CohortSignUpFirstOrder

Two Python applications: building a cohort from simulated data and counting venues in cities using Foursquare API. A report in pdf is included in "exercise-1"

• Building a cohort of Signup's vs first order's. Folder: "exercise-1".
• The number of laundry, hairdressing and fitness stores per 100k habitants. Folder: "exercise-2".

Building a cohort of Signup's vs first order's

The folder exercise-1 contains the Python code library_simulate_data.py to generate a population of user ID's and signup timestamps:

User ID	sign_up_timestamp
0102V	2018/04/02
...	...

and another random population with coherent orders of the users population:

User ID	order_timestamp
2313Q	2019/01/13
...	...

The order timestamps are generated in a period of 5 weeks after the login date for each user.

The cohort

The goal is to construct from the former two tables a cohort counting: from the users that signed up in week N, the users that made the first order in week N + k. Something like this:

Week	number	N+0 [%]	N+1 [%]	N+2 [%]	...
Week 1	58	45	25	7	...
Week 2	12	34	23	9	...
...	...	...	...	...	...

The python script main_cohort_signup.py contains the code to generate the cohort table from two pandas dataframes.

The number of laundry, hairdressing and fitness stores per 100k habitants

In this simple script I use Foursquare API requests to obtain the number of laundry, hairdressing and fitness stores per 100k habitants in two cities of France and Germany. It consists on two scripts contained in folder exercise-2: gymsLaundryBeaty_foursquareAPI.py and tableComparison.py . The output table from the latter script is in latex format. The results obtained from the API requests are stored in .csv files, and extreme benefit from the package foursquare_api_tools is warmly acknowledged (see foursquare_api_tools).