User:Dalmeidad

Week 1: June 4th, 2018

Tuesday (6/5/18)

• Met with Dr. Praveen and Joe, discussed the outline of the project, as well as plans for week and other logistics
• Discussed overview of internship with Dr. Brylow, set up all of my accounts
• First read-through of the 2017 research paper of this topic - "Forecasting Stock Prices using Social Media Analysis"

Wednesday (6/6/18)

• Took notes on 2017 research paper, formulating initial questions and listing topic areas in which to focus on understanding
• Looked through last year's code
• Began reading through last year's references, starting with "Stock Market Prediction System with Modular Neural Networks"

Thursday (6/7/18)

• Met with Dr. Praveen and Joe, solidified ideas of what goals to consider for project
• Began making a research table, classifying readings through ideas/methods and conclusions
• Finished "Stock Market Prediction System with Modular Neural Networks" and added to table
• Finished half of week 1 of Coursera Data Science in Python

Friday (6/8/18)

• Built up basic knowledge about neural networks and the different types
• Brainstormed ways to approach last years project and improve upon it/ make it my own

Week 2: June 11th, 2018

Monday (6/11/18)

• Found around 8 papers pertaining to the project
• Created and filled out an excel spreadsheet covering the papers and their main ideas

Tuesday (6/12/18)

• Met with Dr. Praveen and Joe, solidified plans for bringing in other sources to make model better and considering N day look ahead
• Finished the papers found on Monday
• Started week 1 on Coursera

Wednesday (6/13/18)

• RCR training

Thursday (6/14/18)

• Met with Dr. Praveen briefly
• Finished week 1 Coursera
• Found more papers for background knowledge
• Worked on converting python 2.7 to python 3

Friday (6/15/18)

• Worked on better understanding pandas and other packages being used in previous year's code
• Started week 2 of Coursera

Week 3: June 18th, 2018

Monday (6/18/18)

• Met with P.H.D student Joseph Coelho and worked through Scott's code from last year's REU
• Continued to work independently on understanding the code

Tuesday (6/19/18)

• Met with Dr. Praveen and Joe, discussed the main goal and steps to take in improving the project
• Worked on making slight alterations the code and seeing the results
• Looked over some of the past readings I had done

Wednesday (6/20/18)

• Finished week 2 of Coursera
• Developed a plan as to how to implement changes to Scott's code in following weeks

Thursday (6/21/18)

• Worked on better understanding the method of finding and analyzing content of URLs in twits

Friday (6/22/18)

• Mostly added ability to parse to URLs
• Copied over Scott's code into a clean Jupyter Notebook and added important comments for myself

Week 4: June 25th, 2018

Monday (6/25/18)

• Finished the code to parse and obtain sentiment from URLs
• Began running tests, realized that running the code was very time consuming and began to look for a way around this

Tuesday (6/26/18)

• Met with Dr. Praveen and discussed my current progress on the code as well as steps to take during the coming weeks
• Began to write code to take all twit data and extract sentiment from the URLs, then save the data to a CSV file for future use

Wednesday (6/27/18)

• Worked on Midterm presentation
• Practiced Midterm presentation

Thursday (6/28/18)

• Presented and watched presentations of Midterm presentation

Friday (6/29/18)

• Worked on week 3 of Coursera

Week 5: July 2nd, 2018

Monday (7/2/18)

• Finished week 3 of Coursera

Tuesday (7/3/18)

• Added to code ability to export data taken from URL's into csv's

Wednesday (7/4/18)

• Brainstormed ways in which to include sentiment score from the URL's into the final prediction model
• LaTeX presentation, plus some further research to try to familiarize myself with it

Thursday (7/5/18)

• Started week 4 of Coursera

Friday (7/6/18)

• Worked on transferring files to lab computer and configuring applications to run sentiment extractor on them
• Read through 'Twitter Mood Predicts the Stock Market' and summarized

Week 6: July 9th, 2018

Monday (7/9/18)

• Began learning about tweepy and setting up a twitter application
• Finalized lab computer prep, just require certain packages to be installed

Tuesday (7/10/18)

• Read different articles to see which parts of the text were more important to a sentiment score
• Started creating personalized sentiment analyzer using the Loughran and McDonald Sentiment Word List

Wednesday (7/11/18)

• Researched creating sentiment analyzers
• Worked on the sentiment analysis code

Thursday (7/12/18)

• Met with Advisor
• Worked on note sheet from readings
• Started looking at Naive Bayes Classification tutorial