What makes a comic book valuable?

Comics books from the Golden, like the first appearance of Superman, and Silver age of comics are getting sold for millions in auctions.

Modern-age comic books can also get really valuable within a couple of years (e.g. the first issue of The Walking Dead, also a hit TV series, is valued around 1800$).

Unfortunately no one can predict whether a comic is going to rise in value beforehand! Could ML methods help with that?

The Data

This dataset was scraped from comicbookrealm.com, a free price-guide website for american comic-books, at the beginning of January 2022. It includes almost all data relevant to a comic issue like issue number, date, cover price, current value, synopsis, creators, characters, publisher, volume type,cover image link, cover variant, print number etc.

I plan to repeat the scraping from time to time in order to update the dataset with latest issues.

Content

The dataset can be downloaded in .csv and .parquet format for a more convenient size: ** Note: it all columns giving a link the address of the website https://comicbookrealm.com/ should be added at the beginning of each string.**

Take a look at a typical series link to understand better all the relevant info

comic_data.csv: Includes all information about each single issue that could be scraped. In detail: title: the series title title_link: the link to the title in the website issue_link: the link to each issue of the title cover_link: the link to the cover image of the issue issue: the number of the issue cover_date: the date (month and year) the issue was released cover_price: the price in which the issue was sold current_value: the highest price the issues was ever sold hist_price_link: the link for the historical prices as recorder on the website searched: the times the issue was searched in the site's database owned: the times a user logged the issue as owned pages: number of pages in issue rating: the average rating by the site's users rating count: the number of rating votes ISBN-UPC: the unique issue code est_print_run: the estimated number of copies printed variant_of: if the issue is a variant cover of preview: if the issue is a preview synopsis: brief description of the plot and/or issue edition e.g. if it is a limited edition or exclusive etc. contributors_names: lists of the names of the contributors contributors_roles: lists that correspond to the role of each contributor (e.g. script, inks etc.) characters: List of characters appearing in the issue as well as extra info (e.g. secret identity, link to character photo, special event like first appearance, death etc.) check on the the website for an example on the link above pub_id: the unique publisher's id as given by the website 'volume': volume number of the title or whether it is a one-shot, mini-series, limited-series etc. years: The year span the title was published issues_total: Total number of issues in the series pub_name: The name of the publisher pub_titles_total: total number of titles released by the publisher pub_issues_total : total number of issues released by the publisher

Inspiration

What makes a comic valuable? What common characteristics do valuable comics share? What other data would be needed to put odds on a comic book rising in value?

Let me know of your ideas and I can enrich the current dataset with more relevant data like sales data, historical prices etc. All comments and suggestions are welcome!