Since Great Britain left the European Union on the 31st of January this year, the employees of 30 banks and 20 financial service providers are going to move to Frankfurt. At least 1500 new jobs were created in the banking sector in Frankfurt. Since a good connection to restaurants, green spaces and cultural offers is more and more important for townspeople it is important to provide information about the different boroughs and areas of the financial capital of Germany. Employers are interested in a smooth relocation of employees so that they can return to work as soon as possible.
The goal is to compare the areas of London with the boroughs of Frankfurt, using data science techniques and machine learning. The employers and employees can use the information to find places in Frankfurt that are quite similar (e.g. same diversity of venues) to their home in London.
This repo uses the Foursquare API to get information about top venues of different boroughs of Frankfurt.
I also used Folium to create an interactive map of Frankfurt and geopy to get coordinates.
Scikit-learn was used to create a classification model.
My thanks to Alex Aklson for leading the ‘Applied Data Science’-Course and for teaching different tools and techniques to make this exciting work possible!
For the London and Frankfurt data there exist Wikipedia pages that got a table of the bouroughs.
https://en.wikipedia.org/wiki/List_of_London_boroughs
https://de.wikipedia.org/wiki/Liste_der_Stadtteile_von_Frankfurt_am_Main
So I am going to scrape the page to get the data and I have to structure the data. Then I am going to add latitude and longitude using geopy and I will export the datasets as csv.
After we got the data in a right structured format, we apply some data analysis to the data to get more Insights about the two locations. More concretely I will create folium maps to get a visual feedback of the data I am working with. Then I will use Foursquare API to get the top 100 venues of the boroughs and I will apply the logistic regression machine learning algorithm to classify the boroughs. After that I will assign the London areas to the boroughs of Frankfurt.
The report contains the prepared data, assumptions, conclusion and a discussion.