In [1]:
%reload_ext autotime
import pandas as pd
from glob import glob
import json
import os
from bertopic import BERTopic
from bertopic.representation import KeyBERTInspired
import plotly.io as pio
pio.renderers.default = 'notebook'
pd.options.plotting.backend = "plotly"
df = pd.read_csv("country_calendar_metadata.csv")
df["Recording Date"] = df["Recording Date"].replace("0000-00-00", "2008-10-23")
def get_transcript(file):
file = os.path.basename(file.replace("?download=1", ""))
transcript_filename = f"transcripts/{file}.json"
with open(transcript_filename) as f:
return json.load(f)["text"]
df["transcript"] = df.link.apply(get_transcript)
#df.index = pd.to_datetime(df["Recording Date"])
df.to_csv("country_calendar_metadata_with_transcripts.csv")
df
✔️ 10.6 s (2024-05-21T14:37:21/2024-05-21T14:37:31)
2024-05-21 14:37:26.429595: I tensorflow/core/util/port.cc:113] oneDNN custom operations are on. You may see slightly different numerical results due to floating-point round-off errors from different computation orders. To turn them off, set the environment variable `TF_ENABLE_ONEDNN_OPTS=0`. 2024-05-21 14:37:26.487729: I tensorflow/core/platform/cpu_feature_guard.cc:210] This TensorFlow binary is optimized to use available CPU instructions in performance-critical operations. To enable the following instructions: AVX2 AVX512F AVX512_VNNI FMA, in other operations, rebuild TensorFlow with the appropriate compiler flags. 2024-05-21 14:37:27.329129: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Could not find TensorRT
Out[1]:
| File Name | Name | Description | Duration | Type | Subject | Age Level | Channel | Recording Date | Recording Time | id | link | size_bytes | transcript | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 0 | etv/country_calendar_-_air_new_zealand_uniform... | country calendar - air new zealand uniforms 23... | from www.throng.co.nz : this episode of nation... | 00:21:33 | Library | NaN | G | lostplusfound | 2008-10-23 | 00:00:00 | 61580 | https://cdn.etv.org.nz/etv/country_calendar_-_... | 145371146 | Music Music Music Music Music Music Music Mus... |
| 1 | etv/TV_One_1900_21022009_20090221193153256qtst... | Country Calendar - Series 2009 - Episode 1 | 'Lyn of Taranaki' Country Calendar starts its ... | 00:29:57 | Documentary | Agriculture & Horticulture, Business, Ecology ... | Unrated | TV 1 | 2009-02-21 | 19:00:00 | 12564 | https://cdn.etv.org.nz/etv/TV_One_1900_2102200... | 49973516 | Lin Webster's life has taken many unusual twi... |
| 2 | etv/TV_One_1900_28022009_20090228193154256qtst... | Country Calendar - Series 2009 - Episode 2 | 'Shared Fruit' New Zealand and Chinese scienti... | 00:29:56 | Documentary | Agriculture & Horticulture, Asian Studies, Bus... | Unrated | TV 1 | 2009-02-28 | 19:00:00 | 12697 | https://cdn.etv.org.nz/etv/TV_One_1900_2802200... | 52854708 | Yunnan, tucked away in the remote southwest o... |
| 3 | etv/TV_One_1900_07032009_20090307193229256qtst... | Country Calendar - Series 2009 - Episode 3 | 'Grassroots Leader' Bruce Rollinson is a late-... | 00:29:56 | Documentary | Agriculture & Horticulture, Business, Economic... | Unrated | TV 1 | 2009-03-07 | 19:00:00 | 12812 | https://cdn.etv.org.nz/etv/TV_One_1900_0703200... | 51865839 | The season on Mount Ropiku was the best for y... |
| 4 | etv/TV_One_1900_14032009_20090314193119768qtst... | Country Calendar - Series 2009 - Episode 4 | 'High-flying Fur' What's as fine as merino woo... | 00:29:56 | Documentary | Agriculture & Horticulture, Biology, Business,... | Unrated | TV 1 | 2009-03-14 | 19:00:00 | 12939 | https://cdn.etv.org.nz/etv/TV_One_1900_1403200... | 154209978 | With patches of bush and some good keen hunte... |
| ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... |
| 511 | etv/hyundai_country_cale_tv1_20240414_1900 | Hyundai Country Calendar - Season 2024 - 14 Ap... | NaN | 00:30:33 | Documentary, Reality | Agriculture & Horticulture, Business, Careers,... | G | TV 1 | 2024-04-14 | 19:00:00 | 518923 | https://cdn.etv.org.nz/etv/hyundai_country_cal... | 353531481 | Always a favourite on every rural road. Hyund... |
| 512 | etv/hyundai_country_cale_tv1_20240421_1900 | Hyundai Country Calendar - Season 2024 - 21 Ap... | NaN | 00:30:21 | Documentary, Reality | Agriculture & Horticulture, Business, Careers,... | G | TV 1 | 2024-04-21 | 19:00:00 | 518922 | https://cdn.etv.org.nz/etv/hyundai_country_cal... | 344594242 | Always a favourite on every rural road. Hyund... |
| 513 | etv/hyundai_country_cale_tv1_20240428_1900 | Hyundai Country Calendar - Season 2024 - 28 Ap... | NaN | 00:30:10 | Documentary, Reality | Agriculture & Horticulture, Business, Careers,... | G | TV 1 | 2024-04-28 | 19:00:00 | 518972 | https://cdn.etv.org.nz/etv/hyundai_country_cal... | 338647527 | Always a favourite on every rural road. Hyund... |
| 514 | etv/hyundai_country_cale_tv1_20240505_1900 | Hyundai Country Calendar - Season 2024 - 5 May... | NaN | 00:30:45 | Documentary, Reality | Agriculture & Horticulture, Business, Careers,... | G | TV 1 | 2024-05-05 | 19:00:00 | 519110 | https://cdn.etv.org.nz/etv/hyundai_country_cal... | 352367660 | Always a favourite on every rural road. Hyund... |
| 515 | etv/hyundai_country_cale_tv1_20240512_1900 | Hyundai Country Calendar - Season 2024 - 12 Ma... | NaN | 00:30:53 | Documentary, Reality | Agriculture & Horticulture, Business, Careers,... | G | TV 1 | 2024-05-12 | 19:00:00 | 519287 | https://cdn.etv.org.nz/etv/hyundai_country_cal... | 348651686 | Always a favourite on every rural road. Hyund... |
516 rows × 14 columns
In [2]:
df["Year"] = df["Recording Date"].str[:4]
✔️ 1.8 ms (2024-05-21T14:37:31/2024-05-21T14:37:31)
In [3]:
df["year_in_name"] = df.Name.str.extract("(\d{4})")
✔️ 2.76 ms (2024-05-21T14:37:31/2024-05-21T14:37:31)
In [4]:
mismatches = df[(df.Year != df.year_in_name) & ~df.year_in_name.isna()]
df.loc[mismatches.index, "Recording Date"] = df.year_in_name + "-01-01"
df.index = pd.to_datetime(df["Recording Date"])
df
✔️ 20.9 ms (2024-05-21T14:37:32/2024-05-21T14:37:32)
Out[4]:
| File Name | Name | Description | Duration | Type | Subject | Age Level | Channel | Recording Date | Recording Time | id | link | size_bytes | transcript | Year | year_in_name | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Recording Date | ||||||||||||||||
| 2008-10-23 | etv/country_calendar_-_air_new_zealand_uniform... | country calendar - air new zealand uniforms 23... | from www.throng.co.nz : this episode of nation... | 00:21:33 | Library | NaN | G | lostplusfound | 2008-10-23 | 00:00:00 | 61580 | https://cdn.etv.org.nz/etv/country_calendar_-_... | 145371146 | Music Music Music Music Music Music Music Mus... | 2008 | 2008 |
| 2009-02-21 | etv/TV_One_1900_21022009_20090221193153256qtst... | Country Calendar - Series 2009 - Episode 1 | 'Lyn of Taranaki' Country Calendar starts its ... | 00:29:57 | Documentary | Agriculture & Horticulture, Business, Ecology ... | Unrated | TV 1 | 2009-02-21 | 19:00:00 | 12564 | https://cdn.etv.org.nz/etv/TV_One_1900_2102200... | 49973516 | Lin Webster's life has taken many unusual twi... | 2009 | 2009 |
