plada 0.0.11

PlaDa: Platform for Data market

plada

PlaDa (Platform for Data market)

PlaDa is a simulation platform designed to model and analyze data marketplaces. It provides tools for setting up market models, configuring simulation parameters, running simulations, and analyzing results.

PlaDaは、データ市場のモデル化や分析を行うためのシミュレーションプラットフォームです。PlaDaを使うことで、市場のモデル化、シミュレーションのパラメータ設定、シミュレーションの実行、結果の分析が行えます。

Install

PlaDa is available on PyPI and can be installed using pip.

PlaDaはPyPIから利用でき、pipを使ってインストールできます。

Using pip

$ pip install plada
$ python
>> import plada

Using pip in a Jupyter Notebook

When installing within a Jupyter Notebook, use the !pip command:

!pip install plada
import plada

For more detailed examples, refer to the test.ipynb notebook.

詳細については、test.ipynbをご覧ください。

Usage

Step1：Set up the Market Model.

Note: The model must be a Graph object and must contain "variables".

model = nx.read_graphml("test.graphml")

Step2：Configure Settings

Define the configuration for the simulation, market, and agents.

config = {
    "Simulation":{
        "num_iterations": 10,
        "num_steps": 10,
        "isPrice": True,
    },
    "Market":{
        "model": model,
    },
    "Agent": {
        "num_buyers": 10,
        "strategy_weights": {
            "random": 0.0,
            "related": 0.0,
            "ranking": 1.0,
        },
        "new_buyer_probability": 0.8,
    }
}

Step3：Run the Simulation

Initialize the Saver and Runner classes with the configuration settings and logger. Then, execute the main simulation process.

saver = Saver()
runner = Runner(settings=config, logger=saver)

runner.main()

Explanation

buyer.py

• Configuration of Buyer Agent
• Manages the state and strategies of buyers, updates budget, and saves purchased data.

market.py

• Market Class: Manages the overall market.
• Data Class: Manages the data.
• Variable Class: Manages the variables.
• Primarily responsible for updating prices and related market dynamics.

simulator.py

• Setup of the Purchase Simulation Workflow
• Organizes the information to be logged (requires improvement)

runner.py

• Execution of the Simulation

logger.py

• Saving the Results of the Simulation

analyzer.py

• Basic Analysis of the Results
• Visualizes the distribution of purchase counts.