Climate Change Chatbot Analytics

Includes all software necessary for the analysis of the chatbot.

Code Entry Points

• All code to create an excel-file to analyse from a postgres Tracker Store is at the location src/climate_change_bot/analytics/import
• The script to start the plotly dash application src/climate_change_bot/analytics/dashboard.py
• Global Storage of the data as pandas dataframe src/climate_change_bot/analytics/store.py
• All pages of the application are defined in src/climate_change_bot/analytics/pages
• All components for the pages of the application are defined in src/climate_change_bot/analytics/components

Running the Code

Only described for linux.

• Install first python 3.10.
• Navigate in console to analytics directory
• Create a virtual environment and install the packages in the requirements.txt.
• Run the following code if you want to import the data from a postgres tracker store. Be aware that all model_id must be listed in model_versions.csv. Skip this step if you use an already created excel-file.

export DB_HOST=url_track_store
export DB_PORT=port_of_the_tracker_store
export DB_USER=user_tracker_store
export DB_PASSWORD=password_of_post
export DB_NAME=database_name
export OUTPUT_FILE_NAME=conversations.xlsx

cd src/climate_change_bot/analytics/import/

python3 import_events.py

• Navigate in console to analytics directory if you executed previous step
• Run the following code to start the analytic tool on http://127.0.0.1:8050/ . Set the location to the excel-file to analyse.

export PYTHONPATH="${PYTHONPATH}:your_path/analytics/src/climate_change_bot"
export FILE_NAME=../data/conversations_prod.xlsx # Change if your file is at another location

cd src/
python3 climate_change_bot/analytics/dashboard.py