Recent Updates (11-27-2017) I've posted a [clustered full dataset](https://www.kaggle.com/jeffkao/fcc-net-neutrality-comments-clustered). This might give you a boost on the work you're doing with the data! I've posted a [vectorized subset w/ 100,000 data points](https://www.kaggle.com/jeffkao/proc_17_108_level0_unclustered_sample_vectorized/) sampled _after_ manual reduction of the dataset after EDA. Context Cleaned data used in researching public comments for FCC Proceeding 17-108 (Net Neutrality Repeal). Data collected from the beginning of submissions (April 2017) until Oct 27th, 2017. The long-running comment scraping script suffered from a couple of disconnections and I estimate that I lost ~50,000 comments because of it. Even though the Net Neutrality Public Comment Period ended on August 30, 2017, the FCC ECFS system continued to take comments afterward, which were included in the analysis. I did a write-up on the results here: https://hackernoon.com/more-than-a-million-pro-repeal-net-neutrality-comments-were-likely-faked-e9f0e3ed36a6 Content Document id, Text of the post, and number of duplicates found for that text was all that was necessary to generate the results. Text is raw & unchanged from the original. I'm working hard to get online a fuller set of data w/ other important metadata fields. Cleaned-up notebooks used are available on [github](https://github.com/j2kao/fcc_nn_research). I am posting the notebook for Exploratory Data Analysis first, and will include others as they are cleaned up. Please share with the rest of us what interesting insights you glean from the data! Tweet at me [@jeffykao](https://twitter.com/jeffykao).