README
The author of this codebase approaches the Qanon phenomenon from a skeptical perspective.
If you are simply looking for the screenshots.csv dataset and don't want to reproduce it, go to our google sheet.
SUMMARY
This python 3 codebase allows a user to create a dataset of metadata for the images posted by Q. It also produces a dataset that subsets the Q image metadata to focuse on 29 images marked "Screenshot." 27 of these images have time information and it is likely this metadata was added by Q's own device rather than inherited from another source. In any case, we have not been able to find the source for the metadata online using reverse image source and metadata extraction tools.
The author was able to run the code in python 3.8.5.
DETAILED INSTRUCTIONS
On ubuntu linux, the shell script install_exiftool.sh should download and install exiftool. On other environments you should follow the installation process on the exiftool website (https://exiftool.org/).
The shell script scrape_metadata.sh assumes this repository is located in the user's home directory. The script should set up folders, set up a virtual environment in the directory, download images and metadata, and create two datasets. The initial run of the exiftool command should appear at image-metadata/data/metadata.csv. The final output of this script is a dataset of image metadata decorated with some information about the drops, which is useful for analysis. This dataset will be located at image-metadata/analysis/txt/metadata_decorated.csv.
Those who wish to download the images and initial posts.json separately can do so from https://qalerts.net/media/ and https://qalerts.app/data/json/posts.json for now. (Take note: we are unaffiliated with qalerts, which is a Qanon-promoting website.)
To create the screenshots dataset, activate the virtual environment and run python3 main.py --create_data_set. The dataset will be output to image-metadata/analysis/txt/screenshots.csv.
Finally, the ipython notebook image-metadata-analysis.ipynb reads in the screenshot data and performs analysis on its timestamps.
FURTHER INFORMATION
You can reach the code authors at abigail.wxyz@gmail.com and robertamourgoogs@gmail.com.