Measure local reputation of global banking

HearHere analyzes the opinion generated by banks around their branches to improve their acceptance in each community.

About

Requeriments

Install the needed libraries for the project. They are also included in requirements.txt file.

pip3 install requests
pip3 install pandas
pip3 install numpy
pip3 install bs4
pip3 install googlemaps
pip3 install twint

First steps

Open the main.ipynb file with Jupyter Notebook and follow the instructions.

A valid Google Cloud Platform API key is needed to update the database. If you have one, you can edit the api_keys.py file with it.

Files structure

Code

• main.ipynb: Jupyter file with HH queries setup and visualization
• locations_scraper.py: Downloads data from web
• twitter_scraper.py: Downloads tweets
• files_updater.py: Call the scrapers to update the databases
• google_maps_api.py: Conects HH with Google Cloud
• config.py: Project paths configuration
• api_keys.py*

Folders

• data: Data from scraping
• images

Attributions

Data sources

• BBVA - Branch and Cashier Locator - BBVA branches locations
• Twitter - Related tweets

Connected APIs:

• Google Geocoding API - Full addresses and coordinates

Road map (next steps)

• Incorporate Google Maps reviews from Place_ID (in progress)
• Improve the amount and quality of information retrieved with APIs (in progress)
• Add new banks so that they can be compared with each other
• Improve code efficiency
• Expand the tool to other sectors
• Port the tool out of the development environment

Acknowledgements

HearHere was the final project of Ironhack's Data Analysis bootcamp. Thank you to everyone who made it possible.