Tweets about Apple

Diego Duque

Abstract

The purpose of this project is to understand and analyze tweets related to Apple. This company has been getting more and more popular over the years. It is a contemporary reference for high-tech. Using Python brings us the opportunity to massively understanding tweets related to anything, in this case, related to Apple.

Design

The project is designed to unsupervised learning as a final result. Topic model is also performed and at the end of the code will be provided sentiment analysis related to the dataset.

Data

This public Kaggle dataset will be used: 1624 unique values (tweets).

Methodology

Tools

Sklearn

from sklearn.feature_extraction.text import TfidfVectorizer, ENGLISH_STOP_WORDS
from sklearn.decomposition import NMF
from sklearn.cluster import KMeans
from sklearn.manifold import TSNE

NLTK

from vaderSentiment.vaderSentiment import SentimentIntensityAnalyzer

Numpy
Pandas

import pandas_datareader.data as web

Plotting

import matplotlib.pyplot as plt %matplotlib inline
from mpl_toolkits.mplot3d import Axes3D

Summary

Topic Modeling Conclusions (n = 9):

topic_0: STOCK - AAPL stock price, investing ideas and techinques topic_1: Negative_Sentiment - f*** and hate topic_2: Products: Apple products and technical issues topic_3: Apple_Event - rumors, on new devices, new iPhone topic_4: Chargers - related to iPhone and macbook charger bad quality topic_5: STOCK and trade - AAPL stock, and Apple's global trade topic_6: Positive_Sentiment - happy customers topic_7: Founders - Steve Jobs and Wozniak topic_8: Quality - negative coments again broken products and iOS

Sentiment Analysis:

Positive

Negative

Kmeans Clustering & plot by using TSNE

k = 9 so here shows 9 different clusters by colors as plotting by this dimensionality reduction method called TSNE on 2 dimensions.

dieguque/NLP-Project