| 2009-02-28 | etv/TV_One_1900_28022009_20090228193154256qtst... | Country Calendar - Series 2009 - Episode 2 | 'Shared Fruit' New Zealand and Chinese scienti... | 00:29:56 | Documentary | Agriculture & Horticulture, Asian Studies, Bus... | Unrated | TV 1 | 2009-02-28 | 19:00:00 | 12697 | https://cdn.etv.org.nz/etv/TV_One_1900_2802200... | 52854708 | Yunnan, tucked away in the remote southwest o... | 2009 | 2009 |
| 2009-03-07 | etv/TV_One_1900_07032009_20090307193229256qtst... | Country Calendar - Series 2009 - Episode 3 | 'Grassroots Leader' Bruce Rollinson is a late-... | 00:29:56 | Documentary | Agriculture & Horticulture, Business, Economic... | Unrated | TV 1 | 2009-03-07 | 19:00:00 | 12812 | https://cdn.etv.org.nz/etv/TV_One_1900_0703200... | 51865839 | The season on Mount Ropiku was the best for y... | 2009 | 2009 |
| 2009-03-14 | etv/TV_One_1900_14032009_20090314193119768qtst... | Country Calendar - Series 2009 - Episode 4 | 'High-flying Fur' What's as fine as merino woo... | 00:29:56 | Documentary | Agriculture & Horticulture, Biology, Business,... | Unrated | TV 1 | 2009-03-14 | 19:00:00 | 12939 | https://cdn.etv.org.nz/etv/TV_One_1900_1403200... | 154209978 | With patches of bush and some good keen hunte... | 2009 | 2009 |
| ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... |
| 2024-04-14 | etv/hyundai_country_cale_tv1_20240414_1900 | Hyundai Country Calendar - Season 2024 - 14 Ap... | NaN | 00:30:33 | Documentary, Reality | Agriculture & Horticulture, Business, Careers,... | G | TV 1 | 2024-04-14 | 19:00:00 | 518923 | https://cdn.etv.org.nz/etv/hyundai_country_cal... | 353531481 | Always a favourite on every rural road. Hyund... | 2024 | 2024 |
| 2024-04-21 | etv/hyundai_country_cale_tv1_20240421_1900 | Hyundai Country Calendar - Season 2024 - 21 Ap... | NaN | 00:30:21 | Documentary, Reality | Agriculture & Horticulture, Business, Careers,... | G | TV 1 | 2024-04-21 | 19:00:00 | 518922 | https://cdn.etv.org.nz/etv/hyundai_country_cal... | 344594242 | Always a favourite on every rural road. Hyund... | 2024 | 2024 |
| 2024-04-28 | etv/hyundai_country_cale_tv1_20240428_1900 | Hyundai Country Calendar - Season 2024 - 28 Ap... | NaN | 00:30:10 | Documentary, Reality | Agriculture & Horticulture, Business, Careers,... | G | TV 1 | 2024-04-28 | 19:00:00 | 518972 | https://cdn.etv.org.nz/etv/hyundai_country_cal... | 338647527 | Always a favourite on every rural road. Hyund... | 2024 | 2024 |
| 2024-05-05 | etv/hyundai_country_cale_tv1_20240505_1900 | Hyundai Country Calendar - Season 2024 - 5 May... | NaN | 00:30:45 | Documentary, Reality | Agriculture & Horticulture, Business, Careers,... | G | TV 1 | 2024-05-05 | 19:00:00 | 519110 | https://cdn.etv.org.nz/etv/hyundai_country_cal... | 352367660 | Always a favourite on every rural road. Hyund... | 2024 | 2024 |
| 2024-05-12 | etv/hyundai_country_cale_tv1_20240512_1900 | Hyundai Country Calendar - Season 2024 - 12 Ma... | NaN | 00:30:53 | Documentary, Reality | Agriculture & Horticulture, Business, Careers,... | G | TV 1 | 2024-05-12 | 19:00:00 | 519287 | https://cdn.etv.org.nz/etv/hyundai_country_cal... | 348651686 | Always a favourite on every rural road. Hyund... | 2024 | 2024 |
516 rows × 16 columns
In [5]:
keywords = ["dirt", "soil", "ground", "land", "earth"]
for word in keywords:
# \b means a word boundary, so we only get complete words, and not words like dirty, or grounded, or zealand
df[word] = df.transcript.str.lower().str.count(rf"\b{word}\b")
df
✔️ 1.41 s (2024-05-21T14:37:32/2024-05-21T14:37:33)
