The goal of this project was to find out about early market reaction to iphone X which has been announced on Spetember 12, 2017. This was achieved by mining and analyzing people opinions on a relevent website (https://techcrunch.com/2017/09/13/i-dont-want-the-new-iphone-x-and-i-cant-be-alone/).
Users' online reviews and comments from www.techcrunch.com for iPhone X which will be released in November 2017.
People opinions sentiments for new iPhone X
For running this code you need to have Python 3.6 and install sklearn, lxml, requests, and pandas packages on your machine.
A- Crawling online comments from web, cleaning up data and extracting important information from html content:
I developed a Python web crawler to pull out people's comments from techcrunch.com websites where they discussed the new iPhone. This crawler downloads all reviews and comments posted by people. I used lxml and requests Python libraries to extract important data from HTML contents.
B- Running sentiment analysis using my developed sentiment analysis model:
For the second step, I used Amazon user reviewes for a verity of unlocked mobile phones. Using ratings and text of reviews, I developed a logistic regression model to distinguish postitive and negative reviews. I used "CountVectorizer" from sklearn.feature_extraction.text library to find probabilty of 1-word and 2-words combinations in postitive and negative reviews. After verifying accuracy of the model which is around 97% and finding 10 most commen words in postive and negative reviews, I applied the developed model to the extract comment list.
I collected 128 comments( including replies to comments) in total where 78 positive, and 50 of them were labelled negative respect to the topic (iPhone X). Below is the details:
pos=60.9%
neg=39.1%
Above results show how people thought about the iPhone X release after its demo on September 12, 2017 and it does not reflect the developer's opinion.
If you have any question regard to this project, please contact me @ "ehsan DOT sadeghi AT gmail DOT com".