Research project for ML and NLP: Analyze news articles about App Inventor

[halatmit]

Suggestion from Benji:

Factiva is a business information and research tool (owned by Dow Jones). It enables you to query news that is up to 5 years old. This could give us insight regarding how App Inventor has been represented in the news over the past 5 years.

I was inspired by Vitomir Kovanovic, whom I met earlier this week. He used Factiva to access news data on MOOCs and write What public media reveals about MOOCs: A systematic analysis of news reports. He did topic modeling to understand common themes in the public discourse on MOOCs.

Factivia is available via MIT libraries (google "mit factiva). I attached the data from a query for "app Inventor" from all sources for the past 5 years (29 June 2011 - 1 July 2016). n=445, but there are over 100 articles listed as duplicate so the count may be closer to 300 depending on duplicates are handled.

This could be a potential student project for somebody interested in machine learning and natural language processing. AppInventorNewsFactiva.zip

[shivakrishna9]

I am interested in this NLP project. Can I submit a proposal for this as part of GSoC?

[sjais789]

I am interested in the project but i have some queries like,is it compulsory to use Factiva tool,as it lags behind some news publisher and it's not containing all the articles related to MIT app inventor.

[sjais789]

I have applied LDA algorithm to generate topic based on the word frequency for the first document file of (AppInventorNewsFactiva.zip) provided by @halatmit. I will soon generate topic on all the document files. https://github.com/sjais789/LDA_MIT_App_Inventor Any suggestions to improve the result @halatmit @shivakrishna9 .

`