add topics
Browse files
app.py
CHANGED
|
@@ -63,15 +63,11 @@ stopword = stopword + context # add some frequent words in the documents
|
|
| 63 |
clean_text = df_clean['description'].apply(lambda x: clean_up1(x, stopword, pos))
|
| 64 |
docs = clean_text.apply(lambda x: " ".join(x)).tolist()
|
| 65 |
|
| 66 |
-
topic_model = BERTopic(
|
| 67 |
-
nr_topics=6,
|
| 68 |
-
top_n_words=30,
|
| 69 |
-
low_memory=True,
|
| 70 |
-
n_gram_range=(1, 2))
|
| 71 |
|
| 72 |
-
topics,
|
|
|
|
| 73 |
|
| 74 |
-
topic_fig = topic_model.visualize_barchart(n_words=10)
|
| 75 |
|
| 76 |
timestamps = df_clean.index
|
| 77 |
topics_over_time = topic_model.topics_over_time(docs, topics, timestamps,
|
|
@@ -86,6 +82,8 @@ topics_over_time.set_index('Timestamp', inplace=True)
|
|
| 86 |
topics_over_time['year'] = topics_over_time.index.year
|
| 87 |
topic_per_year = topics_over_time.groupby(['year'])['Words'].apply(lambda x: x.str.cat(sep=' '))
|
| 88 |
|
|
|
|
|
|
|
| 89 |
fig1, ax = plt.subplots()
|
| 90 |
sns.countplot(ax=ax, x='year', data=df_clean, palette='viridis');
|
| 91 |
|
|
|
|
| 63 |
clean_text = df_clean['description'].apply(lambda x: clean_up1(x, stopword, pos))
|
| 64 |
docs = clean_text.apply(lambda x: " ".join(x)).tolist()
|
| 65 |
|
| 66 |
+
topic_model = BERTopic.load("bertopic")
|
|
|
|
|
|
|
|
|
|
|
|
|
| 67 |
|
| 68 |
+
with open('topics', 'rb') as f:
|
| 69 |
+
topics = pickle.load(f)
|
| 70 |
|
|
|
|
| 71 |
|
| 72 |
timestamps = df_clean.index
|
| 73 |
topics_over_time = topic_model.topics_over_time(docs, topics, timestamps,
|
|
|
|
| 82 |
topics_over_time['year'] = topics_over_time.index.year
|
| 83 |
topic_per_year = topics_over_time.groupby(['year'])['Words'].apply(lambda x: x.str.cat(sep=' '))
|
| 84 |
|
| 85 |
+
topic_fig = topic_model.visualize_barchart(n_words=10)
|
| 86 |
+
|
| 87 |
fig1, ax = plt.subplots()
|
| 88 |
sns.countplot(ax=ax, x='year', data=df_clean, palette='viridis');
|
| 89 |
|
topics
ADDED
|
Binary file (1.17 kB). View file
|
|
|