LoraBot - Tool for analyzing activity in chatbots (Telegram, Slack, Discord, Vk, WhatsApp, Viber and all other)
created in Python.
Article about it in English: https://medium.com/@alekskram/analytics-of-chatbots-in-python-e044024c0db7
Article about it in Russian: https://habr.com/ru/post/653535/
- lorabot directory with code that will help you in your analytics
- docker-compose.yml for create database for analytics near your Project
- .env stores credentials for accessing analytics to the database and a password for accessing analytics from the bot
- *_bot.py with embedded analytics (pay attention to telegram_bot.py, there are the most examples)
- First of all, you need to download LoraBot as zip or using git:
and put near your bot.
git clone https://github.com/aleksspevak/lorabot.git
- Create postgresql database for LoraBot by running the following command
and put database near your bot. If you don't have docker-compose installed, you can read how to install here: https://docs.docker.com/compose/install/
docker-compose up -d
- Install the libraries, that need for LoraBot:
pip3 install -r requirements.txt
- Set password for accessing analytics in .env file:
ANALYTICS_PASSWORD=lorabot - Initialize LoraBot in your bot:
from lorabot import LoraBot lora_bot = LoraBot("your bot's name")
- Set where you need LoraBot functions, to start getting information for analytics as in examples:
Great! Now you get analytics from your bot.
#to track new users, for example, in a telegram, it is appropriate to put a function in the processing functions of the /start command lora_bot.user(USER_ID, NAME_OF_REGION) #to track commands, menu messages, messages lora_bot.message(TEXT, TEXT_TYPE, USER_ID) #to track events lora_bot.event(EVENT, EVENT_TYPE, USER_ID) #to track review lora_bot.review(REVIEW, USER_ID) #to track bot assessment lora_bot.assessment(RATING_IN_INT_FORMAT, USER_ID)
- To get analytics, make some kind of conditional branching, for example in a file telegram_bot.py a condition has been made to receive a message with a keyword, after that the password is checked and after that the bot owner gets access to analytics:
If you have your own database or want to deploy it in the cloud as a SaaS, you need to create a user yourself, run the create_tables.sql file and put credentials in the .env file.
Below describe all functions for analytic in LoraBot.
Notice that some function has parameters to set period of analytics,volume of returning data or to set message/event type.
You can learn more about parametrs on function's documentation.
Also some functions return only text information, but there are functions that return photo and text information.
Here some examples how return information from analytics in telegram bot:
#Return total information of users that using your bot(only text)
info = lora_bot.analyze_total(START_PERIOD, END_PERIOD)
bot.send_message(message.chat.id, info)
#Return information about daily active users(photo+text)
photo, info = lora_bot.analyze_dau(START_PERIOD, END_PERIOD)
bot.send_message(message.chat.id, info)
bot.send_photo(message.chat.id, photo)period_start - beginning of the analysis period
period_end - end of the analysis period
message_type - type of message from the user
event_type - type of event from the user
messages_for_funnel - array of messages in right order for funnel
events_for_funnel - array of events in right order for funnel
volume - amount of data to show
#1 analyze_total - Return total information about your users, their messages and events.
Parametrs: period_start, period_end.
#2 analyze_user_number_accumulation - Analyzes number of users with accumulation
Parametrs: period_start, period_end.
#3 analyze_new_user - Analyzes number of new registered users
Parametrs: period_start, period_end.
#4 analyze_hour_activity - Analyzes number of message by hours activity
Parametrs: period_start, period_end.
#5 analyze_dau - Analyzes number of active users by days
Parametrs: period_start, period_end.
#6 analyze_wau - Analyzes number of active users by weeks
Parametrs: period_start, period_end.
#7 analyze_mau - Analyzes number of active users by month
Parametrs: period_start, period_end.
#8 analyze_yau - Analyzes number of active users by year
Parametrs: period_start, period_end.
#9 analyze_messages_number - Analyzes number of messages by days.
Parametrs: period_start, period_end, message_type.
#10 analyze_messages - Return messages.
Parametrs: period_start, period_end, message_type, volume.
#11 analyze_messages_type - Analyzes messages by types.
Parametrs: period_start, period_end.
#12 analyze_messages_funnel - Create funnel for messages.
Parametrs: messages_for_funnel, period_start, period_end.
#13 analyze_events_number - Analyzes number of events by days.
Parametrs: period_start, period_end, event_type.
#14 analyze_events - Return events.
Parametrs: period_start, period_end, event_type, volume.
#15 analyze_events_type - Analyzes events by types.
Parametrs: period_start, period_end.
#16 analyze_events_funnel - Create funnel for events.
Parametrs: events_for_funnel, period_start, period_end.
#17 analyze_assessment - Analyzes assessment.
Parametrs: period_start, period_end.
#18 analyze_review - Return reviews.
Parametrs: period_start, period_end, volume.
#19 analyze_language - Analyzes what language users use.
Parametrs: period_start, period_end.
#20 analyze_bots_users - Analyzes number of users in all bots.
Parametrs: No.
There is a function to which you can pass an SQL query directly from the bot or put your query in it. This is sql_query function and it accepts only one parameter - a query.
info = lora_bot.sql_query(YOUR_SQL)
bot.send_message(message.chat.id, info, reply_markup=your_markup)
Database schema where information on users, messages, events and reviews is stored:
Enjoy!