A compact proof‑of‑concept that shows how Revolut Points could become a daily habit instead of an after‑thought.
consumer platform: https://revpoints-plus-7kuw.vercel.app/ vendor platform: https://revpoints-plus-wqpj.vercel.app/
Looks at what the user already buys and recommends partner vendors where paying with points makes obvious sense.
- Ingest the most recent transactions (mock CSV shipped with the repo).
- Find partner vendors in similar categories or with overlapping products.
- Rank by relevance and value of the points offer.
- Display the short list inside a Streamlit mobile‑style UI.
That’s the entire loop—lightweight, fast, and transparent.
| Stakeholder | Benefit |
|---|---|
| User | Clear places to spend points, translated into cash‑like savings. |
| Partner vendor | Extra footfall from a highly‑targeted audience at no acquisition cost. |
| Revolut | Higher points‑redemption rates and stickier daily engagement. |
# clone & enter
$ git clone https://github.com/your‑org/revpoints‑plus.git
$ cd revpoints‑plus
# create venv
$ python -m venv .venv && source .venv/bin/activate
# install deps
$ pip install -r requirements.txt
# run demo
$ streamlit run app.pyOpen http://localhost:8501 and browse the "Spend Points" feed.
transactions.csv– mock user transactions.partners.csv– hypothetical partner catalogue with point offers.recommender.py– k‑NN content‑based recommender (under 100 lines).app.py– Streamlit UI shell.
Tweak the CSVs to see different recommendation results.
Released under the MIT License.
