We are doing that following a Social Movement, and have accumulated very big quantities of tweets (about 3 months following-up: 500.000 registers), but there is a moment in which there is a cut-off when downloading data (tweets). We use a server for this. Any experience on this? We are very interested in reading your contributions in this area. We did a similar interdisciplinary research with Delicious (Bookmarking services) and now we are improving methods for extracting big data, automatic coding and classification, and analysis in Twitter following a Social Movement (see link in the case of Deliciuos).
http://hdl.handle.net/10272/6414
I am assuming you do not have software and are unable to generate your own so here is what might work or at least give you some direction:
1. Prepare an excel sheet with all relevant classifications or research variables.
2. Identify the key words which correspond with each research classification.
3. Go through the tweets and cut and paste all relevant text material onto a regular Word Document (docx), the newer the version the better.
4. Using the Word filter function, open up a key word search and enter your key words (one at a time); this should result in a numerical value.
5. Enter the numerical values for each word onto your excel sheets for analysis.
6. Assuming this gets you some of the information you require, you will have to do reliability and validity tests to ensure the data is applicable.
I used this technique to research political biases in newspapers during elections.
If you have any questions, let me know......Good luck.
- Badges
- Science topic
- Similar topics
- Social Science
- Media
- Social Media