Skip to content

Latest commit

 

History

History
73 lines (63 loc) · 3.44 KB

README.md

File metadata and controls

73 lines (63 loc) · 3.44 KB
Raw

MEDIASCOPE DATA LOADER

🛠️ Description

This Python project is designed to load data using Mediascope API

Key Features:

  • Load data even if network is not stable
  • Create settings for multiple reports and load them at once
  • Easy and versatile report settings
  • Automated slicing of a long period into parts for faster calculation and loading
  • Supports multiple target audience
  • Asynchronous data loading improves performance
  • Export data in CSV

⚙️ Setup

  • Download the package

  • Configure virtual environment. At the moment the package uses Python v3.12

  • Be sure to install all required packages

    • Run pip install -r requirements.txt to install all the requirements.

  • Configure Mediascop API connection settings

    • Create mediascope.json file and put it in settings/connections/ folder
    • Configure your Mediascope API connection settings (for more details please refer to Mediascope API settings): 
        {
    "username": "you username",
    "passw": "you password",
    "client_id": "client_id",
    "client_secret": "00000000-0000-0000-0000-000000000000",
    "auth_server": "https://auth.mediascope.net/.....",
    "root_url": "https://api.mediascope.net/...."
  }

  • Configure the report settings

    • The report settings are located in the settings/reports folder
    • Copy one of the yaml files with the predefined report settings, rename and modify it according to your needs
    • The key parameters with comments are highlighted below: 
      'report_subtype': 'DYNAMICS_BY_SPOTS', # defines the loader type
'data_lang': 'en', # language of the output data
'category_name': 'cars', # the report output directory
'folder':'raw_data', # the report output subdirectory
'multiple_files': True, # recommended to set True for periods longer that a month
............
'period': {
 'date_filter': ['2020-01-01', '2023-12-31'], # if set explicitly, 'period_num' will be ignored if set to Null, will use 'period_num' to get period from current date 
 'last_time': { # This block is used if need to calculate the date referencing to current e.g. 3 months ago
   'period_type': 'm', 
   'period_num': '3',
   'include_current': True # Try to include this month or week if it is accessible in the base
 }
 ............
    • You can set up here filters and slices following Mediascope API guides
    • Define the name and path to your report settings in the main.py as a parameter REPORT_SETTINGS = 'settings/reports/YOUR_REPORT_NAME.yaml'

  • Loader types:

    • Using CrossTab:
      • DYNAMICS_BY_SPOTS: tv spots by advertisers and brands
      • DYNAMICS_BY_SPOTS_DICT: distinct brands, advertisers to be used to clear data with data_cleaner module
      • TOP_NAT_TV_ADVERTISERS: get top advertiser on air
      • TOP_NAT_TV_PROGRAMS: top programs
      • NAT_TV_CHANNELS_BA: get buying audience dynamics per channel
    • Using TimeBand:
      • NAT_TV_CHANNELS_ATV: average time view dynamic per channel
      • NAT_TV_CHANNELS_TVR: tvr dynamic per channel
      • NAT_TV_CHANNELS_SOC_DEM: load average age, share of women and share per channel

Run the main.py