus):
|
103 |
+
if potus is not None:
|
104 |
+
# create a Counter object from the trigrams
|
105 |
+
potus_df = df[df["potus"] == potus]
|
106 |
+
# decode the tokens-nostop column from a byte array to a list of string
|
107 |
+
trigrams = (
|
108 |
+
potus_df["tokens-nostop"]
|
109 |
+
.apply(lambda x: list(ngrams(x, n_grams)))
|
110 |
+
.apply(Counter)
|
111 |
+
.sum()
|
112 |
+
)
|
113 |
+
# get the most common trigrams
|
114 |
+
common_trigrams = trigrams.most_common(10)
|
115 |
+
# unzip the list of tuples and plot the trigrams and counts as a bar chart
|
116 |
+
trigrams, counts = zip(*common_trigrams)
|
117 |
+
# join the trigrams into a single string
|
118 |
+
trigrams = [" ".join(trigram) for trigram in trigrams]
|
119 |
+
# create a dataframe from the trigrams and counts
|
120 |
+
trigrams_df = pd.DataFrame({"trigrams": trigrams, "counts": counts})
|
121 |
+
fig4 = px.bar(
|
122 |
+
trigrams_df,
|
123 |
+
x="counts",
|
124 |
+
y="trigrams",
|
125 |
+
title=f"Top {n_grams}-grams",
|
126 |
+
orientation="h",
|
127 |
+
height=400,
|
128 |
+
)
|
129 |
+
return fig4
|
130 |
|
131 |
+
if president != "All" and president is not None:
|
132 |
+
gr.Plot(plotly_bar, inputs=[grams, president])
|
133 |
+
|
134 |
+
def plotly_line(president):
|
135 |
+
if president != "All" and president is not None:
|
136 |
+
potus_df = df[df["potus"] == president]
|
137 |
+
fig5 = make_subplots(specs=[[{"secondary_y": True}]])
|
138 |
+
fig5.add_trace(
|
139 |
+
go.Scatter(
|
140 |
+
x=potus_df["date"],
|
141 |
+
y=potus_df["word_count"],
|
142 |
+
name="Word Count",
|
143 |
+
),
|
144 |
+
secondary_y=False,
|
145 |
+
)
|
146 |
+
fig5.add_trace(
|
147 |
+
go.Scatter(
|
148 |
+
x=potus_df["date"],
|
149 |
+
y=potus_df["ari"],
|
150 |
+
name="ARI",
|
151 |
+
),
|
152 |
+
secondary_y=True,
|
153 |
+
)
|
154 |
+
# Add figure title
|
155 |
+
fig5.update_layout(title_text="Double Y Axis Example")
|
156 |
+
|
157 |
+
# Set x-axis title
|
158 |
+
fig5.update_xaxes(title_text="xaxis title")
|
159 |
+
|
160 |
+
# Set y-axes titles
|
161 |
+
fig5.update_yaxes(
|
162 |
+
title_text="<b>primary</b> yaxis title", secondary_y=False
|
163 |
+
)
|
164 |
+
fig5.update_yaxes(
|
165 |
+
title_text="<b>secondary</b> yaxis title", secondary_y=True
|
166 |
+
)
|
167 |
+
return fig5
|
168 |
|
169 |
+
# calculate the total number of words in the speech_html column and add it to a new column
|
170 |
+
# if the president is "All", show the word count for all presidents
|
171 |
+
# if the president is not "All", show the word count for the selected president
|
172 |
+
if president != "All" and president is not None:
|
173 |
+
gr.Plot(plotly_line, inputs=[president])
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
174 |
|
175 |
|
176 |
demo.launch(share=True)
|