A scoring engine that combines search demand, competition difficulty, and monetisation potential into a single niche score, with daily automated runs and YouTube intelligence layered on for niches with real demand.
Picking a profitable affiliate niche takes hours of cross-referencing keyword tools, commission rates, and competitor sites — and the answer goes stale every few months.
solution
Python pipeline discovers niches, samples Google Autocomplete for keyword volume, scores affiliate programs against each niche, and stores history so you can watch a niche heat up or cool down. YouTube intelligence kicks in only above a demand threshold so the panel stays focused.
architecture
external → compute → store → ui
outcome
Tracks 33+ niches with automated scoring. Niche research goes from hours to minutes; opportunities surface before the obvious money has been priced in.
stack
PythonFlaskPostgreSQLGoogle AutocompleteYouTube Data APIDocker
capabilities
Niche discovery and de-duplication
Multi-factor scoring (demand · monetisation · density · ease)
Affiliate program matching per niche
YouTube intelligence above demand threshold
Historical trend tracking with score explanations
lessons learned
01Score explanations matter more than scores. The "why" column unlocks trust faster than tweaking weights ever did.
02Sampling Google Autocomplete is a good-enough volume proxy when you can't afford a real keyword tool — and it is harder for competitors to game.
03Adding YouTube data to every niche overwhelmed the UI; gating it behind a demand threshold made the dashboard usable again.