Out[5]:
| File Name | Name | Description | Duration | Type | Subject | Age Level | Channel | Recording Date | Recording Time | ... | link | size_bytes | transcript | Year | year_in_name | dirt | soil | ground | land | earth | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Recording Date | |||||||||||||||||||||
| 2008-10-23 | etv/country_calendar_-_air_new_zealand_uniform... | country calendar - air new zealand uniforms 23... | from www.throng.co.nz : this episode of nation... | 00:21:33 | Library | NaN | G | lostplusfound | 2008-10-23 | 00:00:00 | ... | https://cdn.etv.org.nz/etv/country_calendar_-_... | 145371146 | Music Music Music Music Music Music Music Mus... | 2008 | 2008 | 0 | 0 | 0 | 0 | 0 |
| 2009-02-21 | etv/TV_One_1900_21022009_20090221193153256qtst... | Country Calendar - Series 2009 - Episode 1 | 'Lyn of Taranaki' Country Calendar starts its ... | 00:29:57 | Documentary | Agriculture & Horticulture, Business, Ecology ... | Unrated | TV 1 | 2009-02-21 | 19:00:00 | ... | https://cdn.etv.org.nz/etv/TV_One_1900_2102200... | 49973516 | Lin Webster's life has taken many unusual twi... | 2009 | 2009 | 0 | 2 | 0 | 2 | 1 |
| 2009-02-28 | etv/TV_One_1900_28022009_20090228193154256qtst... | Country Calendar - Series 2009 - Episode 2 | 'Shared Fruit' New Zealand and Chinese scienti... | 00:29:56 | Documentary | Agriculture & Horticulture, Asian Studies, Bus... | Unrated | TV 1 | 2009-02-28 | 19:00:00 | ... | https://cdn.etv.org.nz/etv/TV_One_1900_2802200... | 52854708 | Yunnan, tucked away in the remote southwest o... | 2009 | 2009 | 0 | 0 | 0 | 6 | 0 |
| 2009-03-07 | etv/TV_One_1900_07032009_20090307193229256qtst... | Country Calendar - Series 2009 - Episode 3 | 'Grassroots Leader' Bruce Rollinson is a late-... | 00:29:56 | Documentary | Agriculture & Horticulture, Business, Economic... | Unrated | TV 1 | 2009-03-07 | 19:00:00 | ... | https://cdn.etv.org.nz/etv/TV_One_1900_0703200... | 51865839 | The season on Mount Ropiku was the best for y... | 2009 | 2009 | 0 | 5 | 3 | 8 | 0 |
| 2009-03-14 | etv/TV_One_1900_14032009_20090314193119768qtst... | Country Calendar - Series 2009 - Episode 4 | 'High-flying Fur' What's as fine as merino woo... | 00:29:56 | Documentary | Agriculture & Horticulture, Biology, Business,... | Unrated | TV 1 | 2009-03-14 | 19:00:00 | ... | https://cdn.etv.org.nz/etv/TV_One_1900_1403200... | 154209978 | With patches of bush and some good keen hunte... | 2009 | 2009 | 0 | 0 | 0 | 2 | 1 |
| ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... |
| 2024-04-14 | etv/hyundai_country_cale_tv1_20240414_1900 | Hyundai Country Calendar - Season 2024 - 14 Ap... | NaN | 00:30:33 | Documentary, Reality | Agriculture & Horticulture, Business, Careers,... | G | TV 1 | 2024-04-14 | 19:00:00 | ... | https://cdn.etv.org.nz/etv/hyundai_country_cal... | 353531481 | Always a favourite on every rural road. Hyund... | 2024 | 2024 | 0 | 0 | 0 | 1 | 0 |
| 2024-04-21 | etv/hyundai_country_cale_tv1_20240421_1900 | Hyundai Country Calendar - Season 2024 - 21 Ap... | NaN | 00:30:21 | Documentary, Reality | Agriculture & Horticulture, Business, Careers,... | G | TV 1 | 2024-04-21 | 19:00:00 | ... | https://cdn.etv.org.nz/etv/hyundai_country_cal... | 344594242 | Always a favourite on every rural road. Hyund... | 2024 | 2024 | 0 | 1 | 0 | 4 | 0 |
| 2024-04-28 | etv/hyundai_country_cale_tv1_20240428_1900 | Hyundai Country Calendar - Season 2024 - 28 Ap... | NaN | 00:30:10 | Documentary, Reality | Agriculture & Horticulture, Business, Careers,... | G | TV 1 | 2024-04-28 | 19:00:00 | ... | https://cdn.etv.org.nz/etv/hyundai_country_cal... | 338647527 | Always a favourite on every rural road. Hyund... | 2024 | 2024 | 0 | 0 | 2 | 1 | 0 |
