Analytics

PhilPapers web analytics, Zenodo, and Academia provided the following data.

It counts total hits, unique hits, downloads, and the unique-hits-to-downloads ratio from PhilPapers.

Views, downloads, and views-to-downloads ratio from Zenodo.

Views, downloads, and views-to-downloads ratio from Academia.

Combined views, downloads total, and overall unique hits/view to downloads total ratio from PhilPapers, Zenodo, and Academia.

Geographical spread, location list, total unique cities and countries, and a timeline for the entire series.

References for hits, when available, are in the Google Sheet. Views/hits, downloads, and the views/hits-to-downloads ratio are also available per paper and per platform in the Google Sheet.

I sorted through it, one row and column at a time, in Google Sheets in one browser window, with a data source in another window.

Added the Google Sheet to Data Studio once done, then compiled the graphs in various reports, and finally embedded those graphs in the website.

So entirely manually. No scrapers, no bots, no whatever.

As accurate as it can be. I have spent weeks upon weeks meticulously and painstakingly compiling the data for the utmost accuracy.

That being said, I cannot say with 100% certainty how much might be bots or AI scraping the work, nor how much of it is actual people. However, the sheer spread suggests more actual people than bots or AI.

The Hohenwald, TN USA hits are partially me. I cannot say how much, as PhilPapers doesn’t provide granular enough details, but not every single one is me because the dates do not line up with when I was active. I also want to be clear that this affects hits only.

I have not downloaded my own papers; that should be obvious, as I literally have copies of my own work, but for the sake of being clear, I wanted to state that.

Also, all the data is just from what I can see. I cannot see anything from private sharing via email, Slack, Teams, or any other channel.

In other words, I see the first line of interactions. Someone viewing or downloading. I cannot see if they share it privately with a friend, colleague, or otherwise.

As such, all numbers should be treated as the floor, not the ceiling.

The views and downloads are public on Zenodo, Downloads are public on Philpapers, and I provide a video showcasing the hits from Philpapers/Philpeople web analytics. Views are also public on Academia. So you can compare the metrics if you want.

Google Trends data, while it exists, doesn’t cleanly translate into useful data, as it’s hard to say what’s actually related to my work and what's completely unrelated.

Domain Interest as well. Academia gives some, but it’s 0.001% of my total data, so it’s not included as it’s so little and messy.

There are also a few other completely different types of data that cannot be turned into a graph at all or presented in any form.

Otherwise, you see 95% of what I see.

Because I believe in transparency, I wanted to compile the data myself anyway and had been doing some of it for a while.

Also, I got nothing to lose by being so transparent, and most importantly, because why not?

Yes, the data will be updated periodically, providing new information for the graphs, and a new version of the spreadsheet will be uploaded when it happens.

A copy of the current Google Sheet is available below as a zipped folder with HTML files for each sheet, and previous versions can be found at the link below that one, when available.