Graphs

Topic Classification with deep learning

1. R script to train a GLMnet model using 1.6 million labeled Twitter tweets Result: Sentiment Max. Accuracy of 87.67% (Positivity). 2. R notebook of Week 5 (the naive approach) Result: 74% Positivity 3. Compare the sentiment scores of the two methods, Result: ML sentiment score is more accurate compared to naive approach

Network graph of Samsung Galaxy Twitter users. 1. Network Measures 2.Histogram of Node Degree 3.Network Diagram 4.Highlighting Degrees and layouts (Fruchterman Reingold, Graphopt, Kamada Kawai) 5. Hubs and Authorities 6.Community detection 7.Obseravtions

1. Twitter account and set up : developer API credentials. Collecting Twitter data from its API on a specific search term. 2. Creating a word cloud of the tweets {the more a specific word appears in a source of textual data of tweets , the bigger and bolder it appears in the word cloud.}

(1). Word cloud to show what users talk about Zynga on Twitter, which gives important information at a glance i.e., the more a specific word appears in the Zynga twitter data, the bigger and bolder it appears in the word cloud. (2). Histogram to show the Zynga User's Sentiment. (3). Histogram to show Zynga users' involvement on each weekday (4). Recommendations for Zynga to increase its monthly active users.

1. Word cloud to show what users talk about Zynga on Twitter, which gives important information at a glance i.e., the more a specific word appears in the Zynga twitter data, the bigger and bolder it appears in the word cloud. 2. Histogram to show the Zynga User's Sentiment

1. Word cloud to show what users talk about Zynga on Twitter, which gives important information at a glance i.e., the more a specific word appears in the Zynga twitter data, the bigger and bolder it appears in the word cloud.

Charts presenting profile of Trump twitters (a) User gender distribution by country (b) User posting count by country

(a) A histogram to show the distribution of sentiment scores of Samsung tweets related "sports". (b) A pie chart to show the percentage of Positive, Negative and Neutral tweets.

Pie chart plot of percentage of tweets mentioning Health related topics and a word cloud of those tweets.

Word cloud plot of the trump data users from New York City & Washington and respective list of word frequency.

The calculated average MSRP of each product.