| 2024-05-05 | etv/hyundai_country_cale_tv1_20240505_1900 | Hyundai Country Calendar - Season 2024 - 5 May... | NaN | 00:30:45 | Documentary, Reality | Agriculture & Horticulture, Business, Careers,... | G | TV 1 | 2024-05-05 | 19:00:00 | ... | https://cdn.etv.org.nz/etv/hyundai_country_cal... | 352367660 | Always a favourite on every rural road. Hyund... | 2024 | 2024 | 0 | 0 | 0 | 5 | 0 |
| 2024-05-12 | etv/hyundai_country_cale_tv1_20240512_1900 | Hyundai Country Calendar - Season 2024 - 12 Ma... | NaN | 00:30:53 | Documentary, Reality | Agriculture & Horticulture, Business, Careers,... | G | TV 1 | 2024-05-12 | 19:00:00 | ... | https://cdn.etv.org.nz/etv/hyundai_country_cal... | 348651686 | Always a favourite on every rural road. Hyund... | 2024 | 2024 | 0 | 0 | 0 | 0 | 1 |
516 rows × 21 columns
In [6]:
df[keywords].describe()
✔️ 20.4 ms (2024-05-21T14:37:33/2024-05-21T14:37:33)
Out[6]:
| dirt | soil | ground | land | earth | |
|---|---|---|---|---|---|
| count | 516.000000 | 516.000000 | 516.000000 | 516.000000 | 516.000000 |
| mean | 0.182171 | 1.139535 | 1.220930 | 3.451550 | 0.269380 |
| std | 0.569237 | 2.519806 | 1.526983 | 3.661871 | 0.549592 |
| min | 0.000000 | 0.000000 | 0.000000 | 0.000000 | 0.000000 |
| 25% | 0.000000 | 0.000000 | 0.000000 | 1.000000 | 0.000000 |
| 50% | 0.000000 | 0.000000 | 1.000000 | 2.500000 | 0.000000 |
| 75% | 0.000000 | 1.000000 | 2.000000 | 5.000000 | 0.000000 |
| max | 5.000000 | 19.000000 | 12.000000 | 29.000000 | 4.000000 |
In [7]:
df.resample("1M")[keywords].sum().plot(height=800)
✔️ 255 ms (2024-05-21T14:37:33/2024-05-21T14:37:34)
In [8]:
topic_model = BERTopic(representation_model=KeyBERTInspired(), min_topic_size=5)
topic_model.fit(df.transcript)
topic_model.get_topic_info()
✔️ 17.7 s (2024-05-21T14:37:34/2024-05-21T14:37:52)
Out[8]:
| Topic | Count | Name | Representation | Representative_Docs | |
|---|---|---|---|---|---|
| 0 | -1 | 209 | -1_farm_home_country_over | [farm, home, country, over, going, weve, bit, ... | [ At home on every rural road, Hyundai Country... |
| 1 | 0 | 36 | 0_fishing_fish_boat_catch | [fishing, fish, boat, catch, around, bit, abou... | [ At home on every rural road, Hyundai Country... |
| 2 | 1 | 30 | 1_fruit_orchard_trees_tree | [fruit, orchard, trees, tree, weve, are, years... | [ Always a favourite on every rural road. Hyun... |
| 3 | 2 | 18 | 2_horse_horses_riding_farm | [horse, horses, riding, farm, bit, work, going... | [ At home on every rural road, Hyundai Country... |
| 4 | 3 | 16 | 3_farm_land_country_place | [farm, land, country, place, bit, home, very, ... | [ Just what is it about Country Canada that's ... |
| 5 | 4 | 15 | 4_cows_farm_milking_dairy | [cows, farm, milking, dairy, milk, work, make,... | [ At home on every rural road, Hyundai Country... |
| 6 | 5 | 15 | 5_winery_vineyard_wines_wine | [winery, vineyard, wines, wine, grapes, zealan... | [ Good night. It's been a fickle season for gr... |
| 7 | 6 | 14 | 6_farm_deer_dogs_work | [farm, deer, dogs, work, dog, bit, around, goi... | [ At home on every rural road, Hyundai Country... |
| 8 | 7 | 14 | 7_farm_bit_weve_things | [farm, bit, weve, things, going, ha, take, to,... | [ Always a favourite on every rural road. Kyun... |
| 9 | 8 | 13 | 8_beekeeping_bees_bee_hive | [beekeeping, bees, bee, hive, hives, honey, bu... | [ Always a favourite on every rural road. Hyun... |
