Lucas Oskorep
26 lines
673 B
Python
26 lines
673 B
Python
from textblob import TextBlob
|
|
import pandas as pd
|
|
import preprocessor as p
|
|
from pathlib import Path
|
|
from glob import glob
|
|
|
|
def add_sent_analysis(file, dest):
|
|
df = pd.read_csv(file)
|
|
pol = []
|
|
subj = []
|
|
for index, row in df.iterrows():
|
|
tweet = row["tweet.text"]
|
|
tweet = p.clean(tweet)
|
|
tb = TextBlob(tweet)
|
|
pol.append(tb.polarity)
|
|
subj.append(tb.subjectivity)
|
|
|
|
df["pred_polarity"] = pol
|
|
df["pred_subj"] = subj
|
|
df.to_csv(dest, quoting=2, index=False)
|
|
|
|
for file in glob("./data/*"):
|
|
print(file)
|
|
print()
|
|
add_sent_analysis(file, "./processed_data/" + Path(file).name)
|
|
# default accuracy is 72% - NICE! |