The bar market in Moscow is saturated, and classic guides offer too broad a selection. If you want to play a quiz or participate in other activities—2GIS or Yandex Maps won’t show you where you can join in. You have to follow the social media accounts of specific bars and track their announcements. Conversely, it’s unpleasant when you’ve searched for a cozy spot to quietly talk with friends, only to stumble into a «social bomb.» Too much time is spent searching for that perfect bar for the evening.Project Goal: To create a Telegram bot that selects the ideal venue for a user to enjoy alcoholic beverages.

Mapping and Geo-ServicesYandex Maps / 2GIS — Name, address, coordinates, hours of operation, category, rating, number of reviews, phone number, website, photos, social media links. — Frequency: Rating — every 14 days; new bars — weekly (this does not refer to the initial database creation).Bar Social Media and Channels1. Bar Telegram Channels • What it provides: Announcements for quizzes, karaoke nights, parties, and special offers — the primary source for events. • Pros: Bars actively manage their presence here, high freshness, machine-readable text. • Cons: Requires a list of channels and regular updates, parsing via Telegram API/bot, possible private channels. • Frequency: Daily.2. Bar VK Communities • What it provides: Posts with announcements, events (VK Events), discussions, and reviews. • Pros: Open API, numerous bars, supports geolocation. • Cons: Some bars have moved to Telegram, posts can be unstructured. • Frequency: Daily.3. Bar Websites • What it provides: Listings, menus, rules (deposit, dress code), and reservation information. • Pros: Official source, structured listings. • Cons: Many bars lack websites or do not update theirs. • Frequency: Weekly.Quiz Listings (QuizPlease, Quizium, «Quiz, Please!») • What it provides: Quiz schedules per bar, including date, time, price, and registration link. • Pros: Structured data, events ready to go. • Cons: They only cover their specific format. • Frequency: Weekly.Direct Agreements with Bars • What it provides: Up-to-date listings, exclusive events, discounts for bot users, and data verification. • Pros: Maximum accuracy, potential for monetization. • Cons: Requires negotiation and time.

Human Curation The responsibility of editors, content managers, and experts. Humans are responsible for data collection, quality assurance, and decision-making.

Data Collection and Curation: • Visiting bars. Meeting with owners and bartenders • Reviewing bar social media to extract announcements for quizzes and activities • Creating a brief initial description based on criteria: noisy/quiet, for dates/for groups/for a single user • Verification: checking if the data is current, if the bar is still open, or if the format has changed • Developing event filters: activity type, regularity, participation requirements • Reviewing the selection of events with low levels of informational accuracy.Product Development: • Designing dialogue scenarios within the bot. The user inputs criteria for venue selection • Setting up prioritized attributes for bot surveys • Moderating user reviews and complaints • Partnerships with bars. Inputting current announcements and discounts into the program • Commercial agreements with bar owners

Machine Processing The responsibility area of servers, scripts, and neural networks. The machine is responsible for collection, scaling, speed, and mathematics.

Data Collection: • Point de-duplication • Geocoding addresses and calculating distances from the metro or a user-specified point • Gathering event informationProcessing: • Automatic rating updates • Filtering based on user query parametersSelection and Delivery: • Ranking bars based on the user’s request • Considering user history • Creating a weekly selection of «Top 5» and «Anti-ranking» based on user reviews • Notifications about new bars, events, or happenings based on prior client requests • Event reminders