HuaRenCa
Back to Forum
Community

A Quantitative Credit Card Recommendation Tool Based on Multi-Attribute Decision Making (MADM)

宇惠
宇惠

1 week ago

Try it online: https://card-compass.netlify.app (free; no registration required; your candidate list, preferences, and any data you modify are stored only in your own browser)

Why I built this tool

When faced with too many credit card options, many people (myself included) often fall into "choice paralysis": Is this card actually useful to me? Which is more worthwhile compared with another card? We often have to spend a lot of time reading guides, calculating effective annual fees, and evaluating rewards rates. So some people go straight to posting questions, but often still don't get satisfying answers, because whether a credit card is good or bad is a highly subjective question. I'm sure everyone has often seen posts with titles like these:

Newbie questions (click to learn more details)

Choosing a card based on spending habits or travel needs (click to learn more details) To replace the cumbersome mental weighing process with a more rigorous, quantitative approach, I developed this quantitative credit card recommendation tool based on multi-attribute decision making (MADM). It aims to fill the gap in credit card selection tools and help users scientifically and objectively choose the card that best suits them based on their personal spending structure and preferences.

Some AI Slop and Methodology (click to learn more details) Usage tutorial

Summary (click to learn more details) Card Compass is a tool that helps you choose US credit cards: select a few candidate cards, tell it what you care about (sign-up bonus, effective annual fee, dining/grocery/travel rewards, lounge access...), and it will produce a ranking that is uniquely yours based on your own preferences. Every score can be broken down to show which factors contributed to it.

Links

Try it online: https://card-compass.netlify.app (no registration required; your candidate list, preferences, and any data you modify are stored only in your own browser)

Summary

This project is still in its early stages, so there are likely to be some bugs. Reports are welcome.

A Quantitative Credit Card Recommendation Tool Based on Multi-Attribute Decision Making (MADM)

4
1.9K

Comments (4)

Your avatar
Sign in to comment
暮露成霜
暮露成霜1 week ago

You mentioned in the main text that each score can be broken down to see which factors contributed to it, and that's something I care about. I'd like to follow up: how are the weights set—do you let users drag sliders to set priorities themselves, or is there a default template that can then be fine-tuned? Also, if two cards have very close total scores but big differences in their sub-scores, how does the tool handle that? Would it prompt something like "these two actually have different strengths" instead of forcing a ranking?

陈佳明
陈佳明1 week ago

The data is stored only in your browser locally, so it will be lost if you switch devices or clear your cache—keep that in mind.

蔡文浩
蔡文浩1 week ago

In MADM, even a slight change in weights can flip the rankings. It's worth flagging the items most sensitive to weight changes on the results page, so users don't treat the rankings as absolute conclusions.

叶塞
叶塞1 week ago

In reality, many people get stuck when choosing a card because their future spending is uncertain—for example, whether they'll travel more for work or open a new spending category. If a tool allowed users to set a range for different spending categories instead of only entering a fixed monthly spending amount, it might better reflect what people are actually unsure about.