Forecast retail, buylist, and inventory movements of Magic, the Gathering trading cards.
Major trading card game websites like tcg player and card kingdom publish pricing and buylist information of Magic, the Gathering cards in secondary markets. These scripts scrape and analyze their publications to forecast retail, buylist, and inventory movements of Magic, the Gathering trading cards. Automating the collection of price discrepancies between different market's valuation of specific Magic, the Gathering trading cards creates opportunities for profit, whether from foreign markets tokyomtg or generating your own buylist that appear above retail, but are actually below current buylist offers via cardsphere.
All of these scripts are written in R (Rstudio), and most will require some familiarization with Docker containers to recreate.
remDr = remoteDriver(remoteServerAddr = "IP4", port = 4445L, browser = "chrome")Above IP4 will need to be inputed according to your own Docker Containers.
drive_auth(email = "GMAIL")
drive_auth(use_oob=TRUE)All "GMAIL" elements will need to be replaced with your personal gmail credentials to read and write from your drive.
username$sendKeysToElement(list("EMAIL"))
passwd$sendKeysToElement(list("PASSWORD"))Login credentials for specific sites will also be required.
| Lines | Purpose |
|---|---|
| 1-50 | Needed Packages and Functions |
| 51-448 | Data Scraping (Post, Rvest, RSelenium) |
| 449-996 | Create Analytical Reports For Each Site |
| 997-1436 | Consolidate KPI's Down to 4 |
| 1437-1578 | Compare TCG & CK Granularly |
| 1579-1637 | Create Personal Buylist on Cardsphere |
| Lines | Purpose |
|---|---|
| 1-75 | Needed Packages and Functions |
| 77-129 | Determining Card Legality by Format |
| 130-770 | Review Format Retail Avgs w/ ggplot |
| 770-1457 | Review Format Buylist Avgs w/ ggplot |
| 1458-1594 | Merge Above Avgs w/ ggplot into one graph |
| Lines | Purpose |
|---|---|
| 1-55 | Needed Packages and Functions |
| 56-154 | Prepare Data in Needed Time Series Format |
| 155-219 | Holts-Winter Forecast |
| 220-293 | Feed Forward Neural Network Forecast |
| 294-374 | Arima Forecast |
| 375-475 | Join Forecasts, Choose By Lowest MAPE |
| 476-531 | Growth/Decline List Based Off MAPE & RMSE |
| Lines | Purpose |
|---|---|
| 1-51 | Needed Packages and Functions |
| 56-71 | Create Ranking Buckets for Buylist Demand |
| 72-117 | Request API Data & Apply Ranking Buckets |
| 118-166 | Scrape Cardkingdom Bestseller List |
| 167-253 | Back Calculate Throughput Algorithm |
| Lines | Purpose |
|---|---|
| 1-38 | Needed Packages and Functions |
| 41-239 | Scrape Major Japanese Vendor Tokyomtg |
| 241-278 | Analyze against US Markets for Arbitrage |
- Obtains base level card information for connecting scrapes between all websites. mtgjson vastly outperforms scryfall in terms of time to complete, but the option is yours.
- Transfer data from CSV trees to Bigquery Database
- Review daily outputs from Wizards of the Coast to gauge player demand.