| 10 | 9 | 13 | 9_zealand_day_island_all | [zealand, day, island, all, very, going, just,... | [ Rest of your night, let's hope the ABs serve... |
| 11 | 10 | 12 | 10_calf_calves_bull_farm | [calf, calves, bull, farm, bulls, very, get, b... | [ At home on every rural road, Hyundai Country... |
| 12 | 11 | 12 | 11_farm_muster_home_country | [farm, muster, home, country, going, around, w... | [ I wouldn't change my life for anything. You ... |
| 13 | 12 | 11 | 12_cherry_cherries_orchard_fruit | [cherry, cherries, orchard, fruit, what, get, ... | [ Always a favourite on every rural road. Hyun... |
| 14 | 13 | 11 | 13_sheep_farm_wool_bit | [sheep, farm, wool, bit, work, going, country,... | [ At home on every rural road, Hyundai Country... |
| 15 | 14 | 11 | 14_milking_dairy_milk_cows | [milking, dairy, milk, cows, farm, cheese, buf... | [ Lin Webster's life has taken many unusual tw... |
| 16 | 15 | 9 | 15_vegetable_produce_grow_growing | [vegetable, produce, grow, growing, soil, make... | [ Always a favourite on every rural road. Hyun... |
| 17 | 16 | 9 | 16_goats_goat_sheep_farm | [goats, goat, sheep, farm, milking, milk, bit,... | [ At home on every rural road, Hyundai Country... |
| 18 | 17 | 8 | 17_lambs_lamb_sheep_farm | [lambs, lamb, sheep, farm, home, make, life, f... | [ At home on every rural road, Hyundai Country... |
| 19 | 18 | 8 | 18_sheep_wool_farm_zealand | [sheep, wool, farm, zealand, merino, bit, are,... | [ Always a favourite on every rural road. Hyun... |
| 20 | 19 | 8 | 19_farm_eggs_egg_chickens | [farm, eggs, egg, chickens, organic, hens, chi... | [ At home on every rural road, Hyundai Country... |
| 21 | 20 | 7 | 20_farm_home_flowers_going | [farm, home, flowers, going, pretty, day, alwa... | [ Always a favourite on every rural road. Kyun... |
| 22 | 21 | 6 | 21_music_zealand_valley_land | [music, zealand, valley, land, gold, farm, wha... | [ Music Music Music Music Music Music Music Mu... |
| 23 | 22 | 6 | 22_pigs_pig_pork_farm | [pigs, pig, pork, farm, zealand, bit, business... | [ At home on every rural road, Hyundai Country... |
| 24 | 23 | 5 | 23_farming_farm_sheep_zealand | [farming, farm, sheep, zealand, wool, blair, o... | [ At home on every rural road, Hyundai Country... |
In [9]:
topic_model.visualize_topics()
✔️ 1.48 s (2024-05-21T14:37:52/2024-05-21T14:37:53)
In [10]:
topics_over_time = topic_model.topics_over_time(df.transcript, df.index, nr_bins=10)
topics_over_time
✔️ 7.72 s (2024-05-21T14:37:53/2024-05-21T14:38:01)
Out[10]:
| Topic | Words | Frequency | Timestamp | |
|---|---|---|---|---|
| 0 | 4 | cows, cow, milking, milk, bullocks | 1 | 1986-12-18 08:29:45.600 |
| 1 | 6 | deer, helicopters, helicopter, choppers, tim | 1 | 1986-12-18 08:29:45.600 |
| 2 | -1 | ron, rons, build, get, work | 1 | 1990-09-26 14:24:00.000 |
| 3 | 7 | musical, instruments, live, instrument, farm | 1 | 1990-09-26 14:24:00.000 |
| 4 | 11 | muster, musterers, hills, horses, sheep | 1 | 1994-06-22 04:48:00.000 |
| ... | ... | ... | ... | ... |
| 99 | 17 | lambs, sheep, farmers, farm, zealand | 1 | 2020-08-16 09:36:00.000 |
| 100 | 18 | sheep, wool, farm, merino, hops | 4 | 2020-08-16 09:36:00.000 |
| 101 | 19 | farm, eggs, egg, chickens, birds | 3 | 2020-08-16 09:36:00.000 |
| 102 | 20 | farm, home, dogs, going, morning | 2 | 2020-08-16 09:36:00.000 |
| 103 | 22 | pigs, pig, farm, brood, bees | 1 | 2020-08-16 09:36:00.000 |
104 rows × 4 columns
In [11]:
topic_model.visualize_topics_over_time(topics_over_time, top_n_topics=20, height=600)
✔️ 62.4 ms (2024-05-21T14:38:01/2024-05-21T14:38:01)