NAHDI · X DEEP LISTENING ·
24 Months of Public Conversation Across Saudi Arabia
This case study never fabricates. Real, source-linked posts are shown; aggregate scores activate only when a full dataset is connected.
Real, entity-resolved X posts collected live via Bright Data, each linked to its public source. Toggle original and English. Flags mark in-window vs historical context.
Not every "Nahdi" is the brand. Every post carries a relationship, a topic and an entity-confidence score. Only high and medium confidence influence analysis.
Editable in one config file.
The entity-resolution layer.
Not a decorative number. Seven weighted components, a disclosed formula, and an honest status: the score activates when the 24-month mention corpus is connected.
Thirteen administrative regions. Regions carrying a real signal in the collected sample are marked; the rest honestly read "no geo data connected".
Kept strictly separate, so you always know which is which.
Current official Nahdi properties. Contextual only, not merged into the X Brand Health score.
Source, window, queries, rules, formula and limitations, in the open.